Fast network. Fast Ethernet technology. DSAP and SSAP fields

Ethernet, but also to the equipment of other, less popular networks.

Ethernet and Fast Ethernet Adapters

Characteristics of adapters

Network Adapters (NIC, Network Interface Card) Ethernet and Fast Ethernet can conjugate with a computer through one of standard interfaces:

  • ISA tire (Industry Standard Architecture);
  • pCI bus (Peripheral Component InterConnect);
  • tire PC Card (it is PCMCIA);

Adapters designed for the ISA system bus (highway), not so long ago were the main type of adapters. The number of companies producing such adapters was large, which is why the devices of this type were the cheapest. Adapters for ISA are manufactured 8- and 16-bit. 8-bit adapters are cheaper, and 16-bit - faster. True, the exchange of information on the ISA bus cannot be too fast (in the limit - 16 MB / s, real - not more than 8 MB / s, and for 8-bit adapters - up to 2 MB / s). Therefore, the Fast Ethernet adapters requiring the efficient operation of high exchange rates for this system tire is practically not available. The ISA tire goes into the past.

The PCI bus is now practically pushed out the ISA bus and becomes the main extension bus for computers. It provides an exchange of 32- and 64-bit data and has a high bandwidth (theoretically up to 264 MB / s), which quite satisfies the requirements not only Fast Ethernet, but also more quickly Gigabit Ethernet. The fact that the PCI bus is applied not only in IBM PC computers, but also in PowerMac computers. In addition, it supports the automatic configuration of the Plug-and-Play equipment. Apparently, in the near future the majority will be focused on the PCI bus network adapters. The lack of PCI compared to the ISA bus is that the amount of its expansion slots in the computer is usually small (usually 3 slots). But just network adapters Connected to PCI first.

The PC Card Tire (the old PCMCIA name) is used so far only in the portable computers of the Notebook class. In these computers, the internal tire of PCI is usually not displayed out. The PC Card interface provides a simple connection to the computer miniature extension cards, and the exchange rate with these boards is sufficiently high. However, more and more laptop computers are equipped with built-in network adaptersSince the ability to access the network becomes an integral part of the standard set of functions. These built-in adapters are again connected to the inside bus PCI computer.

When choosing network adapter An oriented on a bus, first of all, make sure that the free slots of the extension of this tire is in the computer, including the network. It should also be estimated to be the complexity of the installation of the acquired adapter and the outlook the output of the board of this type. The latter may be needed in the event of an adapter output.

Finally, there are still network adaptersconnected to a computer via parallel (printer) port LPT. . The main advantage of this approach is that it does not need to open the computer case for connecting adapters. In addition, in this case, the adapters do not occupy computer resources, such as interrupt channels and PDP, as well as memory addresses and I / O devices. However, the speed of exchanging information between them and the computer in this case is significantly lower than when using the system tire. In addition, they require more processor time to exchange with the network, thereby slowing the work of the computer.

IN lately more and more computers in which network adapters Embed into a system fee. The advantages of this approach are obvious: the user should not buy a network adapter and install it in the computer. It is enough just to connect the network cable to the external connector of the computer. However, the disadvantage is that the user cannot choose the adapter with the best characteristics.

To other major characteristics network adapters You can attribute:

  • method of configuring the adapter;
  • size installed on the board buffer memory and exchange modes with it;
  • the ability to install on the permanent memory chipboard for remote download (BOOTROM).
  • the ability to connect an adapter to different types of transmission media ( twisted para, thin and thick coaxial cable, fiber optic cable);
  • used by the adapter transmission speed over the network and the presence of the switching function;
  • the possibility of applying a full-duplex exchange mode adapter;
  • adapter compatibility (more precisely, adapter driver) with network software used.

Configuring the adapter by the user was used mainly for the adapters designed for the ISA bus. Configuration implies configuration to the use of computer system resources (I / O addresses, interrupt channels and direct memory access, buffer memory addresses and remote download memory). Configuration can be carried out by installing in the desired position of the switches (jumpers) or using the Configuration DOS program attached to the adapter (Jumperless, Software Configuration). When you start such a program, the user is invited to establish the configuration of the equipment using simple menu: Select the adapter parameters. The same program allows you to produce self-test adapter. The selected parameters are stored in the non-volatile memory of the adapter. In any case, when selecting parameters, it is necessary to avoid conflicts with system devices Computer and with other extension boards.

Configuring the adapter can be performed and automatically in Plug-and-Play mode when the computer is powering on. Modern adapters usually support exactly this mode, so the user can easily install them.

In the simple adapters, the exchange with the internal buffer memory of the adapter (Adapter RAM) is carried out through the address space of I / O devices. In this case, no additional configuration of memory addresses is required. The basic address of the buffer memory operating in the memory mode must be set. It is attributed to the top of the computer's top memory (

In the ComputerPress test laboratory, testing was tested for use in 10/100 Mbps workstations with Fast Ethernet network cards for the PCI bus. The most common cards currently with a 10/100 MBT / S bandwidth were selected, since, first, they can be used in Ethernet networks, Fast Ethernet and in mixed networks, and, secondly, the promising Gigabit Ethernet technology ( Bandwidth up to 1000 Mbps) is still used to be most often used to connect powerful servers to network core network equipment. It is extremely important what quality passive network equipment (cables, sockets, etc.) is used on the network. It is well known that if for Ethernet networks there is enough cable on twisted pair of category 3, then 5 category is needed for Fast Ethernet. Scattering a signal, poor noise security can significantly reduce throughput network.

The purpose of testing was to define primarily the effective performance index (Performance / Efficiency Index Ratio in the future P / E-index), and only then - the absolute value of the bandwidth. The P / E-index is calculated as the ratio of the bandwidth of the network card in Mbit / c to the degree of loading of the central processor in percent. This index is a sectoral standard for determining the performance of network adapters. It was introduced in order to take into account the use of network card resources of the central processor. The fact is that some manufacturers of network adapters are trying to achieve maximum performance by using for network operations of a larger number of computer processor cycles. The minimum processor load and relatively high throughput are of great importance for the execution of critical business and multimedia applications, as well as real-time tasks.

Cards were tested, which are currently more often used for workstations in corporate and local networks:

  1. D-Link DFE-538TX
  2. SMC EtherPower II 10/100 9432TX / MP
  3. 3Com Fast Etherlink XL 3C905B-TX-NM
  4. Compex RL 100atx
  5. INTEL ETHEREXPRESS PRO / 100 + MANAGEMENT
  6. CNET PRO-120
  7. NetGear Fa 310tx
  8. Allied TeleSyn AT 2500TX
  9. Surecom EP-320X-R

The main characteristics of the tested network adapters are shown in Table. one . Let us explain some terms that are used in the table. The automatic determination of the connection speed means that the adapter itself determines the maximum possible operation speed. In addition, in the case of supporting the velocity auto definition, no additional configuration during the transition from Ethernet to Fast Ethernet is not necessary. That is, the system administrator does not require reconfigure the adapter and overload drivers.

