Corporate information systems (CIS). The concept of networks. Corporate information systems. The structure and purpose of the CIS. Characteristic. Requirements for the organization of the corporate information system. Processes. Multilevel organization of corporate information systems Concept of corporate system and network
network of a large enterprise). Before discussing the characteristic features of each of the listed types of networks, let us dwell on the factors that force enterprises to acquire their own computer network.What gives the enterprise the use of networks
This question can be clarified as follows:
- When to deploy in an enterprise computer networks preferable to using stand-alone computers or multi-machine systems?
- What new opportunities are emerging in the enterprise with the advent of the computer network?
- Finally, does an enterprise always need a network?
If you do not go into details, then the ultimate purpose of using computer networks at the enterprise is to increase the efficiency of its work, which can be expressed, for example, in increasing profits. Indeed, if computerization reduced the cost of producing an existing product, shortened the development time for a new model, or accelerated customer service, this means that the enterprise really needed a network.
Conceptual advantage of networks, which follows from their belonging to distributed systems, before autonomous computers is their ability to perform parallel computing... As a result, in a system with several processing units, in principle, it is possible to achieve productivity exceeding the maximum currently possible performance of any separate, no matter how powerful, processor. Distributed systems potentially have a better performance / cost ratio than centralized systems.
Another obvious and important advantage of distributed systems is their higher fault tolerance... Under fault tolerance one should understand the ability of the system to perform its functions (maybe not in full) in the event of failures of individual hardware elements and incomplete data availability. Redundancy is the basis for increased fault tolerance in distributed systems. Redundancy of processing units (processors in multiprocessor systems or computers in networks) allows, if one node fails, to reassign tasks assigned to it to other nodes. To this end, dynamic or static reconfiguration procedures can be provided in a distributed system. V computer networks some datasets may be duplicated on external storage devices several computers on the network, so that if one of them fails, the data remains available.
The use of geographically distributed computing systems is more consistent with the distributed nature of applications in some subject areas, such as automation technological processes, banking, etc. In all these cases, there are separate consumers of information scattered over some territory - employees, organizations or technological installations. These consumers autonomously solve their tasks, so they should be provided with their own computing facilities, but at the same time, since the tasks they solve are logically closely interconnected, their computing facilities should be combined into a common system. The optimal solution in such a situation is the use of a computer network.
For the user, distributed systems also provide such advantages as the ability to share data and devices, as well as the ability to flexibly distribute work throughout the system. This separation of costly peripheral devices - such as high-capacity disk arrays, color printers, plotters, modems, optical disks - in many cases it is the main reason for deploying a network in an enterprise. A user of a modern computer network works at his computer, often not realizing that he is using the data of another powerful computer located hundreds of kilometers away. He sends e-mail via a modem connected to a communications server shared by several departments in his enterprise. The user has the impression that these resources are connected directly to his computer, or "almost" connected, since they require minor additional steps to work with compared to using their own resources.
Recently, another incentive for the deployment of networks began to prevail, much more important in modern conditions than the cost savings due to the division of expensive hardware or software among the employees of the corporation. This motive was the desire to provide employees with prompt access to extensive corporate information. In the face of tough competition in any sector of the market, the company ultimately wins, whose employees can quickly and correctly answer any question of the client - about the capabilities of their products, about the conditions for their use, about solving various problems, etc. In a large enterprise, even a good manager hardly knows all the characteristics of each of the manufactured products, especially since their nomenclature can be updated every quarter, if not month. Therefore, it is very important that the manager has the opportunity from his computer connected to corporate network, say, in Magadan, transfer the client's question to a server located in the central office of the enterprise in Novosibirsk, and promptly receive an answer that satisfies the client. In this case, the client will not apply to another company, but will continue to use the services of this manager.
Using the network leads to improvement communications between employees of the enterprise, as well as its customers and suppliers. Networks reduce the need for businesses to use other forms of communication, such as telephone or mail. It is often the ability to organize e-mail that is one of the reasons for deploying a computer network in an enterprise. New technologies are becoming more and more widespread, which make it possible to transfer not only computer data, but also voice and video information through network communication channels. Corporate network, which integrates data and multimedia information, can be used for organizing audio and video conferencing, in addition, on its basis, its own internal telephone network can be created.
Benefits of using networks
- An integral advantage is an increase in the efficiency of the enterprise.
- Ability to perform parallel computing, due to which productivity can be increased and fault tolerance.
- More consistent with the distributed nature of some applications.
- Sharing data and devices.
- Possibility of flexible distribution of work throughout the system.
- Online access to extensive corporate information.
- Improving communications.
Problems
- Complexity of system and application software development for distributed systems.
- Performance issues and reliability data transmission over the network.
- Security issue.
Of course, when using computer networks there are also problems associated mainly with the organization of effective interaction of individual parts of a distributed system.
First, there are software problems: operating systems ah and applications. Programming for distributed systems is fundamentally different from programming for centralized systems. So, a network operating system, performing in the general case all the functions of managing local computer resources, moreover, it solves numerous problems associated with the provision of network services. The development of network applications is complicated by the need to organize the joint work of their parts running on different machines. A lot of the hassle is also provided by ensuring the compatibility of the software installed on the network nodes.
Secondly, there are many problems associated with transporting messages over communication channels between computers. The main tasks here are to ensure reliability (so that the transmitted data is not lost or distorted) and performance (so that data exchange occurs with acceptable delays). In the structure of the total cost of a computer network, the costs of solving "transport issues" make up a significant part, while in centralized systems these problems are completely absent.
Thirdly, these are issues related to security, which are much more difficult to solve in a computer network than in a stand-alone computer. In some cases, when security is especially important, it is better to refuse to use the network.
There are many more pros and cons, but the main proof of the effectiveness of the use of networks is the indisputable fact of their ubiquity. Today it is difficult to find an enterprise that does not have at least a one-segment network of personal computers; more and more networks with hundreds of workstations and dozens of servers appear; some large organizations are acquiring private global networks that unite their branches located thousands of kilometers away. In each specific case, there were reasons for creating a network, but the general statement is also true: there is still something in these networks.
Department networks
Department networks are networks that are used by a relatively small group of employees working in one department of the enterprise. These employees perform some general tasks, such as accounting or marketing. It is believed that a department can have up to 100-150 employees.
The main goal of the department network is separation local resources such as applications, data, laser printers and modems. Typically, departmental networks have one or two file servers, no more than thirty users (Figure 10.3), and are not subnetted. Most of the enterprise traffic is localized in these networks. Department networks are usually created on the basis of any one network technology - Ethernet, Token Ring. Such a network most often uses one or at most two types of operating systems. A small number of users allows peer-to-peer network operating systems such as Windows 98 to be used in departmental networks.
Rice. 10.3.
Network management tasks at the department level are relatively simple: adding new users, fixing simple failures, installing new nodes, and installing new software versions. Such a network can be managed by an employee who devotes only part of his time to performing the duties of an administrator. Most often, the network administrator of the department does not have special training, but is the person in the department who understands computers best of all, and by itself it turns out that he is engaged in network administration.
There is another type of networks that are close to department networks - workgroup networks. These networks include very small networks, including up to 10-20 computers. The characteristics of workgroup networks are practically the same as the characteristics of departmental networks described above. Properties such as network simplicity and homogeneity are most pronounced here, while departmental networks may in some cases approach the next largest type of network - campus networks.
Campus networks
Campus networks get their name from the English word campus. It was on the territory of university campuses that it was often necessary to combine several small networks into one large one. Now this name is not associated with student campuses, but is used to refer to the networks of any enterprises and organizations.
Campus networks(Figure 10.4) unite many networks of different departments of the same enterprise within a single building or one territory covering an area of several square kilometers. However, global connections are not used on campus networks. The services of such a network include interoperability between departmental networks, access to shared enterprise databases, access to shared fax servers, high-speed modems, and high-speed printers. As a result, employees in each department of the enterprise gain access to some files and resources of networks of other departments. Campus networks provide access to corporate databases no matter what types of computers they are located on.
Rice. 10.4.
It is at the campus network level that the problems of integrating heterogeneous hardware and software arise. The types of computers, network operating systems, network hardware in each department may differ. Hence the complexities of managing campus networks. In this case, administrators should be more qualified, and the means of operational management of the network - more effective.
Enterprise networks
Corporate networks also called enterprise-wide networks, which corresponds to the literal translation of the term "enterprise-wide networks" used in the English literature to refer to this type of network. Enterprise networks ( corporate networks) unite a large number of computers in all territories of an individual enterprise. They can be complex and can cover a city, region, or even a continent. The number of users and computers can be measured in thousands, and the number of servers - in hundreds, the distances between networks of individual territories are such that you have to use corporate network different types of computers will definitely be used - from mainframes to personal computers, several types of operating systems and many different applications. Inhomogeneous parts corporate network should work as a whole, providing users with the most convenient and simple access to all the necessary resources.
Enterprise networks ( corporate networks) unite a large number of computers in all territories of an individual enterprise. For corporate network are characteristic:
- scale - thousands of user computers, hundreds of servers, huge volumes of data stored and transmitted over communication lines, a variety of applications;
- high degree of heterogeneity - different types of computers, communication equipment, operating systems and applications;
- use of global connections - the networks of branches are connected using telecommunication means, including telephone channels, radio channels, satellite communications.
The emergence corporate networks is a good illustration of the well-known postulate about the transition from quantity to quality. When connecting separate networks of a large enterprise with branches in different cities and even countries into a single network, many of the quantitative characteristics of the united network cross a certain critical threshold, beyond which a new quality begins. Under these conditions, the existing methods and approaches to solving traditional problems of smaller-scale networks for corporate networks proved to be unusable. Tasks and problems have come to the fore that in the networks of workgroups, departments and even campuses either were of secondary importance or did not appear at all. An example is the simplest (for small networks) task - maintaining credentials about network users.
The easiest way to solve this is to put the credentials of each user in the local credentials database of each computer, to which resources the user should have access. When an attempt is made to access, this data is retrieved from the local account base and, based on this, access is granted or denied. In a small network of 5-10 computers and about the same number of users, this method works very well. But if there are several thousand users on the network, each of whom needs access to several dozen servers, then, obviously, this solution becomes extremely ineffective. The administrator must repeat the operation of entering the credentials of each user several dozen times (according to the number of servers). The user himself is also forced to repeat the logon procedure every time he needs access to the resources of the new server. A good solution to this problem for a large network is to use a centralized help desk that stores Accounts all network users. The administrator performs the operation of entering user data into this database once, and the user performs the logical login procedure once, and not to a separate server, but to the entire network.
When moving from a simpler type of networks to a more complex one - from department networks to corporate network- the coverage area is increasing, it becomes more and more difficult to maintain connections between computers. As the scale of the network grows, the requirements for its reliability, performance, and functionality increase. An increasing amount of data circulates over the network, and it is necessary to ensure their safety and security, as well as availability. All this leads to the fact that corporate networks are built on the basis of the most powerful and diverse hardware and software.
Corporate network management systems have not been around for very long. One of the first systems for this purpose, which became widespread, was the SunNet Manager software product released in 1989 by SunSoft. SunNet Manager was focused on communication equipment management and network traffic control. It is these functions that are most often referred to when talking about a network management system.
INTRODUCTION
2
2
3
4
4
5
6 Fiber optic connection
6
CONCLUSION
11
Files: 1 file
INTRODUCTION | |
1 The concept of "Corporate networks" | |
2 Corporate network structure | |
3 Equipment for corporate networks | |
4 Layered representation of the corporate network | |
5 Communication channels of the corporate network | |
6 Fiber optic connection | |
CONCLUSION | |
LIST OF USED LITERATURE APPENDIX |
Introduction
Corporate network management systems have not been around for very long. One of the first systems for this purpose, which became widespread, was the SunNet Manager software product released in 1989 by SunSoft. SunNet Manager was focused on communication equipment management and network traffic control. It is these functions that are most often referred to when talking about a network management system. In addition to network management systems, there are management systems for other elements of the corporate network: OS management systems, DBMS, corporate applications. Telecommunication network management systems are also used: telephone, as well as primary networks of PDH and SDH technologies.
Regardless of the object of management, it is desirable that the management system performs a number of functions that are defined by international standards that summarize the experience of using management systems in various fields. There are recommendations ITU-T X.700 and the related standard ISO 7498-4, which divide the tasks of the management system into five functional groups:
network configuration and naming management;
error handling;
performance and reliability analysis;
safety management;
accounting of network operation.
1. The concept of "Corporate networks"
A corporate network is a system that provides information transfer between various applications used in the corporate system. A corporate network includes thousands of various components: computers of various types, system and application software, network adapters, hubs, switches and routers, and cabling. The main task of system integrators and administrators is to ensure that this cumbersome and very expensive system copes as best as possible with the processing of information flows circulating between enterprise employees and allows them to make timely and rational decisions that ensure the enterprise's survival in fierce competition. And since life does not stand still, the content of corporate information, the intensity of its flows and the ways of processing it are constantly changing. The latest example of a dramatic change in the technology of automated processing of corporate information in plain sight - it is associated with the unprecedented growth in popularity of the Internet in the last 2-3 years.
A corporate network, as a rule, is geographically distributed, i.e. uniting offices, divisions and other structures located at a considerable distance from each other. The principles by which a corporate network is built are quite different from those used to create a local network. This limitation is fundamental, and when designing a corporate network, all measures should be taken to minimize the amount of data transferred. For the rest, the corporate network should not impose restrictions on which applications and how they process the information carried over it.
The main stages of the process of creating a corporate information system can be distinguished:
conduct an information survey of the organization;
based on the results of the survey, select the architecture of the system and the hardware and software for its implementation. based on the results of the survey, select and develop the key components of the information system;
corporate database management system;
system of automation of business operations and workflow;
electronic document management system;
special software;
decision support systems.
2. The structure of the corporate network
To connect remote users to the corporate network, the simplest and most affordable option is to use a telephone connection. ISDN networks can be used where possible. In most cases, global data transmission networks are used to connect network nodes. Even where it is possible to lay leased lines (for example, within the same city), the use of packet switching technologies allows to reduce the number of necessary communication channels and - which is important - to ensure the compatibility of the system with existing global networks.
Connecting a corporate network to the Internet is justified if you need access to the appropriate services. It is worth using the Internet as a data transmission medium only when other methods are unavailable and financial considerations outweigh the requirements of reliability and security. If you will only be using the Internet as a source of information, it is best to use dial-on-demand technology. in such a way of connection, when the connection to the Internet site is established only on your initiative and at the time you need. This dramatically reduces the risk of unauthorized access to your network from the outside.
The structure of the corporate network is shown in Figure 1.
Figure 1 - Corporate network
3. Equipment for corporate networks
A corporate network is a rather complex structure that uses various types of communication, communication protocols and methods of connecting resources.
All equipment of data transmission networks can be conditionally divided into two large classes - peripheral, which is used to connect end nodes to the network, and backbone or backbone, which implements the basic functions of the network (circuit switching, routing, etc.). There is no clear boundary between these types - the same devices can be used in different capacities or combine both functions. It should be noted that high requirements are usually imposed on backbone equipment in terms of reliability, performance, number of ports, and further expandability. Peripheral equipment is a necessary component of any corporate network. The functions of backbone nodes can be assumed by the global data transmission network, to which resources are connected. As a rule, backbone nodes as part of a corporate network appear only in those cases when leased communication channels are used or their own access nodes are created.
4. Layered representation of the corporate network
It is useful to think of a corporate network as a complex system consisting of several interacting layers. At the base of the pyramid, representing the corporate network, there is a layer of computers - information storage and processing centers, and a transport subsystem (Figure 2), which ensures reliable transmission information packages between computers.
Figure 2 - Hierarchy of corporate network layers
Above the transport system, there is a layer of network operating systems that organizes the work of applications in computers and provides the resources of its computer for general use through the transport system.
Various applications work on the operating system, but due to the special role of database management systems, which store basic corporate information in an organized way and perform basic search operations on it, this class of system applications is usually separated into a separate layer of the corporate network.
At the next level, system services operate, which, using the DBMS as a search tool the information you need among the millions and billions of bytes stored on disks, provide end users with this information in a decision-friendly form, and also perform some common information processing procedures for enterprises of all types. These services include the WorldWideWeb service, e-mail system, collaboration systems, and many others.
And, finally, the upper level of the corporate network is represented by special software systems that perform tasks specific to a given enterprise or enterprises of a given type. Examples of such systems are bank automation systems, organizations accounting, computer-aided design, process control, etc.
The ultimate goal of the corporate network is embodied in top-level applications, but for them to work successfully, it is absolutely essential that the subsystems of the other layers are clearly performing their functions.
5. Communication channels of the corporate network
The first problem that has to be solved when creating a corporate network is the organization of communication channels. Communication channels - are created through communication lines using complex electronic equipment and communication cables.
Communication cable is a lengthy product of the electrical industry. There are many different types of LAN cables:
thin coaxial cables;
thick coaxial cables;
shielded twisted pairs that look like electrical wiring;
unshielded twisted pairs;
fiber optic cables that can operate over long distances and at faster speeds than other types of cables. However, their wiring and network adapters are quite expensive for them.
Communication cables (and a host of other things) are used to build communication lines. The length of communication lines ranges from tens of meters to tens of thousands of kilometers. Any more or less serious communication line, except for cables, includes: trenches, wells, couplings, crossings over rivers, sea and oceans, as well as lightning protection (as well as other types of protection) of lines.
Communication channels will be organized along the lines already built. In this case, the channels, according to the nature of the transmitted signals, can be analog or digital. So, on the same communication line, you can simultaneously create both analog and digital channels, functioning separately. Moreover, if the line, as a rule, is built and rented out all at once, then the channels are introduced gradually. It is already possible to provide communication through the line, but such use of extremely expensive structures is very ineffective. Therefore, channeling equipment is used. The number of channels is increased gradually by installing more and more powerful channelization equipment (sometimes they say - multiplexing, especially in relation to digital channels).
6. Fiber optic connection.
6.1 Optical communication systems.
Fiber optic communication lines are a form of communication in which information is transmitted over optical dielectric waveguides known as "optical fiber".
Optical fiber is currently considered the most advanced physical medium for transmitting information, as well as the most promising medium for transmitting large flows of information over long distances. The grounds to think so follow from a number of features inherent in optical waveguides.
6.2 Physical features.
1. Broadband optical signals due to extremely high carrier frequency (Fo = 10 ** 14 Hz). This means that information can be transmitted via an optical communication line at a rate of about 10 ** 12 bit / s or Terabit / s. The data transfer rate can be increased by transmitting information in two directions at once, since light waves can propagate in one fiber independently of each other.
2. Very small (in comparison with other media) attenuation of the light signal in the fiber. The best fiber samples have an attenuation of 0.22 dB / km at a wavelength of 1.55 μm, which makes it possible to build communication lines up to 100 km in length without signal regeneration.
The concept of "corporate communication system" has long been established and entrenched. And so firmly that we often stopped even thinking about its semantic (they say, semantic) content. On the eve of the autumn conference "Corporate communication systems - lessons of convergence", organized by our magazine, we propose to expand our understanding of communication networks of enterprises and institutions, and at the same time to think about further ways of their development and improvement.
And since there are probably as many points of view on corporate networks and systems as there are people involved in them, we considered it reasonable to turn directly to the "primary sources" and find out what meaning the leading Ukrainian experts fill this concept with and what the collective mind of mankind thinks about this. called the Internet.
We asked the experts, whose opinions are posted in the boxes, to focus their answers on the definition of the term "corporate communication system" and the directions of its migration at the present time.
O seemingly, corporate network - it is, first of all, the network of the enterprise. Unlike the operator's network or home network... The purpose of these networks is different. At the very least, corporate communication systems are designed to serve employees of the enterprise and do not provide any services to third-party organizations and citizens (except perhaps for personal phone calls and the use of the World Wide Web for non-production purposes). An enterprise can be both large and small, profitable or unprofitable, consisting of a single office or many branches in one country or around the world. In which case is it correct to speak of a corporate network, and in which not? Indeed, in a small enterprise at one site, we will be dealing with a relatively simple network. And if an enterprise has many geographically distributed branches, then the network can acquire a very complex architecture and developed service capabilities.
To resolve all these doubts, let's turn to the origins. Term "corporation" comes from Latin corporatio - association ... Therefore, if an enterprise consists of one office and there is nothing more to combine in it except for computers with printers, then there seems to be no need to talk about a corporation.
But let's remember that the concept of "corporate communication system", or "corporate network" (enterprise network), came to us from the West. Before that, the domestic term “ institutional or industrial communication systems ". The appearance in those days of the term PBX (office-production automatic telephone exchange) once again indicates that we are talking about enterprise networks.
Intuitively, we all understand somewhere what a corporate network is. But sometimes it is useful to plunge into more subtle philological and linguistic spheres. After all, the hour is uneven and it may turn out that many concepts are used by us only because “everyone says so,” nothing more, and their innermost meaning has long been lost.
In this regard, we will try to understand the etymology of the term "corporate communication network". What is a corporation? The Internet provides many definitions of a corporation. Let's choose the most interesting ones.
A corporation [Latin corporatio - association, community] is a form of organization of entrepreneurial activity, providing for shared ownership of participants, an independent legal status and concentration of management functions in the hands of professional managers (managers) who work for hire. Distinguish between public and private corporations.
This is probably the simplest and most accessible definition. Here's another one though.
Corporation (legal) is a common name for many types of unions that have an internal organization that unites the members of the union into one whole, who are the subject of rights and obligations, a legal entity. The expression of the will of the corporation is the general meeting of its members, the executive body is the board. Distinguish between public and private corporations. The former include territorial unions, for example, an urban, rural community, local estate unions; to the second - trade unions, commercial and industrial societies, etc., acting on the basis of special charters.
The legal definition expands rather well on the previous one.
A corporation (in social psychology) is an organized group characterized by isolation, maximum centralization and authoritarian leadership, opposing itself to other social communities on the basis of its narrowly individualistic and narrow-group interests. Interpersonal relationships in a corporation are mediated by asocial and often antisocial value orientations. Personalization of an individual in a corporation is carried out through the depersonalization of other individuals.
Well it should be twisted like that. Sounds like a prosecutor's accusatory speech (God forbid).
So a corporation is an association. Moreover, the union of companies, branches, structural divisions and even employees of one enterprise. In other words, corporate network is really a synonym enterprise networks .
Here I would like to make an important reservation. In everyday practice, they often talk about enterprise-wide networks or department. In this case, it is understood that various technical solutions, equipment and software are used for such networks. Note: this is a slightly different terminological layer that does not intersect with the subject of this article.