The support of the BUS Master mode allows you to transmit data directly between the network card and the computer memory. Thus, the central processor is released to perform other operations. This property has become the standard de facto. No wonder all known network cards support Bus Master mode.

Remote inclusion (Wake On LAN) allows you to turn on the PC on the network. That is, it is possible to serve PCs at no time. For this purpose, three-pin connectors are used on a system board and a network adapter that are connected by a special cable (included in the package). In addition, it is necessary to special control software. Wake On LAN technology is developed by the Intel-IBM Alliance.

The full duplex mode allows you to transmit the data at the same time in both directions, half duplex - only in one. Thus, the maximum possible bandwidth in full-duplex mode is 200 Mbps.

The DMI interface (Desktop Management Interface) makes it possible to receive information on configuration and PC resources using Network Management.

WFM Specification Support (Wired For Management) provides network adapter with network management and administration software.

To remotely download the OS computer over the network, network adapters are supplied special memory BOOTROM. This makes it possible to effectively use non-free workstations on the network. In most tested cards, only a socket for installing BOOTROM was present; The BOOTROM microcircuit itself is usually separately ordered by the option.

ACPI support (Advanced Configuration Power Interface) reduces power consumption. ACPI is a new technology that ensures the operation of the power management system. It is based on the use of both hardware and software. In principle, Wake on LAN is an integral part of ACPI.

Profitability raising means allow you to increase the efficiency of the network card. The most famous of them are Parallel Tasking II 3Com and Adaptive Technology company. These funds are usually patented.

Support for basic operating systems is provided by almost all adapters. The main OS includes: Windows, Windows NT, NetWare, Linux, SCO Unix, Lan Manager and others.

Service support level is rated by the availability of documentation, diskettes with drivers and the ability to download latest versions Drivers from the company's website. The packaging plays the last role. From this point of view, the best, in our opinion, are the D-Link network adapters, Allied Telesen and Surecom. But in general, the support level was satisfactory for all cards.

Usually, the warranty extends to the entire operation time of the network adapter (lifelong warranty). Sometimes it is limited to 1-3 years.

Testing technique

All tests used the latest versions of the network card drivers that were loaded from the Internet servers of the relevant manufacturers. In the case when the network card driver allowed any settings and optimization, the default settings were used (except for the Intel network adapter). Note that the richest additional features And the features have the functions and the corresponding 3com and Intel drivers.

Performance measurement was performed using the Novell Perform3 utility. The principle of the utility is that the small file is rewritten from the workstation to the shared network drive of the server, after which it remains in the server's file cache and for a specified period of time repeatedly reads from there. This allows you to achieve the interaction of the type of memory-network memory and eliminate the effect of delays associated with disk operations. The utility settings include the initial file size, the final file size, the size step and the test time. The Novell Perform3 utility displays performance values \u200b\u200bwith different size files, medium and maximum performance (in KB / C). The following parameters were used to configure the utility:

  • File initial size - 4095 bytes
  • Final File Size - 65 535 byte
  • File increment step - 8192 bytes

Test time with each file was set to twenty seconds.

In each experiment, a pair of identical network cards was used, one of which worked on the server, and the other on the workstation. It seems that this does not match common practice, since the servers usually use specialized network adapters equipped with a number of additional functions. But it is in this way that the same network cards are installed on the server and workstations - all the well-known test laboratories of the world are tested (Keylabs, Tolly Group, etc.). The results are obtained slightly lower, but the experiment turns out to be clean, since only the analyzed network cards work on all computers.

Compaq DeskPro En Client Configuration:

  • pentium II 450 MHz processor
  • cache 512 kb
  • rAM 128 MB
  • winchester 10 GB
  • operating system Microsoft Windows. NT Server 4.0 C 6 A SP
  • tCP / IP protocol.

Compaq Deskpro EP Server Configuration:

  • celeron 400 MHz processor
  • rAM 64 MB
  • winchester 4.3 GB
  • microsoft Windows NT Workstation 4.0 C 6 A SP
  • tCP / IP protocol.

Testing was carried out under conditions when computers were connected directly to the UTP Category crossover cable 5. During these card tests, the 100Base-TX Full Duplex mode was operated. In this mode, bandwidth is slightly higher due to the fact that part of the service information (for example, a reception confirmation) is transmitted simultaneously with useful information, the amount of which is estimated. Under these conditions, it was possible to fix quite high bandwidth values; For example, for the adapter 3COM FAST Etherlink XL 3C905B-TX-NM on average 79.23 Mbps.

The workload of the processor was measured on the server using windows utilities NT Performance Monitor; The data was recorded in the log file. The Perform3 utility launched on the client not to affect the server processor load. Intel Celeron was used as a computer-server processor, the performance of which is significantly lower than the performance of Pentium II and III processors. Intel Celeron was used intentionally: the fact is that because the loading of the processor is determined with a sufficiently large absolute error, in the case of large absolute values, the relative error turns out to be less.

After each test, the Perform3 utility places the results of its work in the text file as a set of data of the following type:

65535 bytes. 10491.49 Kbps. 10491.49 Aggregate Kbps. 57343 bytes. 10844.03 Kbps. 10844.03 Aggregate Kbps. 49151 bytes. 10737.95 Kbps. 10737.95 Aggregate Kbps. 40959 bytes. 10603.04 Kbps. 10603.04 Aggregate Kbps. 32767 bytes. 10497.73 Kbps. 10497.73 Aggregate Kbps. 24575 bytes. 10220.29 Kbps. 10220.29 Aggregate Kbps. 16383 bytes. 9573.00 Kbps. 9573.00 Aggregate Kbps. 8191 bytes. 8195.50 Kbps. 8195.50 Aggregate Kbps. 10844.03 Maximum Kbps. 10145.38 Average KBP.

The file size corresponding to the bandwidth for the selected client and for all clients is (in this case, the client is only one), as well as the maximum and average bandwidth throughout the test. The obtained average values \u200b\u200bfor each test were translated from KBIAT / C to MBIT / C by the formula:
(Krib x 8) / 1024,
and the P / E index value was calculated as the ratio of bandwidth to the processor load in percent. In the future, the average value of the P / E index was calculated according to the results of three dimensions.

Using the Perform3 utility on Windows NT Workstation, the following problem appeared: In addition to writing to a network drive, the file was also recorded in the local file cache, from where it was later read very quickly. The results were impressive, but unreal, since data transfer as such on the network was not performed. In order for the applications to perceive shared network drives as ordinary local disksThe operating system uses a special network component - redirector, redirecting I / O request over the network. Under normal working conditions, when executing a file recording procedure to a shared network disc, the redirector uses the Windows NT caching algorithm. That is why when writing to the server, an entry is also entry into the local file cache of the client machine. And for testing it is necessary that caching is carried out only on the server. In order for the caching client-client, there were no parameter values \u200b\u200bin the Windows NT registry, which made it possible to disable caching produced by the redirector. This is how it was done:

  1. Path in Registry:

    HKEY_LOCAL_MACHINE \\ SYSTEM \\ CURRENTCONTROLSET \\ SERVICES \\ RDR \\ PARAMETERS

    Parameter name:

    UseWriteBehind allows Write-Behind optimization for recorded files

    Type: REG_DWORD.