Corporate communication network
Having decided on the ideas about the corporation, let's move on to communication networks .
Communication network - a set of terminal devices (communication terminals), united by information transfer channels and switching devices (network nodes), providing the exchange of messages between all terminal devices.
However, talking about the communication network as a whole and not mentioning the type of information transmitted over this network will not be entirely correct. Ultimately, all existing networks are intended to transmit a certain type (or several types) of information. Enterprises most often build local area networks (LANs) and telephone networks, each of which uses its own hardware resources.
At the same time, the idea of convergence, having taken possession of the minds of engineers and equipment designers, has gathered around it advocates of comprehensive integration. The brainchild of this idea is multiservice networks, building on the winning concept of using packet networks to transmit multimedia traffic. Therefore, speaking about a corporate network, it is necessary to clarify what type of information will be transmitted in this network - data, voice, video traffic, etc. By the way, the concept of a corporate network is closely related to the concept of systems integration, as an integrated approach to automation of design, production and creation of (corporate) information networks, requiring the solution of technical problems and the implementation of organizational measures.
Large corporate communication systems unite geographically distributed divisions or branches of the enterprise. But if there is only one branch, this is just a simpler, degenerate case. In this case, the corporate network can be designed for data transmission, voice, or be multiservice. Obviously, the services available on branch networks (Internet, e-mail, voice mail, telephony, file transfer, etc.) should be fully implemented in the corporate communications network. Otherwise, it is hardly true to assert that the corporate network fully possesses this or that functionality.
So, the result of the study of the issue can be a definition that has absorbed both the points of view of experts, and opinions borrowed from the Internet, and our own reasoning, namely:
A corporate network (aka departmental) is a communication network used to transfer various types of information within a company or group of companies (corporation) and is not used to provide commercial communication services to third-party organizations and individuals... Such networks are deployed both on the basis of their own infrastructure and using the resources provided by telecommunications operators.
What should be the corporate communications network?
Why does an enterprise need a communications network at all? The question is rhetorical. Probably in order to provide employees of the enterprise with the opportunity perform their duties productively ... This is especially true in the presence of an aggressive competitive environment. A high-quality communication system increases labor productivity through the implementation of a wide range of various services, as well as by ensuring the effective functioning of the information infrastructure of the enterprise.
Architecture and possibilities a corporate network depends on the tasks that are entrusted to it, on the size of the enterprise and the specifics of its activities, as well as on the prospects for further expansion. Currently, the corporate network of a small enterprise contains, as a rule, one or two components - telephone and data transmission. Moreover, telephone services can be implemented directly through a local telecommunications operator (without installing a PBX), and computers are connected into a small local network with Internet access in any available way.
We see that telephony and data transfer small businesses are initially separated. As the enterprise grows, each network develops, but still independently of each other. PBX is added, servers and databases, firewalls and call centers appear. But the voice still (for the time being) remains separated from the data transmission.
Proponents of unification will rightly point out that there are many solutions for small enterprises of the SOHO level, involving the use of IP channels for both telephony and data transmission. Indeed, such solutions can be quite effective, for example, when organizing a remote office. But we will come to this question a little later.
Despite the well-known conservatism of employees of technical departments of enterprises, the principles convergence , the use of a single medium for the transmission of heterogeneous traffic, find more and more adherents. But are all enterprises ready to implement a single multiservice network? Most likely, the answer will be no. And, by and large, this is not a question at all. Indeed, often the enterprise has already built two separate networks, each based on the traditional native architecture and equipment. In most cases, there is no question of using a single IP environment for voice and data transmission within an enterprise. To make such a decision, there must be either sufficiently weighty economic arguments , or arguments of a different kind - convenience, savings in maintenance, or something else.
Enterprise networks of the future
If we are talking only about data transmission and telephony services, then we ourselves are undoubtedly in the captivity of old paradigms. After all, the list of services that can be organized and provided to subscribers of the corporate network is much wider. It is worth recalling video conferencing systems, a single universal mailbox (Unified Messaging), a DECT microcellular communication system. Currently, the issue of convergence of mobile and fixed communication services is quite acute, especially since many manufacturers offer such solutions both at the operator and at the corporate level (see publications in "S&B", 2006, No. 4, p. 78 - 81, “New horizons of corporate communications”, as well as “S&B”, 2006, No. 4, pp.82–85, “FMC, or the New Paradigm of the Convergence Era”). After a while, it will be appropriate to talk about the use of Wi MAX in corporate networks.
The corporate network of the future is an integrated environment that provides various types of services - traditional data transmission, telephony, video conferencing and video broadcasting, access control, security and video surveillance. The necessary components of a corporate network are mobile access and advanced data transmission security.
When discussing the feasibility of certain solutions offered by manufacturers, first of all, one should talk about the possibility and efficiency of performing the production tasks that the enterprise faces. Obviously, the tasks solved in different sectors of the economy differ from each other. Therefore, communication networks of oblenergos, railways, banks, government bodies have their own characteristics. At a certain stage, when the enterprise becomes sufficiently large and cumbersome, proposals for the creation of joint multiservice networks transmitting multimedia traffic. When the future starts knocking on the door more and more insistently, it is quite appropriate to build multi-service corporate next generation networks ... In this case, the enterprise creates a single network designed to transmit heterogeneous traffic. Processing of each type of traffic, as you might expect, falls on specialized systems, often traditional computing resources (servers) with the appropriate software. In this case, the data traffic is locked on servers and in databases. Voice traffic will be consolidated into an IP-PBX. Video traffic - on video conferencing servers. It is not surprising that specialized application servers will be used to handle different types of traffic.
Technology does not stand still, and creative thought cannot be stopped at all. Time will pass, and the traditional ways of organizing corporate communication systems will be replaced by more modern ones, providing the deployment of a whole range of new services and new applications. These solutions will pave the way for business and IT leaders. The victory of the new generation multiservice networks will be conditioned, first of all, by the prospects that they will open for business. In this case, the cost of the solution will cease to play a decisive role. After all, once the advantage of replacing a bicycle with a car was also questioned. But time has made its own adjustments. For the new opportunities provided by modern communication systems will be an order of magnitude higher than those offered today.
Who doubts that time is the most powerful innovation factor?
Vladimir SKLYAR
|
“... A promising direction of development
|
|
The modern corporate communication system today consists of a universal network infrastructure and intelligent services that guarantee the effective integration of communication systems and business processes of the enterprise. The versatility of the infrastructure allows you to increase the speed of information exchange by using the most suitable transmission medium. |
|
"... Give me a connection, period ..." |
|
The very concept of "corporate communication system" has not undergone any significant transformations and, as before, implies a complex of technical, organizational, technical and organizational solutions and measures to ensure sustainable management of corporate forces and funds, as well as interaction with other structures through their corporate communication networks and / or public communication networks. |
|
“… Corporate communication system as
|
|
The modern corporate communication system includes the following interconnected components: |
|
"... Corporate communication systems
|
|
The corporate communication system is one of the main systems that ensure the efficiency of the business of any company. It must solve several key tasks, namely: to increase the efficiency of employees' work, optimizing the interaction between them and providing effective means of communication; improve the quality of interaction with the company's clients, ensuring high-quality processing and distribution of external calls; and lower operating costs through IP solutions, effective management tools, and minimized downtime. The modern corporate communication system today is not just a telephone communication system and a data transmission network. Such a system should be an integrated environment aimed at solving all communication problems of users, regardless of their location (inside or outside the office) and the available means of communication at their disposal. Corporate communication systems are developing towards the convergence of services and the provision of new communication opportunities that become available to users. These are video conferencing, collaboration on documents, indication of availability in real time, etc. As many companies increasingly employ people away from their offices, the demands on corporate mobility functions are increasing. The convergence of communication services in action may look, for example, as the ability to use all business telephony functions (dialing an internal office subscriber by name, call forwarding, conference calls, etc.) available in the office on a desktop device, also from a mobile phone outside the office over a GSM or Wi-Fi network; or access to corporate e-mail and the availability of colleagues both from a web browser and using a communicator while traveling, and so on. |
|
“... an employee of a modern corporation
|
|
The essence of any modern technology is the ability and ability to migrate. This also applies to communication systems. From a large, heavy and very expensive "piece of iron" with a manufacturer's oath of "investment protection" and the ability to modernize - to easy and flexible solutions. Only the approach was not settled: many multi-tasking systems under one control and monitoring, or one "multifunctional harvester". |
|
“... Flexibly and efficiently provide
|
|
Today, scientific and technological progress, especially in the field of IT technologies, is extremely rapid. And whatever function we try to designate as an indicator that the communication system under consideration is modern, as a new, more modern function or technology appears. Communication systems are developing very quickly. Therefore, I would nevertheless become attached to the needs of the corporation's business. That is, a communication system can be considered modern if it allows you to flexibly and quickly solve all the "continuously growing" tasks of the company's business. 1. Almost all large corporations want to have not a network of disparate subsystems (which we affectionately call a "zoo"), but a single geographically distributed telecommunications system. Such a system is easier to monitor, administer, and provide security, and license, and scale, and increase functionality, etc., etc. It is more flexible and allows for quick readjustment to the changing business environment of the company. Yesterday we were proud of the unified systems, consisting of only 7 subdivisions, dispersed throughout Ukraine. And today some of our unified communication systems already have more than 200 of them. Imagine the scale of the problem if, for example, you want to update a similarly sized system of individual PBXs. If there are 250 working days in a year, then this is at least a year. In our case (when the system is unified), such a procedure will take only a few minutes. |
|
“… Communication as a control system module |
|
In my opinion, give an unambiguous definition of modern corporate communications rather difficult, since this concept includes many aspects. |
Introduction. From the history of network technologies. 3
The concept of "Corporate networks". Their main functions. 7
Technologies used to create corporate networks. fourteen
The structure of the corporate network. Hardware. 17
Methodology for creating a corporate network. 24
Conclusion. 33
List of used literature. 34
Introduction.
From the history of network technologies.
The history and terminology of corporate networks is closely related to the history of the birth of the Internet and the World Wide Web. Therefore, it does not hurt to remember how the very first network technologies appeared, which led to the creation of modern corporate (departmental), territorial and global networks.
The Internet began in the 1960s as a US Department of Defense project. The increased role of the computer has given rise to the need for both the separation of information between different buildings and local networks, and the maintenance of the overall performance of the system in the event of failure of individual components. The Internet is based on a set of protocols that allow WANs to send and transmit information to each other independently; if one node of the network is unavailable for some reason, the information reaches its final destination through other nodes that are currently operational. The protocol developed for this purpose is called the Internetworking Protocol (IP). (The acronym TCP / IP stands for the same.)
Since then, the IP protocol has become generally accepted in the military as a way to make information publicly available. Since many of these agencies' projects were carried out in various research groups at universities around the country, and the way information was exchanged between heterogeneous networks proved to be very effective, the application of this protocol quickly spread outside the military departments. It began to be used both in NATO research institutes and in European universities. Today, the IP protocol, and therefore the Internet, is the universal global standard.
In the late eighties, the Internet faced a new challenge. At first, the information was either emails or simple data files. For their transmission, appropriate protocols were developed. Now there is whole line files of a new type, usually united under the name multimedia, containing both images and sounds, and hyperlinks that allow users to navigate both within one document and between different documents containing related information.
In 1989, the Laboratory of Elementary Particle Physics of the European Center for Nuclear Research (CERN) successfully launched a new project, the purpose of which was to create a standard for the transmission of this kind of information over the Internet. The main components of this standard were the multimedia file formats, hypertext files, and the protocol for receiving such files over the network. The file format was named HyperText Markup Language (HTML). It was a simplified version of the more general Standard General Markup Language (SGML). The request service protocol is called HyperText Transfer Protocol (HTTP). In general, it looks like this: the server running the program serving the HTTP protocol (HTTP demon) sends HTML files at the request of Internet clients. These two standards formed the basis for a fundamentally new type of access to computer information. Standard multimedia files can now not only be retrieved at the user's request, but also exist and be displayed as part of another document. Since the file contains hyperlinks to other documents that may be located on other computers, the user can access this information with a slight click of the mouse button. This fundamentally removes the complexity of accessing information in a distributed system. Multimedia files in this technology are traditionally called pages. A page is also called information that is sent to the client machine in response to each request. The reason for this is that a document usually consists of many separate parts linked by hyperlinks. This division allows the user to decide for himself which parts he wants to see in front of him, saves him time and reduces network traffic. A software product that is used directly by a user is usually called a browser (from the word browse - to graze) or a navigator. Most of them allow you to automatically get and display a specific page that contains links to documents that the user accesses most often. This page is called the home page and is usually accessed by a separate button. Each non-trivial document is usually supplied with special page, similar to the "Contents" section of the book. This is where the study of the document usually begins, which is why it is also often called the home page. Therefore, in general, a home page is understood as a certain index, an entry point into information of a certain kind. Usually the name itself includes a definition of this section, for example, Microsoft Home Page. On the other hand, each document can be accessed from many other documents. All the space of documents referring to each other on the Internet was called the World Wide Web (the world wide web is an acronym for WWW or W3). The document system is completely distributed, and the author does not even have the ability to trace all the links to his document that exist on the Internet. The server providing access to these pages can register all those who read such a document, but not those who link to it. The situation is the opposite of the existing printed matter in the world. Many research areas have periodically published indexes of articles on a topic, but it is impossible to trace everyone who reads a given document. Here we know those who read (had access) to the document, but we do not know who referred to it. interesting feature consists in the fact that with such a technology it becomes impossible to keep track of all the information available through the WWW. Information appears and disappears continuously, in the absence of any central control. However, this should not be intimidated, the same is happening in the world of printed materials. We do not try to accumulate old newspapers if we have fresh ones every day, and the effort is negligible.
Client software products that receive and display HTML files are called browsers. The first graphical browser was called Mosaic and was made at the University of Illinois. Many of the modern browsers are based on this product. However, due to the standardization of protocols and formats, any compatible software product can be used. Viewers exist on most major client systems capable of supporting smart windows. These include MS / Windows, Macintosh, X-Window systems, and OS / 2. There are also viewing systems for those operating systems where windows are not used - they display text fragments of documents that are being accessed on the screen.
The presence of viewers on such heterogeneous platforms is essential. The operating environments on the author's machine, server, and client are independent of each other. Any client can access and view documents created using HTML and related standards, and transmitted through an HTTP server, regardless of the operating environment in which they were created or where they came from. HTML also supports form design and feedback functions. This means that the user interface, both when querying and retrieving data, allows you to go beyond the point-and-click principle.
Many stations, including Amdahl, have written interfaces for interaction between HTML forms and legacy applications, creating a universal client user interface for the latter. This makes it possible to write client-server applications without thinking about client-level coding. In fact, programs are already emerging in which the client is viewed as a viewer. An example is Oracle Corporation's WOW interface, which replaces Oracle Forms and Oracle Reports. Although this technology is still very young, it is already capable of changing the situation in the field of information management as much as the use of semiconductors and microprocessors in their time changed the world of computers. It allows us to turn functions into separate modules and simplify applications, taking us to a new level of integration that is more in line with the business functions of the enterprise.
Information overload is the curse of our time. The technologies that were created to alleviate this problem have only made it worse. This is not surprising: it is worth looking at the contents of the trash cans (regular or electronic) of an ordinary employee dealing with information. Even aside from the inevitable heap of advertising "junk" in the mail, most of the information is sent to such an employee just "in case" he needs it. Add to this the "out of date" information, which is likely to be needed, but later - and here's the main contents of the trash can. The employee will most likely store half of the information that "may be needed" and all the information that is likely to be needed in the future. When the need arises, he will have to deal with a cumbersome, poorly structured archive personal information, and at this stage additional difficulties may arise due to the fact that it is stored in files of different formats on different media. The advent of photocopiers has made the situation of "suddenly needed" information even worse. The number of copies, instead of decreasing, only increases. Email only made the problem worse. Today, the "publisher" of information can create his own, personal mailing list and, with the help of one command, send an almost unlimited number of copies "in case" they may be needed. Some of these distributors of information understand that their lists are worthless, but instead of fixing them, they put a note at the beginning of the message with the following content: "If you are not interested ..., destroy this message." The letter will still hammer mailbox, and the addressee, in any case, will have to spend time familiarizing with it and destroying it. The exact opposite of "might be useful" information is "timely" information, or information that is in demand. They expected help from computers and networks in working with this type of information, but so far they have not been able to cope with it. There used to be two main methods of delivering timely information.
When using the first of them, information was distributed between applications and systems. To gain access to it, the user had to study and then constantly perform many complex access procedures. When access was obtained, each application required its own interface. When faced with such difficulties, users usually simply refused to receive timely information. They were able to master access to one or two applications, but they were not enough for the rest.
To solve this problem, some enterprises tried to accumulate all the distributed information on one main system. As a result, the user received a single access method and a single interface. However, since in this case all the requests of the enterprise were processed centrally, these systems grew and became more complex. More than ten years have passed, and many of them are still not filled with information due to the high cost of entering and maintaining it. There were other problems here as well. The complexity of such unified systems made them difficult to modify and use. To support the discrete data of transaction processes, tools have been developed to manage such systems. Over the past decade, the data we deal with has become much more complex, making it difficult to maintain information. The changing nature of information needs and how difficult it is to change in this area has given rise to these large, centrally managed systems that inhibit query execution at the enterprise level.
Web technology offers a new approach to delivering information on demand. Since it supports authorization, publication and management of distributed information, new technology does not lead to the complexity of old centralized systems. Documents are written, maintained, and published directly by authors without having to ask programmers to create new data entry forms and reporting programs. With the new viewing systems, the user can access and view information from distributed sources and systems using a simple, unified interface, without having the slightest idea of the servers they are actually accessing. These simple technological changes will revolutionize information infrastructures and fundamentally change the way our organizations operate.
The main distinguishing feature of this technology is that the control of the information flow is in the hands not of its creator, but of the consumer. If the user has the ability to easily receive and view information as needed, it will no longer have to be sent to him "in case" if it is needed. The publishing process can now be independent of the automatic dissemination of information. This includes forms, reports, standards, meeting scheduling, sales support tools, training materials, schedules, and a host of other documents that tend to clog our trash cans. For the system to work, as mentioned above, you need not only a new information infrastructure, but also a new approach, a new culture. As creators of information, we must learn to publish it without distributing it, as users - to be more responsible in identifying and tracking their information requests, actively and effectively receiving information when we need it.
The concept of "Corporate networks". Their main functions.
Before talking about private (corporate) networks, you need to define what these words mean. Recently, this phrase has become so widespread and fashionable that it began to lose its meaning. In our understanding, a corporate network is a system that provides information transfer between various applications used in a corporation's system. Based on this completely abstract definition, we will consider various approaches to creating such systems and will try to fill the concept of a corporate network with concrete content. At the same time, we believe that the network should be as versatile as possible, that is, allow the integration of existing and future applications with the lowest possible costs and restrictions.
A corporate network, as a rule, is geographically distributed, i.e. uniting offices, divisions and other structures located at a considerable distance from each other. Often, corporate network nodes are located in different cities and sometimes countries. The principles by which such a network is built are quite different from those used to create a local network, even covering several buildings. The main difference is that geographically distributed networks use rather slow (today - tens and hundreds of kilobits per second, sometimes up to 2 Mbit / s) leased communication lines. If, when creating a local network, the main costs are for the purchase of equipment and laying of cables, then in wide-area networks the most significant cost element is the rent for the use of channels, which grows rapidly with an increase in the quality and speed of data transmission. This limitation is fundamental, and when designing a corporate network, all measures should be taken to minimize the amount of data transferred. For the rest, the corporate network should not impose restrictions on which applications and how they process the information carried over it.
By applications, we mean here both system software - databases, mail systems, computing resources, file service, etc. - and the means with which the end user works. The main tasks of a corporate network are the interaction of system applications located in various nodes, and access to them by remote users.
The first problem that has to be solved when creating a corporate network is the organization of communication channels. If within the same city you can count on leasing dedicated lines, including high-speed ones, then when moving to geographically distant nodes, the cost of leasing channels becomes simply astronomical, and their quality and reliability often turns out to be very low. A natural solution to this problem is to use existing global networks. In this case, it is sufficient to provide channels from offices to the nearest network nodes. In this case, the global network will take over the task of delivering information between nodes. Even when creating a small network within one city, one should keep in mind the possibility of further expansion and use technologies that are compatible with existing global networks.
Often the first, if not the only, such network that comes to mind is the Internet. Using the Internet in corporate networks Depending on the tasks being solved, the Internet can be considered at different levels. For the end user, it is primarily a worldwide system for the provision of information and postal services. The combination of new technologies for access to information, united by the concept of the World Wide Web, with the cheap and generally accessible global computer communication system, the Internet, has actually spawned a new mass media, which is often referred to simply as the Net. Anyone who connects to this system perceives it simply as a mechanism that gives access to certain services. The implementation of this mechanism turns out to be absolutely insignificant.
When using the Internet as the basis for a corporate data network, a very interesting thing emerges. It turns out that the Network is not a network. This is precisely the Internet - the interconnection. If we look inside the Internet, we will see that information passes through many completely independent and mostly non-profit nodes, connected through the most diverse channels and data networks. The rapid growth of services provided on the Internet leads to overloading of nodes and communication channels, which sharply reduces the speed and reliability of information transmission. At the same time, Internet service providers do not bear any responsibility for the functioning of the network as a whole, and communication channels are developing extremely unevenly and mainly where the state considers it necessary to invest in this. Accordingly, there is no guarantee of the quality of the network, the speed of data transfer, or even the simple reachability of your computers. For tasks in which reliability and guaranteed delivery time of information are critical, the Internet is far from the best solution. In addition, the Internet binds users to a single protocol - IP. It's good when we use standard applications working with this protocol. The use of any other systems with the Internet is not easy and expensive. If you need to provide mobile users with access to your private network, the Internet is also not the best solution.
It would seem that there should be no big problems here - Internet service providers are almost everywhere, take a laptop with a modem, call and work. However, a supplier, say, in Novosibirsk, has no obligation to you if you are connected to the Internet in Moscow. He does not receive money for services from you and, of course, will not provide access to the network. Either you need to conclude an appropriate contract with him, which is hardly reasonable if you are on a two-day business trip, or call from Novosibirsk to Moscow.