    Meaning: 0 (default: 1)

  2. Path in Registry:

    HKEY_LOCAL_MACHINE \\ SYSTEM \\ CURRENTCONTROLSET \\ SERVICES \\ LANMANWORKSTATION \\ PARAMETERS

    Parameter name:

    Utilizentcaching indicates whether the redirector will use the Windows NT cache manager to caching the contents of the files.

    Type: REG_DWORD Value: 0 (default: 1)

INTEL ETHEREXPRESS PRO / 100 + MANAGEMENT network adapter

The throughput of this card and the level of processing the processor turned out to be almost the same as in 3com. Below are the options for setting the parameters of this map.

The new Intel 82559 controller installed on this card provides very high performance, especially in Fast Ethernet networks.

The technology that Intel uses in its Intel EtherExpress Pro / 100 + card, is named Adaptive Technology. The essence of the method is to automatically change the time intervals between Ethernet packages, depending on the network load. With an increase in the network load increase, the distance between individual Ethernet packages is dynamically increasing, which reduces the number of collisions and increase bandwidth. With a small network load when the probability of mala collisions, temporary gaps between the packages are reduced, which also leads to an increase in performance. To the greatest degree of advantage of this method should be manifested in large consolidating Ethernet segments, that is, in cases where hubs predominate in the network topology, and not switches.

New intel technologyThe named Priority Packet allows you to adjust traffic passing through network card, in accordance with the priorities of individual packages. This makes it possible to raise the data transfer rate for critical applications.

Support is provided virtual local networks VLAN (IEEE 802.1Q standard standard.

On the board only two indicators - work / compound, speed 100.

www.intel.com.

Network adapter SMC Etherpower II 10/100 SMC9432TX / MP

In the architecture of this card used two perspective technologies SMC SimultAssking and Programmable Interpacket Gap. The first technology is similar to 3Com Parallel Tasking technology. Comparing testing results for the cards of these two manufacturers, it can be concluded about the degree of efficiency of the implementation of these technologies. Note also that this network card showed the third result and performance and P / E index, ahead of all the cards except 3com and Intel.

On the four map lED indicator: Speed \u200b\u200b100, Transmission, Connection, Duplex.

Address of the main Web site of the company: www.smc.com

Ethernet, despite
For his whole success, never was elegant.
Network boards have only rudimentary
The concept of intelligence. They really
first send the package, but only then
Look, whether the data passed anyone else
Simultaneously with them. Someone compared Ethernet with
society in which people can communicate
with each other only when everyone screams
At the same time.

Like Him
Predecessor, Fast Ethernet uses a method
CSMACD data transfer (Carrier Sense Multiple Access WITH
Collision Detection - multiple access to the environment with
Control the carrier and detection of collisions).
For this long and incomprehensible acronym
Hiding a very simple technology. When
Ethernet should send a message, then
At first she is waiting for the onset of silence, then
sends the package and simultaneously listening, not
Has anyone sent a message
Simultaneously with him. If this happened, then
Both packages do not reach the addressee. If a
the collisions were not, and the fee should continue
transmit data, it still waits
several microseconds before again
Try to send a new portion. it
Made for other fees as well
could work and no one could capture
Channel monopolis. In the case of collision, both
devices shut down on a small
time generated
randomly and then take
New attempt to transfer data.

Because of the collisions
Ethernet nor Fast Ethernet will never be able to achieve
its maximum performance 10
or 100 Mbps. As soon as starts
increase network traffic, temporary
Delays between sending individual packages
reduced, and the number of collisions
Increases. Real
Ethernet performance may not exceed
70% of its potential bandwidth
abilities and maybe even lower if the line
Seriously overloaded.

Ethernet uses
package size 1516 bytes, which is beautiful
approached when he was only created.
Today it is considered a disadvantage when
Ethernet used to interact
servers because servers and links
tend to share big
the number of small packages that
Overloads the network. In addition, Fast Ethernet
imposes limit on the distance between
connected devices - no more than 100
meters and it causes
Additional caution
Designing such networks.

First Ethernet was
Designed on the basis of tire topology,
when all devices connected to the total
Cable, thin or thick. Application
Twisted pair only partially changed the protocol.
Using coaxial cable
The collision was determined immediately by all
stations. In the case of twisted pair
used "jam" signal as soon as
Station determines the collision, then it
sends a signal to the concentrator last
turns in turn "jam" to all
Connected to it devices.

In order to
Reduce overload, Ethernet Standard Network
split into segments that
combine with the help of bridges and
routers. This allows you to transmit
There are only necessary traffic between segments.
Message transmitted between two
stations in one segment will not
transferred to another and will not be able to call it
Overload.

Today is
Building a central highway,
Unifying servers are used
Switchable Ethernet. Ethernet switches can
Consider as high-speed
multiport bridges that are able
independently determine which one
Ports addressed package. Switch
browsing packet headers and such
makes a table defining
where is one or another subscriber with such
physical address. This allows
Limit the packet distribution area
and reduce the likelihood of overflow,
Sending it only to the right port. Only
broadcast packets are sent by
All ports.

100Baset.
- older brother 10baset

Idea technology
Fast Ethernet was born in 1992. In August
next year group of manufacturers
Communicated in the FAST Ethernet Union (Fast Ethernet Alliance, FEA).
The purpose of FEA was as soon as possible
Formal approval FAST Ethernet from Committee
802.3 Institute of Electrical Engineering Engineers and
Radioelectronics (Institute of Electrical and Electronic
Engineers, IEEE), since this particular committee
Engaged in standards for Ethernet. Luck
accompanied new technology and
Supporting it Alliance: in June 1995
All formal procedures have been completed, and
Fast Ethernet technologies assigned the name
802.3u.

Light IEEE
Fast Ethernet is called 100Baset. This is explained
Simple: 100Baset is an extension
10Baset standard with bandwidth from
10 m bit / s to 100 Mbps. The standard 100Baset includes
multiple processing protocol
access with identifying carrier and
CSMA / CD conflict detection (Carrier Sense Multiple
Access with Collision Detection) that is used in
10Baset. In addition, FAST Ethernet can work on
cables of several types, including on
twisted pair. Both of these properties of the new
Standards are very important for potential
buyers, and it is thanks to them 100baset
It turns out to be successful by migrating networks
Based on 10Baset.

The main thing
Commercial argument in favor of 100Baset
is that FAST Ethernet is based on
Inherited technology. As in Fast Ethernet
The same transmission protocol is used.
messages as in old versions of Ethernet, and
Cable systems of these standards
Compatible, to go to 100Baset from 10Baset
Wanted

smaller
Capital investments than to install
other types of high-speed networks. Besides
how 100baset is
Continuation of the old Ethernet standard, all
Tools and procedures
network analysis, as well as all
Software operating on
Old Ethernet networks should be in this standard
Save performance.
Consequently, Wednesday 100Baset will be familiar
Network administrators with experience
with Ethernet. So, training personnel will take
less time and will cost significantly
Cheaper.

PRESERVATION
Protocol

Perhaps,
The greatest practical benefits
technology has brought a decision to leave
Message transfer protocol unchanged.
Message transfer protocol in our case
CSMA / CD, determines how to
transmitted over the network from one node to another
Through the cable system. In the ISO / OSI model
CSMA / CD protocol is part of the level
MEDIA ACCESS CONTROL, MAC).
At this level, the format is determined in
which information is transmitted over the network, and
The way that the network device receives
network access (or network management) for
data transmission.

Name CSMA / CD
You can smash into two parts: Carrier Sense Multiple Access
and Collision Detection. From the first part of the name you can
conclude how node with network
the adapter determines the moment when he
You should send a message. In accordance with
CSMA protocol, network node first "listens"
network to determine if is not transmitted to
The moment is any other message.
If the carrier signal is listening (Carrier Tone),
So at the moment the network is occupied by another
Message - Network node goes into mode
expectations and stays in it until the network
free When the network comes
Silence, the node begins the transfer.
In fact, the data is sent to all nodes.
network or segment, but are accepted only
the node to which they are addressed.

Collision Detection -
The second part of the name - serves to permit
situations where two or more nodes are trying
Transmit messages simultaneously.
According to the CSMA protocol, each ready for
The gear of the node must first listen to the network,
To determine whether it is free. But,
If two nodes listen to the same time,
both they decide that the network is free and start
Transmit your packages at the same time. In this
Situations Transmitted Data
superimposed on each other (network
Engineers call it conflict), and none
from messages does not reach the point
destination. Collision Detection requires a node
listened to the network also after the transfer
Package. If conflict is found, then
The node repeats the transmission through random
the selected time lapse and
He again checks if there was no conflict.

Three kinds of Fast Ethernet

As well as
Conservation of the CSMA / CD protocol, other important
The decision was to design 100Baset such
so that it can be used in it
Cables different types - like those that
used in old versions of Ethernet and
Newer models. Standard defines three
Modifications to ensure work with
Different types of Fast Ethernet cables: 100Basetx, 100Baset4
and 100BaseFX. 100Basetx and 100Baset4 modifications are calculated
on the vitua couple, and 100BaseFX was designed for
optical cable.

Standard 100Basetx
Requires two pairs of UTP or STP. One
Couple serves to transfer, the other - for
reception. These requirements are responsible two
Basic Cable Standards: EIA / TIA-568 UTP
IBM type 1 and STP categories. In 100basetx
Attractive provision
full-duplex mode when working with
network servers, as well as use
just two of four steam eight
cable - two other couples remain
free and can be used in
further to expand opportunities
network.

However, if you
Going to work with 100Basetx using for
this wiring category 5, then you should
Know about its shortcomings. This cable
more than other eight cables (for example
Category 3). In addition, to work with it
Using the use of broken blocks (Punchdown
Blocks), connectors and switting panels,
satisfying the requirements of category 5.
Need to add that to support
Full duplex mode
Install full duplex switches.

Standard 100Baset4.
Differs on softer requirements for
Used cable. The reason for this
The circumstance that in 100Baset4 are used
All four pairs of eight cable: one
for transmission, other for reception, and
The remaining two work as a transmission,
And at the reception. Thus, in 100Baset4 and reception,
and data transfer can be carried out by
Three pairs. Folding 100 Mbps for three pairs,
100BaseT4 reduces the frequency of the signal, so
For its transfer is quite and less
high quality cable. For implementation
100baset4 networks are suitable UTP Category 3 Cables and
5, as well as UTP category 5 and STP type 1.

Advantage
100Baset4 is less hard
Wiring requirements. Cables categories 3 and
4 are more common, and, moreover, they
significantly cheaper than cables
Category 5, which should not be forgotten before
start of installation work. Disadvantages
They are that all four are needed for 100baset4
Couples and that full-duplex mode by this
The protocol is not supported.

Fast Ethernet includes
Also standard for working with multimode
Fiber optic with 62.5-micron core and 125 micron
shell. 100BaseFX standard is oriented in
Main on highway - to connect
Fast Ethernet repeaters within one
building. Traditional advantages
optical cable inherent and standard
100BaseFX: Electromagnetic Resistance
noise, improved data protection and large
Distances between network devices.

RUNNER
On short distances

Although Fast Ethernet and
is a continuation of the Ethernet standard,
The transition from the 10Baset network to 100Baset can not
consider
equipment - for this can
Establish changes in the network topology.

Theoretical
Fast Ethernet network segment diameter limit
is 250 meters; it's just 10
percent of the theoretical size limit
Ethernet network (2500 meters). This limitation
stems from the character of the CSMA / CD protocol and
Transmission rates 100Mbps.

What already
noted earlier transmitting data
The workstation must listen to the network in
time to make sure
The data has reached the destination station.
In the Ethernet network with bandwidth 10
Mbit / s (for example 10Base5) time lapse,
The necessary workstation for
Listening to the network for conflict,
determined by the distance that is 512-bit
Frame (frame size is set in Ethernet standard)
will be held during the processing of this frame on
workstation. For the Ethernet network with bandwidth
10 Mbps capacity is the distance equal
2500 meters.

On the other hand,
The same 512-bit frame (802.3u standard
Specifies the frame of the same size as 802.3, then
There are 512 bits) transmitted by the working
Station in the Fast Ethernet network will be only 250 m,
Before the workstation completes it
Processing. If the receiving station was
removed from the transmitting station on
The distance over 250 m, then the frame could
join conflict with another frame on
lines somewhere on, and transmitting
Station, completing the transfer, no longer
Would perceive this conflict. therefore
Maximum 100Baset network diameter is
250 meters.

To
use a permissible distance
Two repeaters will be required to connect.
All nodes. According to the standard
Maximum distance between the node and
The repeater is 100 meters; In Fast Ethernet,
As in 10baset, the distance between
the hub and workstation is not
must exceed 100 meters. Insofar as
Connectors (repeaters)
Additional delays, real
working distance between nodes can
It is even smaller. therefore
It seems reasonable to take everything
Distances with some reserve.

To work on
long distances will have to be purchased
optical cable. For example, equipment
100BaseFX in half duplex mode allows
Connect the switch with another switch
or the final station located on
Distance to 450 meters from each other.
Setting the full duplex 100BaseFX, you can
Connect two network devices on the
Distance to two kilometers.

AS
Set 100Baset.

Except cables
which we have already discussed, to install Fast
Ethernet will require network adapters for
Workstations and servers, hubs
100Baset and possibly some amount
Switches 100Baset.