Another Internet issue that has been widely discussed in recent years is security. If we are talking about a private network, it seems quite natural to protect the transmitted information from prying eyes. The unpredictability of information paths between many independent Internet sites not only increases the risk that some overly curious network operator might put your data on disk (technically not that difficult), but also makes it impossible to determine the location of the information leak. Encryption solutions only partially solve the problem, since they are mainly applicable to mail, file transfers, etc. Solutions that allow encrypting information in real time with an acceptable speed (for example, when working directly with a remote database or file server) are inaccessible and expensive. Another aspect of the security problem is again related to the decentralization of the Internet - there is no one who can restrict access to the resources of your private network. Since this is an open system where everyone can see everyone, anyone can try to get into your office network and access data or programs. There are, of course, means of protection (for them, the name Firewall is adopted - in Russian, more precisely in German "firewall" - fire wall). However, they should not be considered a panacea - think about viruses and antivirus programs. Any protection can be broken, as long as it pays for the cost of hacking. It should also be noted that you can make a system connected to the Internet inoperable without intruding on your network. There are known cases of unauthorized access to the management of network nodes, or simply using the features of the Internet architecture to disrupt access to a particular server. Thus, the Internet cannot be recommended as the basis for systems that require reliability and closedness. Connecting to the Internet within a corporate network makes sense if you need access to that vast information space, which is actually called the Network.
A corporate network is a complex system that includes thousands of various components: computers of different types, from desktop to mainframe, system and application software, network adapters, hubs, switches and routers, cabling. The main task of system integrators and administrators is to ensure that this cumbersome and very expensive system copes as best as possible with the processing of information flows circulating between the employees of the enterprise and allows them to make timely and rational decisions that ensure the survival of the enterprise in fierce competition. And since life does not stand still, the content of corporate information, the intensity of its flows and the ways of processing it are constantly changing. The latest example of a dramatic change in the technology of automated processing of corporate information in plain sight - it is associated with the unprecedented growth in popularity of the Internet in the last 2-3 years. The changes brought about by the Internet are multifaceted. The WWW hypertext service has changed the way information is presented to a person, collecting all its popular types on its pages - text, graphics and sound. Internet transport - inexpensive and accessible to almost all enterprises (and through telephone networks and single users) - greatly facilitated the task of building a territorial corporate network, while simultaneously highlighting the task of protecting corporate data when it is transmitted through a highly publicly accessible public network with a multi-million "population. ".
Technologies used in corporate networks.
Before setting out the basics of the methodology for building corporate networks, it is necessary to give a comparative analysis of technologies that can be used in corporate networks.
Modern data transmission technologies can be classified according to data transmission methods. In general, there are three main methods of data transmission:
channel switching;
switching messages;
packet switching.
All other methods of interaction are, as it were, their evolutionary development. For example, if we represent data transmission technologies in the form of a tree, then the packet switching branch will be divided into frame switching and cell switching. Recall that packet switching technology was developed over 30 years ago to reduce overhead and improve the performance of existing data transmission systems. The first packet switching technologies, X.25 and IP, were designed with poor quality communication channels in mind. With the improvement in quality, it became possible to use a protocol such as HDLC for transferring information, which found its place in Frame Relay networks. The desire to achieve greater performance and technical flexibility was the impetus for the development of SMDS technology, the possibilities of which were then extended by the standardization of ATM. One of the parameters by which one can compare technologies is the guarantee of information delivery. Thus, X.25 and ATM technologies guarantee reliable delivery of packets (the latter using the SSCOP protocol), while Frame Relay and SMDS operate in a mode where delivery is not guaranteed. Further, the technology can ensure that the data arrives at the recipient in the sequence it was sent. Otherwise, the order must be restored on the receiving side. Packet-switched networks can be pre-connection oriented or simply transfer data to the network. In the first case, both permanent and dial-up virtual connections can be supported. Other important parameters are the availability of data flow control mechanisms, traffic management systems, congestion detection and prevention mechanisms, etc.
Technology comparisons can also be made based on criteria such as the effectiveness of the addressing scheme or routing methods. For example, the addressing used may be geo-based (telephone numbering plan), WAN use, or Hardware... For example, IP uses a 32-bit logical address that is assigned to networks and subnets. The E.164 addressing scheme is an example of a geo-based addressing scheme, and the MAC address is an example of a hardware address. X.25 technology uses a Logical Channel Number (LCN), and the virtual switched connection in this technology uses the X.121 addressing scheme. In Frame Relay technology, multiple VCs can be embedded in one channel, with a single VC being identified by a Data-Link Connection Identifier (DLCI). This identifier is indicated in each transmitted frame. DLCI has local meaning only; in other words, the sender's virtual channel can be identified by one number, while the receiver's can be identified by a completely different one. Virtual dial-up connections in this technology are based on the E.164 numbering scheme. ATM cell headers contain unique VCI / VPI identifiers that change as the cells pass through intermediate switching systems. ATM switched virtual connections can use E.164 or AESA addressing schemes.
Routing of packets on a network can be done statically or dynamically and be either a standardized mechanism for a particular technology or act as a technical basis. Examples of standardized solutions include dynamic routing protocols OSPF or RIP for IP. In relation to ATM technology, the ATM Forum has defined the PNNI switched virtual connection request routing protocol, the distinctive feature of which is the accounting of quality of service information.
The ideal option for a private network would be to create communication channels only in those areas where it is necessary, and transfer over them any network protocols that are required by running applications. At first glance, this is a return to leased communication lines, but there are technologies for building data transmission networks that allow organizing channels inside them that arise only at the right time and in the right place. Such channels are called virtual. A system that combines remote resources using virtual channels can naturally be called a virtual network. Today, there are two main technologies for virtual networks - circuit-switched networks and packet-switched networks. The former include the conventional telephone network, ISDN and a number of other, more exotic technologies. Packet-switched networks are represented by X.25 technologies, Frame Relay and, more recently, ATM. It is too early to talk about the use of ATM in wide area networks. The rest of the types of virtual (in various combinations) networks are widely used in the construction of corporate information systems.
Circuit-switched networks provide a subscriber with several communication channels with a fixed bandwidth per connection. A well-known telephone network provides one communication channel between subscribers. If you need to increase the number of simultaneously available resources, you have to install additional phone numbers, which is very expensive. Even if we forget about the low quality of communication, the limitation on the number of channels and the long connection establishment time do not allow using telephone communication as the basis of a corporate network. For connecting individual remote users, this is a fairly convenient and often the only method available.
Another example of a circuit-switched virtual network is ISDN (Integrated Services Digital Network). ISDN provides digital channels (64 kbps) that can carry both voice and data. A Basic Rate Interface (ISDN) connection includes two of these channels and an additional 16 kbps control channel (this combination is referred to as 2B + D). Use is possible more channels - up to thirty (Primary Rate Interface, 30B + D), but this leads to a corresponding rise in the cost of equipment and communication channels. In addition, the cost of renting and using the network proportionally increases. In general, the restrictions on the number of concurrently available resources imposed by ISDN lead to the fact that this type of communication turns out to be convenient to use mainly as an alternative to telephone networks. In systems with a small number of nodes, ISDN can also be used as the main network protocol. It should only be borne in mind that access to ISDN in our country is the exception rather than the rule.
An alternative to circuit-switched networks is packet-switched networks. When using packet switching, one communication channel is used in a time-sharing mode by many users - much the same as on the Internet. However, unlike networks such as the Internet, where each packet is routed separately, packet-switched networks require establishing a connection between end resources before transmitting information. After establishing a connection, the network "remembers" the route (virtual channel) through which information should be transmitted between subscribers and remembers it until it receives a signal that the connection is broken. For applications operating in a packet switching network, virtual circuits look like ordinary communication lines - with the only difference that their bandwidth and introduced delays vary depending on the network congestion.
The classic packet switching technology is the X.25 protocol. Nowadays it is customary to wrinkle your nose at these words and say: "it is expensive, slow, outdated and not fashionable." Indeed, there are virtually no X.25 networks today that use speeds above 128 kbps. The X.25 protocol includes powerful error correction tools to ensure reliable delivery of information even on bad links and is widely used where quality communication links are not available. In our country, they are not almost everywhere. Naturally, reliability comes at a price — in this case, the speed of the network equipment and the relatively large — but predictable — delays in the dissemination of information. At the same time, X.25 is a universal protocol that allows you to transfer almost any type of data. It is "natural" for X.25 networks to run applications that use the OSI protocol stack. These include systems using the X.400 (e-mail) and FTAM (file exchange) standards, among others. Tools are available to implement OSI-based interoperability Unix systems ... Another standard X.25 networking capability is communication over common asynchronous COM ports. Figuratively speaking, an X.25 network lengthens the cable connected to a serial port, bringing its connector to remote resources. Thus, virtually any application that can be accessed through a COM port can be easily integrated into an X.25 network. As examples of such applications, we should mention not only terminal access to remote host computers, for example, Unix machines, but also the interaction of Unix computers with each other (cu, uucp), systems based on Lotus Notes, e-mail cc: Mail and MS Mail, etc. For LAN interconnection in nodes connected to the X.25 network, there are methods of packing ("encapsulating") information packets from the local network into X.25 packets. Some of the service information is not transmitted in this case, since it can be unambiguously restored on the recipient's side. The standard encapsulation mechanism is the one described in RFC 1356. It allows different LAN protocols (IP, IPX, etc.) to be transmitted simultaneously over a single virtual connection. This mechanism (or an older IP-only implementation of RFC 877) is implemented in almost all modern routers. There are also transmission methods for X.25 and other communication protocols, notably SNA, which is used in IBM mainframe networks, and a number of proprietary protocols from various manufacturers. Thus, X.25 networks offer a universal transport mechanism for transferring information between virtually any application. At the same time, different types of traffic are transmitted over the same communication channel, without knowing anything about each other. By interconnecting LANs over X.25, it is possible to isolate individual parts of the corporate network from each other, even if they use the same communication lines. This facilitates the solution of security and access control problems that inevitably arise in complex information structures. In addition, in many cases there is no need to use complex routing mechanisms, leaving this task to the X.25 network. Today, there are dozens of public X.25 global networks in the world, their nodes are located in almost all major business, industrial and administrative centers. In Russia, X.25 services are offered by Sprint Network, Infotel, Rospak, Rosnet, Sovam Teleport and a number of other providers. In addition to connecting remote sites, X.25 networks always provide means of access for end users. In order to connect to any resource on the X.25 network, the user only needs to have a computer with an asynchronous serial port and a modem. At the same time, there are no problems with authorizing access to geographically remote nodes - firstly, X.25 networks are sufficiently centralized and by concluding an agreement, for example, with Sprint Network or its partner, you can use the services of any of the Sprintnet nodes - and these are thousands of cities around the world, including more than a hundred in the former USSR. Secondly, there is a protocol of interaction between different networks (X.75), which also takes into account payment issues. Thus, if your resource is connected to an X.25 network, you can access it both from your vendor's nodes and through nodes on other networks - that is, from virtually anywhere in the world. From a security perspective, X.25 networks offer some very attractive capabilities. First of all, due to the very structure of the network, the cost of intercepting information on an X.25 network turns out to be high enough to already serve as a good defense. The problem of unauthorized access can also be solved quite effectively by means of the network itself. If any - however small - the risk of information leakage turns out to be unacceptable, then, of course, it is necessary to use encryption tools, including in real time. Today, there are encryption tools designed specifically for X.25 networks that can operate at fairly high speeds - up to 64 kbps. Such equipment is manufactured by Racal, Cylink, Siemens. There are also domestic developments created under the auspices of the FAPSI. The disadvantage of X.25 technology is that there are a number of fundamental speed limits. The first of them is associated precisely with the developed capabilities of correction and recovery. These tools cause delays in the transmission of information and require a lot of processing power and performance from the X.25 hardware, as a result of which it simply "cannot keep up" with fast communication lines. Although there is equipment that has two megabit ports, the actual speed provided by it does not exceed 250 - 300 kbps per port. On the other hand, for modern high-speed communication lines, X.25 correction means are redundant and when they are used, the equipment power is often idle. The second feature that makes X.25 networks to be considered slow is the encapsulation of the LAN protocols (primarily IP and IPX). All other things being equal, the connection of local networks over X.25 is, depending on the parameters of the network, 15-40 percent slower than when using HDLC over a dedicated line. Moreover, the worse the communication line, the higher the loss of productivity. Again, we are dealing with an obvious redundancy: LAN protocols have their own means of correction and recovery (TCP, SPX), but when using X.25 networks, you have to do it again, losing speed.