Adapters
necessary for the organization of the network 100Baset,
The name of Ethernet adapters 10/100 Mbps.
These adapters are capable (this requirement
100BaseT standard) to distinguish 10
Mbit / s from 100 Mbps. To serve the group
servers and workstations translated on
100Baset, also requires a 50Baset concentrator.

When turned on
Server or personal computer with
10/100 Last adapter gives a signal,
informing that it can provide
Bandwidth 100Mbps. If a
host station (most likely it is
there will be a hub) is also designed for
work with 100Baset, it will return a signal in response
which and a hub, and a PC or server
Automatically switch to 100Baset mode. If a
The hub works only from 10Baset, it is not
Submits a response signal, and a PC or server
Automatically go into 10Baset mode.

When
Small-scale configurations 100Baset can
Apply a bridge or switch 10/100, which
ensure the connection of the part of the network operating with
100Baset, with an existing network
10Baset.

Deceptive
RAPIDITY

Summing up everything
The foregoing, we note that, as it seems to us,
Fast Ethernet is the most good for solving problems
High peak loads. For example, if
Someone from users works with CAD or
Image processing programs and
needs to increase bandwidth
ability, then Fast Ethernet may be
Good outlet. However, if
Problems are caused by an excessive number
users online, then 100baset starts
inhibit the exchange of information at about 50 percent
loading the network - in other words, on the same
level as 10Baset. But in the end, it is
After all, nothing more than expansion.

Today it is almost impossible to detect a laptop on sale or motherboard Without an integrated network card, or even two. The connector in all of them is one - RJ45 (more precisely, 8p8c), but the speed of the controller may differ by an order. In cheap models - this is 100 megabits per second (Fast Ethernet), in more expensive - 1000 (Gigabit Ethernet).

If there is no built-in LAN-controller in your computer, then it is most likely an old man on the basis of an Intel Pentium 4 or AMD Athlon XP processor, as well as their "ancestors". Such "dinosaurs" can "make friends" with wired network Only by installing a discrete network card with a PCI connector, since the PCI Express bus during their appearance to the light has not yet existed. But also for PCI bus (33 MHz) "Networks" supporting the most relevant Gigabit Ethernet standard are available, although its throughput may not be enough to fully disclose the high-speed potential of the gigabit controller.

But even in the case of the presence of a 100 megabit integrated network card, the discrete adapter will have to be purchased to those who are going to "prof-upgrade" to 1000 megabits. The best option will be the purchase of a PCI Express controller, which will provide the maximum speed of the network, unless, of course, the corresponding connector is present in the computer. True, many will prefer the PCI card, as they are much cheaper (the cost begins literally from 200 rubles).

What advantages will give in practice the transition from Fast Ethernet on Gigabit Ethernet? How distinguishes the actual data transfer rate of PCI versions of network cards and PCI Express? Enough the speed of ordinary hard disk To fully download a gigabit channel? Answers to these questions you will find in this material.

Test participants

For testing, three cheapest discrete network cards were selected (PCI - Fast Ethernet, PCI - Gigabit Ethernet, PCI Express - Gigabit Ethernet), as they enjoy the greatest demand.

The 100-megabit network PCI card is represented by the ACORP L-100S model (the price begins from 110 rubles), which uses the most popular REALTEK RTL8139D chipset for cheap cards.

A 1000-megabit network PCI card is represented by the ACORP L-1000S model (the price begins from 210 rubles), which is based on the Realtek RTL8169SC chip. This is the only map with the radiator on the chipset - the rest of the testing participants is not required.

1000-megabit network PCI Express map is represented by the TP-Link TG-3468 model (the price begins from 340 rubles). And she did not exception - it is based on the RTL8168B chipset, which is also produced by Realtek.

Exterior Network Card

Chipsets from these families (RTL8139, RTL816X) can be seen not only on discrete network cards, but also integrated on many motherboards.

The characteristics of all three controllers are shown in the following table:

Show Table

The PCI-bus bandwidth (1066 Mbps) theoretically should be sufficiently enough for the "roll" of gigabit network cards until full speed, but in practice it can still not be enough. The fact is that this "channel" is divided by all PCI devices among themselves; In addition, it is broadcast for service information on the maintenance of the tire itself. Let's see if this assumption is confirmed with a real dimension.

Another nuance: the vast majority of modern hard drives have an average read speed of no more than 100 megabytes per second, and often even less. Accordingly, they will not be able to provide a full load of the gigabit channel of the network card, the speed of which is 125 megabytes per second (1000: 8 \u003d 125). By traveling this restriction in two ways. The first is to combine a pair of such hard drives in the RAID-array (RAID 0, STRIPING), while the speed may increase almost twice. The second is to use SSD drives, whose speed parameters are noticeably higher than those of hard drives.

Testing

As a server, a computer was used with the following configuration:

  • processor: AMD Phenom II X4 955 3200 MHz (four-core);
  • motherboard: ASRock A770DE AM2 + (chipset AMD 770 + AMD SB700);
  • rAM: HYNIX DDR2 4 x 2048 GB PC2 8500 1066 MHz (in two-channel mode);
  • video card: AMD Radeon. HD 4890 1024 MB DDR5 PCI EXPRESS 2.0;
  • network card: Realtek RTL8111DL 1000 Mbps (integrated on the motherboard);
  • operating system: Microsoft Windows 7 Home Premium SP1 (64-bit version).

As a client in which test network cards were installed, a computer was used with the following configuration:

  • processor: AMD Athlon 7850 2800 MHz (dual-core);
  • motherboard: MSI K9A2GM V2 (MS-7302, AMD RS780 + AMD SB700 chipset);
  • rAM: HYNIX DDR2 2 x 2048 GB PC2 8500 1066 MHz (in two-channel mode);
  • video card: AMD Radeon HD 3100 256 MB (integrated into chipset);
  • hDD: Seagate 7200.10 160 GB SATA2;
  • operating system: Microsoft Windows XP HOME SP3 (32-bit version).

Testing was performed in two modes: reading and writing through network connection with hard drives (it should show that they can be a "bottle of neck"), as well as with RAM disks in random access memory Computers imitating fast SSD drives. Network cards were connected directly with the help of a three-meter patch cord (eight-tie steam, category 5e).

Data transfer rate (Hard disk - Hard disk, Mbit / s)

The real data transfer rate through a 100-megabit network card ACORP L-100S did not quite slightly reach the theoretical maximum. But both gigabit cards though overtook the first six times, but did not manage to show the highest possible speed. It is perfectly clear that the speed "rigorous" into the performance of seagate 7200 10 hard disks, which, with direct testing on a computer, an average of 79 megabytes per second (632 Mbps).

The fundamental difference in the speed between network cards for the PCI bus (acorp L-1000S) and PCI Express (TP-Link) is not observed in this case, a slight advantage of the latter is quite possible to explain the measurement error. Both controllers worked about sixty percent of their capabilities.

Data transfer rate (RAM drive - RAM disk, Mbps)

The ACORP L-100S expected shown the same low speed and when copying data from high-speed RAM disks. It is clear - the Fast Ethernet standard has long been not consistent with modern realities. Compared to the test mode "Hard disk - Hard disk" The Gigabit PCI card of the ACORP L-1000S was noticeably added in performance - the advantage was about 36 percent. An even more impressive gap showed a TP-LINK TG-3468 network card - an increase was about 55 percent.

Here, the PCI Express bus bandwidth manifested itself - bypassed the ACORP L-1000S by 14 percent, which is no longer wring for an error. The winner did not stretch a little to the theoretical maximum, but also the speed of 916 megabits per second (114.5 MB / s) still looks impressive - this means that it is possible to expect the end of copying almost an order of magnitude less (compared to Fast Ethernet). For example, the time copying time of 25 GB (typical HD rip with good quality) from a computer to a computer will be less than four minutes, and with the adapter of the previous generation - more than half an hour.

Testing has shown that the Gigabit Ethernet network cards are simply a huge advantage (up to tenfold) over Fast Ethernet controllers. If your computers are installed only hard drivesnot combined into a striping array (RAID 0), then the fundamental difference in the speed between PCI and PCI Express cards will not. Otherwise, as well as using productive SSD drives, preference should be given maps with the PCI Express interface, which will ensure the maximum possible data transfer rate.

Naturally, it should be borne in mind that the rest of the devices in the network "tract" (switch, router ...) must support the Gigabit Ethernet standard, and the category of twisted pair (patch cord) should be not lower than 5e. Otherwise, the actual speed will remain at the level of 100 megabits per second. By the way, backward compatibility with the Fast Ethernet standard is saved: You can connect a gigabit network, for example, a laptop with a 100 megabit network card, at the speed of other computers in the network it will not affect.

Introduction

The purpose of the creation of this report was a brief and affordable presentation of the basic principles of work and features of computer networks, on the example of Fast Ethernet.

The network is called a group of connected computers and other devices. The main purpose of computer networks is the sharing of resources and the implementation of interactive communication both within one company and abroad. Resources are data, applications and peripherals, such as an external drive, printer, mouse, modem or joystick. The concept of interactive communication of computers implies messaging in real time.

There are many sets of data standards in computer networks. One set is the Fast Ethernet standard.

Of this material You will learn about:

  • · Fast Ethernet technologies
  • · Switches
  • · FTP cable
  • · Types of compounds
  • · Topologies computer network

In my work, I will show the principles of the network based on the Fast Ethernet standard.

Switching local computing networks (LAN) and FAST Ethernet technology have been developed in response to the need to improve the efficiency of Ethernet network operation. By increasing bandwidth, these technologies can eliminate "bottlenecks" on the network and support applications that require high data transfer rates. The attractiveness of these solutions is that you do not need to choose or another. They are complementary, so the efficiency of the network functioning is most often possible by using both technologies.

The collected information will be useful, both persons starting to learn computer networks and network administrators.

1. Network scheme

2. Fast Ethernet Technology

computer Network Fast Ethernet

Fast Ethernet - ETHERNET technology development result. Based and keeping in inviolability The same CSMA / CD method (collective access with channel poll and collision detection), FAST Ethernet devices operate at a speed, 10 times higher than the Ethernet speed. 100 Mbps. Fast Ethernet provides sufficient bandwidth for applications as automated design and production systems (CAD / CAM), graphics and image processing, multimedia. Fast Ethernet is compatible with 10 Mbps Ethernet, so that the Fast Ethernet integration into your LAN is more convenient to implement the switch, and not the router.

Switch (Switch)

Using switches Many working groups can be interconnected to form a large LAN (see scheme 1). Cheap switches work better than routers, ensuring higher efficiency of LAN functioning. Fast Ethernet Working Groups, including one or two hubs, can be connected via the Fast Ethernet switch to further increase the number of users, as well as coverage of a more extensive area.

As an example, consider the following switch:

Fig. 1 D-LINK-1228 / ME

DES-1228 / ME Switch Series includes custom FAST switches Ethernet level 2 "Premium" class. Possessing advanced functionality, DES-1228 / ME devices are inexpensive solution to create a secure and high-performance network. Distinctive features of this switch are high port density, 4 UPLink Gigabit Port, a small step of changing the settings to control the bandwidth and improved network control. These switches allow you to optimize the network both in terms of functionality and in terms of cost characteristics. The Switches of the DES-1228 / ME series are the best solution for both the functionality and value characteristics.

FTP cable

LAN-5EFTP-BL cableconsists of 4 pairs of single-core copper conductors.

Conductor diameter 24AWG.

Each conductor is concluded in HDPE insulation (high density polyethylene).

Two conductors twisted with a specially selected step make up one twisted pair.

4 twisted pairs are wrapped with polyethylene film and together with a copper single-core grounding conductor enclosed in general screen Foil and PVC shell (PVC).

Direct connection (Straight Through)

It serves:

  • 1. To connect the computer to the switch (hub, switches) via the computer's network card
  • 2. To connect to the switch (hub, switches) network peripheral equipment - printers, scanners
  • 3. For uplink "And on the above-standing switch (hub, switch) - modern switches can automatically customize the inputs in the reception connector

Crossover (Crossover)

It serves:

  • 1. For direct connection of 2 computers to the local network, without the use of switching equipment (hubs, switches, routers, etc.).
  • 2. For uplink, connecting to the above-standing switch in a complex area of \u200b\u200bthe local network, for old types of switches (hubs, switches), they have a separate connector, and a marked "Uplink" or sign H.

Topology Star

Stars - The basic topology of the computer network, in which all network computers are attached to the central node (usually switch), forming a physical network segment. Such a network segment can function both separately and as part of a complex network topology (as a rule, "tree"). The whole exchange of information goes exclusively through a central computer, which is in such a way a very large load is assigned, so it cannot be engaged in anything else. As a rule, it is the central computer that is the most powerful, and it is precisely all the functions for the exchange management. No conflicts on the network with a topology star are not possible, because the management is fully centrally.

application

The classic 10 megabit ethernet arranged most users for about 15 years. However, in the early 1990s, its insufficient bandwidth began to be felt. For computers on the Intel 80286 or 80386 processors with ISA tires (8 MB / s) or EISA (32 MB / s), the bandwidth of the Ethernet segment was 1/8 or 1/32 channel "Memory-Disk", and this was well agreed with the ratio Data volumes processed locally and data transmitted over the network. For more powerful client stations with PCI bus (133 MB / s), this share fell to 1/133, which was clearly not enough. Therefore, many segments of 10 megabit ethernet have become overloaded, the response of servers in them has fallen significantly, and the frequency of the occurrence of collisions has increased significantly, even reducing the useful bandwidth.