It is on these grounds that X.25 networks are declared slow and obsolete. But before talking about the fact that any technology is outdated, it should be indicated for what applications and in what conditions. On low quality links, X.25 networks are quite efficient and offer significant cost and performance gains over leased lines. On the other hand, even if you can count on the rapid improvement in the quality of communication - a prerequisite for the obsolescence of X.25 - then the investment in X.25 equipment will not be lost, since modern equipment includes the possibility of moving to Frame Relay technology.
Frame Relay networks
Frame Relay technology emerged as a means to realize the benefits of packet switching on high-speed lines. The main difference between Frame Relay and X.25 networks is that they exclude error correction between network nodes. The tasks of restoring the information flow are assigned to the terminal equipment and user software. Naturally, this requires the use of sufficiently high-quality communication channels. It is believed that for successful work with Frame Relay, the probability of an error in the channel should be no worse than 10-6 - 10-7, i.e. no more than one bad bit per several million. The quality provided by conventional analog lines is usually one to three orders of magnitude lower. The second difference between Frame Relay networks is that today almost all of them implement only the mechanism of permanent virtual connections (PVC). This means that when connecting to a Frame Relay port, you must determine in advance which remote resources you will have access to. The principle of packet switching - many independent virtual connections in one communication channel - remains here, but you cannot choose the address of any subscriber of the network. All resources available to you are determined when you configure the port. Thus, on the basis of Frame Relay technology, it is convenient to build closed virtual networks used to transfer other protocols by means of which routing is carried out. A "closed" virtual network means that it is completely inaccessible to other users on the same Frame Relay network. For example, in the United States, Frame Relay networks are widely used as the backbone for the Internet. However, your private network can use Frame Relay VCs on the same lines as Inernet traffic - and be completely isolated from it. Like X.25 networks, Frame Relay provides a universal transmission medium for almost any application. The main area of application for Frame Relay today is the interconnection of remote LANs. In this case, error correction and information recovery are performed at the level of LAN transport protocols - TCP, SPX, etc. Losses for encapsulation of LAN traffic in Frame Relay do not exceed two to three percent. Frame Relay LAN encapsulation is described in RFC 1294 and RFC 1490. RFC 1490 also defines Frame Relay for SNA traffic. The ANSI T1.617 Annex G specification describes the use of X.25 over Frame Relay networks. This uses all the addressing, correction and recovery functions X. 25 - but only between end nodes that implement Annex G. A persistent connection over a Frame Relay network in this case looks like a "straight wire" that carries X.25 traffic. The X.25 parameters (packet and window sizes) can be selected to achieve the lowest possible propagation delays and speed loss when encapsulating LAN protocols. The lack of error correction and sophisticated packet switching mechanisms typical for X.25 allows information to be transmitted over Frame Relay from minimal delays... Additionally, it is possible to enable a prioritization mechanism that allows the user to have a guaranteed minimum information transfer rate for a virtual channel. This capability allows Frame Relay to be used to carry delay-critical information such as real-time voice and video. This relatively new feature is gaining popularity and is often the primary consideration in choosing Frame Relay as the backbone of the corporate network. It should be remembered that today Frame Relay services are available in our country in no more than fifteen cities, while X.25 is available in about two hundred. There is every reason to believe that as communication channels evolve, Frame Relay technology will become more widespread - primarily where X.25 networks exist today. Unfortunately, there is no single standard describing how different Frame Relay networks interact, so users are bound to a single service provider. If it is necessary to expand the geography, it is possible to connect at one point to the networks of different suppliers - with a corresponding increase in costs. There are also private Frame Relay networks that operate within a single city or use long-distance - usually satellite - leased lines. Building private networks based on Frame Relay allows you to reduce the number of leased lines and integrate voice and data transmission.
The structure of the corporate network. Hardware.
When building a geographically distributed network, all the technologies described above can be used. To connect remote users, the simplest and most affordable option is to use a telephone connection. ISDN networks can be used where possible. In most cases, global data transmission networks are used to connect network nodes. Even where it is possible to lay leased lines (for example, within the same city), the use of packet switching technologies allows to reduce the number of necessary communication channels and - which is important - to ensure the compatibility of the system with existing global networks. Connecting a corporate network to the Internet is justified if you need access to the appropriate services. It is worth using the Internet as a data transmission medium only when other methods are unavailable and financial considerations outweigh the requirements of reliability and security. If you will only be using the Internet as a source of information, it is best to use dial-on-demand technology. in such a way of connection, when the connection to the Internet site is established only on your initiative and at the time you need. This dramatically reduces the risk of unauthorized access to your network from the outside. The easiest way to provide such a connection is to use dial-up to the Internet site over a telephone line or, if possible, over ISDN. Another, more reliable way to provide a connection on demand is to use a leased line and X.25 protocol, or - much more preferable - Frame Relay. In this case, the router on your end must be configured to disconnect the virtual connection if there is no data for a certain time and re-establish it only when data appears on your end. Widespread connection methods using PPP or HDLC do not provide this opportunity. If you want to provide your information on the Internet - for example, to set up a WWW or FTP server, the connection on demand turns out to be inapplicable. In this case, you should not only use access restriction using Firewall, but also isolate the Internet server from other resources as much as possible. A good solution is to use a single Internet connection point for the entire wide area network, the nodes of which are connected to each other using X.25 or Frame Relay virtual circuits. In this case, access from the Internet is possible to a single site, while users at other sites can access the Internet using a connection on demand.
To transfer data within the corporate network, it is also worth using virtual circuits of packet-switched networks. The main advantages of this approach - versatility, flexibility, security - were discussed in detail above. Both X.25 and Frame Relay can be used as a virtual network when building a corporate information system. The choice between them is determined by the quality of communication channels, the availability of services at the connection points and - not least - financial considerations. Today, the cost of using Frame Relay for long distance communications is several times higher than for X.25 networks. On the other hand, higher data rates and the ability to simultaneously transmit data and voice can be decisive arguments in favor of Frame Relay. On those parts of the corporate network where leased lines are available, Frame Relay technology is more preferable. In this case, it is possible both to interconnect local networks and connect to the Internet, as well as to use those applications that traditionally require X.25. In addition, telephone communication between the nodes is possible over the same network. For Frame Relay, it is better to use digital communication channels, however, even on physical lines or voice frequency channels, you can create a completely effective network by installing the appropriate channel equipment. Good results are obtained by using Motorola 326x SDC modems, which have unique capabilities for data correction and compression in synchronous mode. Thanks to this, it is possible - at the cost of introducing small delays - to significantly increase the quality of the communication channel and achieve an effective speed of up to 80 kbps and higher. Short-range modems can also be used on physical lines of short length, providing sufficiently high speeds. However, high quality of the line is required here, since short-range modems do not support any error correction. RAD short-range modems are widely known, as well as PairGain equipment, which allows achieving a speed of 2 Mbit / s on physical lines about 10 km long. To connect remote users to the corporate network, access nodes of X.25 networks can be used, as well as their own communication nodes. In the latter case, the allocation of the required number of telephone numbers (or ISDN channels) is required, which can be too expensive. If you need to connect a large number of users at the same time, then a cheaper option may be to use the access nodes of the X network. 25, even within the same city.
A corporate network is a rather complex structure that uses various types of communication, communication protocols and methods of connecting resources. From the point of view of the convenience of building and managing the network, one should focus on the same type of equipment from one manufacturer. However, practice shows that there are no suppliers offering the most effective solutions for all emerging problems. A working network is always the result of a compromise - either it is a homogeneous system, not optimal in terms of price and capabilities, or a combination of products from different manufacturers that is more difficult to install and manage. Next, we will look at networking tools from several leading manufacturers and give some recommendations for their use.
All equipment of data transmission networks can be conditionally divided into two large classes -
1.peripheral, which is used to connect end nodes to the network, and
2. backbone or backbone, which implements the basic functions of the network (circuit switching, routing, etc.).
There is no clear boundary between these types - the same devices can be used in different capacities or combine both functions. It should be noted that high requirements are usually imposed on backbone equipment in terms of reliability, performance, number of ports, and further expandability.
Peripheral equipment is a necessary component of any corporate network. The functions of backbone nodes can be assumed by the global data transmission network, to which resources are connected. As a rule, backbone nodes as part of a corporate network appear only in those cases when leased communication channels are used or their own access nodes are created. Peripheral equipment of corporate networks in terms of their functions can also be divided into two classes.
First, these are routers that serve to connect homogeneous LANs (usually IP or IPX) over wide area data networks. In networks using IP or IPX as the main protocol - in particular, in the same Internet - routers are also used as backbone equipment that provides the docking of various communication channels and protocols. Routers can be made both as stand-alone devices and software based on computers and special communication adapters.
The second widely used type of peripheral equipment is gateways), which implement the interaction of applications operating in different types of networks. Enterprise networks primarily use OSI gateways, which provide LAN interoperability with X.25 resources, and SNA gateways, which provide connectivity to IBM networks. A fully functional gateway is always a hardware and software complex, since it must provide the software interfaces necessary for applications. Cisco Systems routers Among the routers, the most well-known products are Cisco Systems, which implement a wide range of tools and protocols used in the interaction of local area networks. Cisco equipment supports a variety of connection methods, including X.25, Frame Relay and ISDN, allowing you to create fairly complex systems. In addition, among the Cisco family of routers, there are excellent servers for remote access to local networks, and in some configurations, gateway functions are partially implemented (what in Cisco terms is called Protocol Translation).
The main area of application for Cisco routers is complex networks using IP or, less often, IPX as the main protocol. In particular, Cisco equipment is widely used in Internet backbones. If your corporate network is primarily intended for interconnecting remote LANs and requires complex IP or IPX routing over heterogeneous communication channels and data networks, then using Cisco equipment is most likely the best choice. The means of working with Frame Relay and X.25 are implemented in Cisco routers only to the extent that is needed to connect local networks and access them. If you want to build your system on the basis of packet-switched networks, then Cisco routers can work in it only as purely peripheral equipment, and many of the routing functions are redundant, and the price, accordingly, is too high. The most interesting for use in corporate networks are the access servers Cisco 2509, Cisco 2511 and the new Cisco 2520 series devices. Their main area of application is access of remote users to local networks via telephone lines or ISDN with dynamic assignment of IP addresses (DHCP). Motorola ISG Equipment Among the equipment designed to work with X.25 and Frame Relay, the most interesting products are produced by the Motorola Information Systems Group (Motorola ISG). Unlike backbone devices used in global data transmission networks (Northern Telecom, Sprint, Alcatel, etc.), Motorola equipment is capable of operating completely autonomously, without a special network control center. The range of capabilities important for use in corporate networks is much wider for Motorola equipment. Of particular note is the advanced hardware and software upgrades that make it easy to adapt the equipment to specific conditions. All Motorola ISG products can operate as X.25 / Frame Relay switches, multi-protocol access devices (PAD, FRAD, SLIP, PPP, etc.), support Annex G (X.25 over Frame Relay), provide SNA protocol conversion (SDLC / QLLC / RFC1490). Motorola ISG equipment can be divided into three groups, differing in the set of hardware and scope.