There is a need to develop a "new" Ethernet, that is, technologies that would be as effective in terms of price / quality ratio with 100 Mbps performance. As a result of searches and research, specialists were divided into two camps, which eventually led to the emergence of two new technologies - Fast Ethernet and L00VG-Anylan. They are distinguished by the degree of continuity with the classic Ethernet.

In 1992, a group of manufacturers of network equipment, including such Ethernet technology, such as Synoptics, 3Com and a number of others, formed a non-commercial association FAST Ethernet Alliance to develop a new technology standard, which was to maintain the features of Ethernet technology as much as possible.

The second camp was headed by Hewlett-Packard and AT & T, which were offered to take advantage of the convenient occasion to eliminate certain known disadvantages of Ethernet technology. After some time, IBM has joined these companies, which contributed proposal to ensure some compatibility with TKen Ring networks.

In the IEEE Institute Committee 802 at the same time a research team was formed to study the technical potential of new high-speed technologies. For the period from the end of 1992 and at the end of 1993, the IEEE group studied 100 megabit solutions proposed by various manufacturers. Along with the offers of Fast Ethernet Alliance, the Group also considered high-speed technology proposed by Hewlett-Packard and AT & T companies.

In the center of the discussion there was a problem of saving random method CSMA / CD access. The Fast Ethernet Alliance proposal retained this method and thereby ensured the continuity and consistency of 10 Mbps and 100 Mbps networks. The HP and AT & T coalition that had supported a significantly smaller number of manufacturers in the network industry than the Fast Ethernet Alliance, offered a completely new access method called Demand Priority. - Priority access on demand. He significantly changed the picture of the behavior of nodes in the network, so I could not fit into ethernet technology And Standard 802.3, and a new IEEE 802.12 Committee was organized for its standardization.

In the fall of 1995, both technologies have become IEEE standards. The IEEE 802.3 Committee adopted the Fast Ethernet specification as standard 802.3 and, which is not an independent standard, and is an addition to the existing standard 802.3 in the form of chapters from 21 to 30. Committee 802.12 adopted L00VG-AnyLAN technology, which uses a new Demand Priority access method And supports frames of two formats - Ethernet and Token Ring.

v. Physical level of technology Fast Ethernet

All differences between the Fast Ethernet technology from Ethernet are focused on the physical level (Fig. 3.20). Mac and LLC levels in Fast Ethernet remained absolutely the same, and they are described by the previous chapters of standards 802.3 and 802.2. Therefore, considering FAST Ethernet technology, we will study only a few options for its physical level.

The more complex structure of the physical level of FAST Ethernet technology is caused by the fact that it uses three options for cable systems:

  • · Fiber optic multimode cable, two fibers are used;
  • · Twisted pair of category 5, two pairs are used;
  • · Twisted pair of category 3, four pairs are used.

Coaxial cable, who gave the world to the first Ethernet network, in the number of allowed data transmission media, Fast Ethernet did not hit. This is a general trend of many new technologies, because at short distances, twisted pair of category 5 allows you to transmit data at the same speed as a coaxial cable, but the network is cheaper and more convenient to operate. At large distances, the optical fiber has a much broader bandwidth than the coaxial, and the network costs turns out not much higher, especially considering the high costs of troubleshooting and troubleshooting in a large cable coaxial system.


Differences Fast Ethernet technology from Ethernet technology

The rejection of the coaxial cable led to the fact that the Fast Ethernet networks always have a hierarchical tree structure constructed at hubs, as well as the L0Base-T / L0Base-F network. The main difference of configurations networks Fast. Ethernet is to reduce the diameter of the network to about 200 m, which is explained by a decrease in the transfer time of the frame of the minimum length 10 times due to an increase in the transmission rate of 10 times compared to 10 megabit Ethernet.

Nevertheless, this circumstance does not greatly prevent the construction of large networks on FAST Ethernet technology. The fact is that the mid-90s is noted not only to the wide distribution of inexpensive high-speed technologies, but also the rapid development of local networks based on switches. When using switches, the Fast Ethernet protocol can operate in full-duplex mode in which there are no restrictions on long length Networks, but only restrictions on the length of physical segments connecting adjacent devices (adapter - switch or switch - switch). Therefore, when creating high-length local networks, the Fast Ethernet technology is also actively used, but only in the full duplex version, together with switches.

This section discusses the half-duplex version of the Fast Ethernet technology, which fully complies with the definition of the access method described in Standard 802.3.

Compared with the options for the physical implementation of Ethernet (and there are six there are six), in the Fast Ethernet, the differences of each option from other deeper is variable as the number of conductors and coding methods. And since the physical options of Fast Ethernet were created simultaneously, and not evolutionally, as for Ethernet networks, it was possible to determine in detail those philodes of the physical layer that do not change from the variant to the option, and those suproes that are specific for each physical environment.

The official standard 802.3 and has set three different specifications for the physical level of Fast Ethernet and gave them the following names:

Fast Ethernet Physical Level Structure

  • · 100Base-TX for a two-party cable on unshielded twisted pair of UTP category 5 or shielded twisted pair STP Type 1;
  • · 100Base-T4 for a four-pane cable on unshielded twisted pair UTP category 3, 4 or 5;
  • · 100Base-FX for multimode fiber optic cable, two fiber are used.

For all three standards, the following statements and characteristics are valid.

  • · Formats of FAST Ethernetee technology differ from frame formats of 10 megabit ethernet technologies.
  • · Intercader interval (IPG) is 0.96 μs, and the bit interval is 10 ns. All temporal parameters of the access algorithm (deferred interval, minimum length transmission time, and the like), measured in bit intervals, remained the same, so the changes to the Standard sections relating to the MAC level were not made.
  • · The sign of the free state of the medium is the transmission of an IDLE symbol of the corresponding redundant code (and not the lack of signals as in Ethernet standards 10 Mbps). The physical layer includes three items:
  • o RECONCILITION SUBLAYER;
  • o independent of the media interface (MILIA INDEPENDENT INTERFACE, MIL);
  • o Physical Layer Device, PHY).

The level of matching is needed so that the MAC level designed for the AUI interface is able to work with the physical level through the MP interface.

The physical layer device (PHY) consists, in turn, from several sublevels (see Fig. 3.20):

  • · Logical coding of data that converts from MAC bytes coming from the 4B / 5B or 8B / 6T code symbols (both codes are used in Fast Ethernet technology);
  • · Pries of physical attachment and a dependence on the physical environment (PMD), which ensure the formation of signals in accordance with the method of physical coding, for example NRZI or MLT-3;
  • · Auto-traveler suite, which allows two interactive ports to automatically select the most efficient mode of operation, for example, half-duplex or full-duplex (this sublayer is optional).

The MP interface supports an independent physical environment of the data exchange between the MAC sublayer and PHY sublayer. This interface is similar to assigning the AUI interface of the classic Ethernet except that the AUI interface was located between the paragraph of the physical signal encoding (for any cable options, the same physical coding method is used - manchester code) and a paragraph of physical connection to the environment, and the MP interface is located between the MAC sublayer and a signal encoding sublevels that are three-FX, TX and T4 in the Fast Ethernet standard.