The first group designed to work as peripheral devices is the Vanguard series. It includes Vanguard 100 (2-3 ports) and Vanguard 200 (6 ports) serial access nodes, as well as Vanguard 300/305 routers (1-3 serial ports and Ethetrnet / Token Ring port) and Vanguard 310 ISDN routers. Vanguard, in addition to a set of communication capabilities, includes the transfer of IP, IPX and Appletalk protocols over X.25, Frame Relay and PPP. Naturally, at the same time, the gentleman's set necessary for any modern router is supported - RIP and OSPF protocols, means of filtering and restricting access, data compression, etc.
The next group of Motorola ISG products includes the Multimedia Peripheral Router (MPRouter) 6520 and 6560, which differ mainly in performance and expandability. In the basic configuration, the 6520 and 6560 have, respectively, five and three serial ports and an Ethernet port, and the 6560 has all high-speed ports (up to 2 Mbps), and the 6520 has three ports up to 80 kbps. MPRouter supports all communication protocols and routing capabilities available for Motorola ISG products. The main feature of MPRouter is the ability to install a variety of additional cards, which is reflected in the word Multimedia in its name. There are serial port cards, Ethernet / Token Ring ports, ISDN cards, Ethernet hubs. The most interesting feature of MPRouter is Voice over Frame Relay. To do this, special cards are installed in it, allowing the connection of ordinary telephone or fax machines, as well as analog (E&M) and digital (E1, T1) automatic telephone exchanges. The number of simultaneously served voice channels can reach two or more dozen. Thus, MPRouter can be used simultaneously as a voice and data integration tool, a router, and an X.25 / Frame Relay node.
The third group of Motorola ISG products is the backbone equipment of global networks. The 6500plus family is expandable, fault-tolerant and redundant to create powerful switching and access nodes. They include various sets of processor modules and I / O modules, allowing for high-performance nodes with 6 to 54 ports. In corporate networks, such devices can be used to build complex systems with a large number of connected resources.
It is interesting to compare Cisco and Motorola routers. We can say that routing is primary for Cisco, and communication protocols are only a means of communication, while Motorola focuses on communication capabilities, considering routing as another service implemented using these capabilities. In general, the routing facilities of Motorola products are poorer than those of Cisco, but sufficient for connecting end nodes to the Internet or corporate network.
Other things being equal, the performance of Motorola products is perhaps even higher, and at a lower price. So the Vanguard 300, with a comparable set of features, turns out to be about one and a half times cheaper than its closest analogue, the Cisco 2501.
Eicon Technology Solutions
In many cases, it is convenient to use the solutions of the Canadian company Eicon Technology as the peripheral equipment of corporate networks. Eicon solutions are based on a universal communication adapter EiconCard that supports a wide range of protocols - X.25, Frame Relay, SDLC, HDLC, PPP, ISDN. This adapter is installed in one of the computers on the local network, which becomes a communication server. This computer can be used for other tasks as well. This is possible due to the fact that EiconCard has a sufficiently powerful processor and its own memory and is able to implement the processing of network protocols without loading the communication server. Eicon software tools allow building both gateways and routers on the basis of EiconCard, work under control of almost all operating systems on the Intel platform. Here we will look at the most interesting ones.
The Eicon family of Unix solutions includes an IP Connect router, X.25 Connect gateways, and SNA Connect. All of these products can be installed on a computer running SCO Unix or Unixware. IP Connect allows IP traffic to be carried over X.25, Frame Relay, PPP or HDLC and is compatible with equipment from other manufacturers such as Cisco and Motorola. The package includes Firewall, data compression and SNMP management tools. The main area of application for IP Connect is connecting application servers and Unix-based Internet servers to the data network. Naturally, the same computer can be used as a router for the entire office in which it is installed. There are several advantages to using an Eicon router instead of "pure hardware" devices. First, it is easy to install and use. From the point of view of the operating system, the EiconCard with IP Connect installed looks like another network card. This makes setting up and administering IP Connect enough a simple matter for anyone who has talked with Unix. Secondly, direct connection of the server to the data transmission network allows to reduce the load on the office LAN and to provide that very single point of connection to the Internet or to the corporate network without installing additional network cards and routers. Third, this "server-centric" solution is more flexible and scalable than traditional routers. There are a number of other benefits that come with combining IP Connect with other Eicon products.
X.25 Connect is a gateway that allows LAN applications to communicate with X.25 resources. This product allows Unix users and DOS / Windows and OS / 2 workstations to connect to remote e-mail systems, databases and other systems. By the way, it should be noted that Eicon gateways today are perhaps the only product widespread in our market that implements the OSI stack and allows you to connect to X.400 and FTAM applications. In addition, X.25 Connect allows you to connect remote users to a Unix machine and terminal applications on stations on a local network, as well as to organize the interaction of remote Unix computers via X.25. Used in conjunction with X.25 Connect standard features Unix, it is possible to implement protocol conversion, i.e. translate Unix access via Telnet into an X.25 call and vice versa. It is possible to connect a remote X.25 user using SLIP or PPP to the local network and, accordingly, to the Internet. In principle, similar protocol translation capabilities are available in Cisco routers running IOS Enterprise software, but this solution is more expensive than Eicon and Unix combined.
Another product mentioned above is SNA Connect. This is the gateway for connecting to IBM mainframe and AS / 400. It is typically used in conjunction with user software - 5250 and 3270 terminal emulators and APPC interfaces - also manufactured by Eicon. Analogs of the solutions discussed above exist for other operating systems - Netware, OS / 2, Windows NT, and even DOS. Of particular note is Interconnect Server for Netware, which combines all of the above capabilities with remote configuration and administration tools and a client authorization system. It includes two products - Interconnect Router, which enables IP, IPX and Appletalk routing, which we believe is the best solution for interconnecting remote Novell Netware networks, and Interconnect Gateway, which provides, in particular, powerful SNA connectivity. Another Eicon product designed to work in the Novell Netware environment is WAN Services for Netware. It is a set of tools that enable Netware applications to be used over X.25 and ISDN networks. Using it in conjunction with Netware Connect allows remote users to connect to the local network via X.25 or ISDN, and also provide an exit from the local network in X.25. There is an option for delivering WAN Services for Netware along with Novell's Multiprotocol Router 3.0. This product is called Packet Blaster Advantage. Packet Blaster ISDN is also available, which does not work with EiconCard, but with ISDN adapters also supplied by Eicon. In this case, various connection options are possible - BRI (2B + D), 4BRI (8B + D) and PRI (30B + D). The WAN Services for NT product is designed to work with Windows NT applications. It includes an IP Router, a means of connecting NT applications to X.25 networks, support for Microsoft SNA Server, and a means of accessing remote users over X.25 to a LAN using Remote Access Server. To connect a Windows NT server to an ISDN network, the Eicon ISDN adapter can also be used in conjunction with the ISDN Services for Netware software.
Methodology for building corporate networks.
Now that we have listed and compared the main technologies that a developer can employ, let's move on to the basic questions and techniques used in designing and developing a network.
Network requirements.
Network designers and network administrators always strive to ensure that three basic network requirements are met, namely:
scalability;
performance;
controllability.
Good scalability is required so that both the number of users on the network and the application software can be changed effortlessly. High network performance is required for most modern applications to function properly. Finally, the network must be manageable enough that it can be reconfigured to meet the ever-changing needs of the organization. These requirements reflect a new stage in the development of network technologies - the stage of creating high-performance corporate networks.
The uniqueness of the new software tools and technology complicates the design of corporate networks. Centralized resources, new classes of programs, other principles of their application, changes in the quantitative and qualitative characteristics of the information flow, an increase in the number of concurrent users and an increase in the power of computing platforms - all these factors must be taken into account in their totality when developing a network. Now there are a large number of technological and architectural solutions on the market, and choosing the most suitable one is a rather difficult task.
In modern conditions, for the correct design of the network, its development and maintenance, specialists must take into account the following issues:
o Change in organizational structure.
When implementing a project, you should not "separate" software specialists and network specialists. When developing networks and the entire system as a whole, a single team of specialists of different profiles is needed;
o Use of new software tools.
It is necessary to get acquainted with the new software at an early stage of network development in order to be able to timely make the necessary adjustments to the tools planned for use;
o Investigation of various solutions.
It is necessary to evaluate various architectural solutions and their possible impact on the operation of the future network;
o Checking networks.
Test the entire network or parts of it early in development. To do this, you can create a prototype of the network, which will allow you to evaluate the correctness of the decisions made. This way you can prevent the appearance of various kinds of "bottlenecks" and determine the applicability and approximate performance of different architectures;
o Choice of protocols.
In order to choose the right network configuration, you need to evaluate the capabilities of the various protocols. It is important to determine how network operations that optimize the performance of one program or suite of programs can affect the performance of others;
o Choice of physical location.
When choosing a place to install servers, you must first of all determine the location of users. Is it possible to move them? Will their computers be connected to the same subnet? Will users have access to the global network?
o Calculation of critical time.
It is necessary to determine the acceptable response times of each application and the possible periods of maximum load. It is important to understand how abnormal situations can affect the health of the network, and determine whether a reserve is needed to organize the continuous operation of the enterprise;
o Analysis of options.
It is important to analyze the different uses of the software on the web. Centralized storage and processing of information often creates additional load at the center of the network, and distributed computing may require strengthening of local networks of workgroups.
Today there is no ready-made, debugged universal methodology, following which, you can automatically carry out the entire range of measures for the development and creation of a corporate network. This is primarily due to the fact that there are no two absolutely identical organizations. In particular, each organization is characterized by a unique leadership style, hierarchy, and business culture. And if we consider that the network inevitably reflects the structure of the organization, then we can safely say that no two identical networks exist.
Network architecture
Before starting to build a corporate network, you must first define its architecture, functional and logical organization and take into account the existing telecommunications infrastructure. A well-designed network architecture helps assess the feasibility of new technologies and applications, serves as a foundation for future growth, determines the choice of network technologies, helps to avoid excessive costs, reflects the connectivity of network components, significantly reduces the risk of incorrect implementation, etc. The network architecture forms the basis of the technical specifications for the network being created. It should be noted that a network architecture differs from a network design in that it, for example, does not define the exact schematic diagram of the network and does not regulate the placement of network components. The network architecture, for example, determines whether some parts of the network will be based on Frame Relay, ATM, ISDN, or other technologies. The network design should contain specific guidelines and estimates of parameters, for example, the required bandwidth, the actual bandwidth, the exact location of the communication channels, etc.
There are three aspects in the network architecture, three logical components:
principles of construction,
network templates
and technical positions.
Design principles are used in network planning and decision making. Principles are a set of simple instructions that describe, in sufficient detail, all aspects of building and operating a deployed network over an extended period of time. As a rule, the formation of principles is based on the corporate goals and basic methods of doing business of the organization.
The Principles provide the primary link between corporate strategy and networking. They serve to develop technical positions and network templates. When developing a technical assignment for a network, the principles of building a network architecture are set out in the section that defines the general goals of the network. The technical position can be viewed as a target description that determines the choice between competing alternative network technologies. The technical position specifies the parameters of the selected technology and gives a description of a single device, method, protocol, service provided, etc. For example, when choosing a LAN technology, you need to consider speed, cost, quality of service, and other requirements. Developing technical positions requires in-depth knowledge of networking technology and careful consideration of the organization's requirements. The number of technical positions is determined by the given granularity, network complexity, and organization scale. The network architecture can be described by the following technical items:
Network transport protocols.
What transport protocols should be used to transfer information?
Network routing.
What routing protocol should be used between routers and ATM switches?
Quality of service.
How will the ability to choose the quality of service be achieved?
Addressing in IP networks and addressing domains.
What addressing scheme should be used for the network, including registered addresses, subnets, subnet masks, forwarding, etc.?
Switching in local networks.
What switching strategy should be used in LANs?
Combining switching and routing.
Where and how switching and routing should be used; how should they be combined?
Organization of a city network.
How should branches of an enterprise located, say, in the same city be connected?
Organization of a global network.
How should the branches of the enterprise communicate over the global network?
Remote access service.
How do remote branch users access the enterprise network?
Network templates are a collection of network structure models that reflect the relationship between network components. For example, for a specific network architecture, a set of templates are created to "show" the network topology of a large branch or WAN, or to show the layering of protocols. Network templates illustrate a network infrastructure that is described by a complete set of technical items. Moreover, in a well-thought-out network architecture, network patterns in terms of granularity can be as close as possible in their content to technical positions. In fact, network templates are a description of the functional diagram of a network section that has specific boundaries; the following basic network templates can be distinguished: for a global network, for a city network, for a central office, for a large branch of an organization, for a branch. Other templates can be developed for network sections that have any particularities.
The described methodological approach is based on studying a specific situation, considering the principles of building a corporate network in their totality, analyzing its functional and logical structure, developing a set of network templates and technical positions. Different implementations of corporate networks can include certain components. In general, a corporate network consists of various departments connected by communication networks. They can be global (WAN) or metropolitan (MAN). Branches can be large, medium and small. A large department can be a center for processing and storing information. A central office is allocated from which the entire corporation is managed. Various service divisions (warehouses, workshops, etc.) can be classified as small branches. Small offices are essentially remote. The strategic purpose of the remote branch is to host sales and technical support closer to the consumer. Communication with customers, which has a significant impact on corporate earnings, will be more productive if all employees can have access to corporate data at any time.
At the first step of building a corporate network, the proposed functional structure is described. The quantitative composition and status of offices and branches is determined. The necessity of deploying your own private communication network is substantiated or the choice of a service provider that is able to meet the requirements is made. The development of the functional structure is carried out taking into account the financial capabilities of the organization, long-term development plans, the number of active network users, running applications, the required quality of service. The development is based on the functional structure of the enterprise itself.
In the second step, the logical structure of the corporate network is determined. Logical structures differ from each other only in the choice of technology (ATM, Frame Relay, Ethernet ...) for building the backbone, which is the central link of the corporation's network. Consider logical structures based on cell switching and frame switching. The choice between these two methods of information transmission is based on the need to provide a guaranteed quality of service. Other criteria can be used as well.
The data transmission backbone must satisfy two basic requirements.
o Ability to connect a large number of low-speed workstations to a small number of powerful, high-speed servers.
o Acceptable speed of response to customer requests.
An ideal backbone should have high reliability of data transmission and an advanced control system. A control system means, for example, the ability to configure the backbone taking into account all local conditions and maintain reliability at such a level that, even if some parts of the network fail, the servers remain available. The listed requirements will probably determine several technologies and the final choice of one of them remains with the organization itself. What matters most is cost, speed, scalability, or quality of service.
The logical structure with cell switching is used in networks with multimedia traffic in real time (video conferencing and high-quality voice transmission). At the same time, it is important to soberly assess how necessary such an expensive network is (on the other hand, even expensive networks are sometimes unable to meet some requirements). If so, then the logical structure of the frame-switched network should be taken as a basis. The logical switching hierarchy uniting two layers of the OSI model can be represented as a three-tier diagram:
The lower layer is used to combine local Ethernet networks,
The middle layer is either a local ATM network, a MAN network, or a WAN backbone.
The top level of this hierarchical structure is responsible for routing.
The logical structure allows you to identify all possible communication routes between individual sections of the corporate network
Cell switching backbone
When using cell switching technology to build a network backbone, the combination of all switches Ethernet layer workgroups are carried out by high-performance ATM switches. Operating at the second layer of the OSI reference model, these switches transmit fixed-length 53-byte cells instead of variable-length Ethernet frames. This networking concept assumes that an Ethernet switch working group Must have an ATM Segmentation and Assembly (SAR) egress port that converts variable-length Ethernet frames to fixed-length ATM cells before forwarding the information to the ATM backbone switch.
For wide area networks, ATM core switches are capable of providing connectivity to remote regions. Also operating at layer 2 of the OSI model, these switches on the WAN can use T1 / E1 (1.544 / 2.0Mbps), T3 (45Mbps), or SONET OC-3 (155Mbps). To provide urban connectivity, a MAN network can be deployed using ATM technology. The same ATM backbone network can be used to communicate between telephone exchanges. In the future, within the framework of the client / server telephony model, these stations can be replaced by voice servers on the local network. In this case, the ability to guarantee the quality of service in ATM networks becomes very important when organizing communication with client personal computers.
Routing
As already noted, routing is the third and most high level in the hierarchical structure of the network. Routing, which operates at the third layer of the OSI reference model, is used to establish communication sessions, which include:
o Communication sessions between devices located in different virtual networks (each network is usually a separate IP subnet);
o Communication sessions that go through the global / city
One strategy for building a corporate network is to install switches at the lower layers of the overall network. The local area networks are then connected using routers. Routers are required to divide a large organization's IP network into many distinct IP subnets. This is to prevent the "broadcast explosion" associated with the operation of protocols such as ARP. To keep unwanted traffic from spreading over the network, all workstations and servers must be split into virtual networks. In this case, routing controls communication between devices belonging to different VLANs.
Such a network consists of routers or routing servers (logical core), a network backbone based on ATM switches, and a large number of Ethernet switches located at the edge. Except in special cases, such as using video servers that connect directly to the ATM backbone, all workstations and servers must be connected to Ethernet switches. This network design will help localize internal traffic within workgroups and prevent such traffic from being pumped through ATM backbone switches or routers. Ethernet switch bundling is done by ATM switches, usually located in the same branch. It should be noted that multiple ATM switches may be required to provide enough ports to connect all Ethernet switches. Typically, in this case, a 155Mbps connection is used over multimode fiber optic cable.
The routers are located away from the ATM backbone switches because these routers need to be moved out of the main communication paths. This design makes routing optional. It depends on the type of communication session and the type of traffic on the network. Routing should be avoided when transmitting live video information as it can introduce unwanted delays. Routing is not necessary for communication between devices located in the same virtual network, even if they are located in different buildings on the territory of a large enterprise.
In addition, even in a situation where routers are required for certain communication sessions, placing routers away from ATM backbone switches minimizes the number of routing hops (routing hops refers to the section of the network from the user to the first router or from one router to another). This not only reduces latency, but also reduces the load on the routers. Routing has become widespread as a local area network communication technology in a global environment. Routers provide a variety of services designed for multi-level control of the transmission channel. This includes the general addressing scheme (at the network layer), independent of how the addresses of the previous layer are formed, as well as the conversion from one frame format of the control layer to another.
Routers make decisions about where to route incoming data packets based on the information they contain about the network layer addresses. This information is retrieved, analyzed, and compared with the contents of the routing tables to determine which port a packet should be sent to. Then, the link-layer address is extracted from the network-layer address if the packet is to be sent to a segment of a network such as Ethernet or Token Ring.
In addition to processing packets, routers update the routing tables in parallel, which are used to determine the destination of each packet. Routers dynamically create and maintain these tables. As a result, routers can automatically respond to changes in network conditions, such as congestion or damage to communication channels.
Determining the route is a rather difficult task. In a corporate network, ATM switches should function in much the same way as routers: communication should take place taking into account the network topology, available routes, and transmission costs. The ATM switch needs this information to select the best route for a particular end-user-initiated communication session. In addition, the definition of a route is not limited to just making a decision about the path that a logical connection will follow after a request for its creation is generated.
The ATM switch can choose new routes if for some reason the communication channels become unavailable. At the same time, ATM switches must provide network reliability at the router level. To create a scalable network with high cost efficiency, it is necessary to move routing functions to the edge of the network and provide traffic switching on its backbone. ATM is the only networking technology that can do this.
To select a technology, you must answer the following questions:
Does the technology provide an adequate quality of service?
Can she guarantee the quality of service?
How expandable will the network be?
Is there a choice of network topology?
Are the services provided by the network cost-effective?
How effective will the management system be?
The answers to these questions determine the choice. But, in principle, in different parts of the network, different technologies... For example, if certain sections require real-time multimedia traffic support or 45 Mbps speed, then ATM is installed in them. If a section of the network requires interactive processing of requests, which does not allow significant delays, then Frame Relay should be used if such services are available in this geographic area (otherwise, you will have to resort to the Internet).
Thus, a large enterprise can connect to the network via ATM, while branch offices connect to the same network via Frame Relay.
When creating a corporate network and choosing a network technology with the appropriate software and hardware, the price / performance ratio should be considered. It is difficult to expect high speeds from cheap technologies. On the other hand, it makes no sense to use the most complex technologies for the simplest tasks. Different technologies must be properly combined to achieve maximum efficiency.
When choosing a technology, one should take into account the type of cable system and the required distances; compatibility with already installed equipment (significant minimization of costs can be achieved if in new system it is possible to turn on the already installed equipment.
Generally speaking, there are two ways to build a high-speed local area network: evolutionary and revolutionary.
The first way is based on the extension of the good old frame relay technology. It is possible to increase the performance of a local network within the framework of this approach by upgrading the network infrastructure, adding new communication channels and changing the method of packet transmission (which is what was done in switched Ethernet). A conventional Ethernet network shares bandwidth, which means that the traffic of all users on the network competes with each other, claiming the entire bandwidth of the network segment. In switched Ethernet, dedicated routes are created, so that real bandwidth of 10 Mbps is available to users.
The revolutionary path involves the transition to radically new technologies, for example, ATM for local networks.
The rich practice of building local networks has shown that the main issue is the quality of service. This is what determines whether the network can work successfully (for example, with applications such as video conferencing, which are increasingly used in the world).
Conclusion.
To have or not to have its own communication network is a “private matter” of each organization. However, if building a corporate (departmental) network is on the agenda, it is necessary to conduct a deep, comprehensive study of the organization itself, the tasks it solves, draw up a clear document flow diagram in this organization and, on this basis, start choosing the most acceptable technology. One of the examples of building corporate networks is the currently widely known Galaxy system.
List of used literature:
1. M. Shestakov "Principles of building corporate data transmission networks" - "Computerra", No. 256, 1997
2. Kosarev, Eremin "Computer systems and networks", Finance and statistics, 1999
3. Olifer V. G., Olifer N. D. "Computer networks: principles, technologies, protocols", Peter, 1999
4. Materials of the site rusdoc.df.ru