The MP connector unlike the AUI connector has 40 contacts, the maximum cable length of the MP is one meter. Signals transmitted over the MP interface have amplitude 5 V.

Physical level 100Base-FX - multimode fiber, two fibers

This specification determines the operation of the FAST Ethernet protocol on the multimode fiber in the half-duplex and complete duplex modes based on a well-proven FDDI encoding scheme. As in the FDDI standard, each node is connected to the network with two optical fibers that are from the receiver (R x) and from the transmitter (T x).

There are many common properties between the L00Base-FX and L00Base-TX specifications, therefore the common property for two specifications will be given under the generalized name L00Base-FX / TX.

While Ethernet with a transmission rate of 10 Mbit / s uses Manchester encoding to represent data when transmitted via cable, a different coding method is defined in the Fast Ethernet standard - 4V / 5B. This method has already shown its efficiency in the FDDI standard and changed to the L00Base-FX / TX specification. In this method, every 4 data bits of Mac sublayer (called symbols) are presented with 5 bits. Excess bit allows you to apply potential codes when representing each of the five bits in the form of electrical or optical pulses. The existence of prohibited combinations of symbols allows you to reject erroneous characters, which increases the stability of the operation of networks with L00Base-FX / TX.

To separate the Ethernet frame from IDLE characters, a combination of Start Delimiter characters is used (pair of symbols J (11000) and to (10001) 4B / 5B code, and after the frame is completed before the first idle character, the T. symbol is inserted.


Continuous data stream 100Base-Fx / TX specifications

After converting 4-bit portions of Mac codes into 5-bit portions of the physical layer, they must be represented as optical or electrical signals in a cable connecting network nodes. L00BASE-FX and L00BASE-TX specification and L00BASE-TX are used for this. various methods Physical coding - NRZI and MLT-3, respectively (as in FDDI technology when working through fiber optic and twisted pair).

Physical level 100Base-TX - twisted pair of DTP Cat 5 or STP Type 1, two pairs

As a data transfer environment, the L00Base-TX specification uses the UTP Category 5 cable or STP Type cable 1. The maximum cable length in both cases is 100 m.

The main differences from the L00Base-FX specification is the use of the MLT-3 method for transmitting 5-bit portions of the 4V / 5B code in twisted pair, as well as the availability of auto-track function (AUTO-NEGOTIATION) to select the port operation mode. The auto-track scheme allows two connected physically devices that support several standards of physical layer, characterized by bit speed and number of twisted pairs, select the most profitable mode of operation. Typically, the auto-bargoing procedure occurs when a network adapter is connected, which can operate at a speed of 10 and 100 Mbps, to a hub or switcher.

The AUTO-NEGOTIATION scheme described below is the L00BASE-T technology standard. Prior to this, manufacturers used various own schemes. automatic definition The speed of operation of interacting ports that were not compatible. AUTO-NEGOTIATION scheme asked as a standard, National Semiconductor is originally called NWAY.

In total, 5 different modes of operation are currently identified, which can support L00Base-TX or 100Base-T4 devices on twisted pairs;

  • · L0Base-T - 2 pairs of category 3;
  • · L0Base-T Full-duplex - 2 pairs of category 3;
  • · L00BASE-TX - 2 pairs of category 5 (or Type 1Astp);
  • · 100Base-T4 - 4 pairs of category 3;
  • · 100Base-TX Full-Duplex - 2 pairs of category 5 (or Type 1A STP).

The L0Base-T mode has the lowest priority in the negotiation process, and the total duplex mode 100Base-T4 is the highest. The negotiation process occurs when the device is turned on, and can also be initiated at any time the device control module.

The device that has begun the AUTO-NEGOTIATION process sends a package of special pulses to its partner. FAST LINK PULSE BURST (FLP)which contains an 8-bit word encoding the proposed interaction mode, starting from the priority supported by this node.

If the partner node supports the AUTO-NEGOTICATION function and can also support the proposed mode, it corresponds to the FLP pulse pack, in which this mode confirms, and the negotiations end. If the partner node can maintain less priority mode, then it indicates it in response, and this mode is selected as a worker. Thus, the most priority general mode of nodes is always selected.

A node that supports only L0Base-T technology, every 16 ms sends Manchester pulses to check the integrity of the line connecting it with the adjacent node. Such a node does not understand the FLP request, which makes it a node with the AUTO-NEGOTIATION function, and continues to send its impulses. The node that received in response to the FLP request only the impulses of the integrity of the line understands that its partner can only work according to the L0Base-T standard, and sets this mode of operation and for itself.

Physical level 100Base-T4 - twisted pair UTP Cat 3, four pairs

The 100Base-T4 specification was designed to be used for high-speed Ethernet available wiring on a twisted pair of category 3. This specification allows you to enhance the total bandwidth due to the simultaneous transmission of the bit streams on all 4 cable pairs.

The 100Base-T4 specification appeared later than other physical level specifications Fast Ethernet. The developers of this technology first wanted to create physical specifications that are closest to the L0Base-T and L0Base-F specifications that operated on two data transmission lines: two pairs or two fibers. For the implementation of work on two twisted pairs I had to go to a better Category 5 cable.

At the same time, the developers of the competing technology L00VG-Anylan initially made a bet on a twisted pair of category 3; The most important advantage was not so much in value, but in the fact that it was already laid in the overwhelming number of buildings. Therefore, after the release of the L00BASE-TX and L00BASE-FX specifications and L00BASE-FX developers, Fast Ethernet technology has implemented their own version of the physical layer for twisted category 3 pairs.

Instead of coding 4B / 5V, in this method, coding 8B / 6t is used, which has a narrower spectrum of the signal and at a speed of 33 Mbit / s fit into the 12 MHz band twisted a pair of category 3 (when encoding 4V / 5V, the signal spectrum does not fit into this strip) . Each 8 bits of the Mac level information are encoded with 6-Tropic figures (Ternary Symbols), that is, numbers having three states. Each tricious digit has a duration of 40 ns. A group of 6-terrible digits is then transmitted to one of the three transmitted twisted pairs, independently and sequentially.

The fourth pair is always used to listen to the carrier frequency in order to detect the collision. Data transfer rate for each of the three transmit steam is 33.3 Mbps, so the total speed of the 100Base-T4 protocol is 100 Mbps. At the same time, due to the accepted method of encoding, the signal change rate at each pair is only 25 mbeda, which allows the use of twisted pair of category 3.

In fig. 3.23 shows the connection of the MDI port of the MDI network adapter 100Base-T4 with the MDI-X of the hub (the console does not say that this connector connects the receiver and the transmitter changes in cable pairs compared to the network adapter connector, which makes it easier to connect pair of wires in the cable. without crossing). Couple 1 -2 always required for data transfer from MDI port to port MDI-X, steam 3 -6 - To receive data from the MDI port from the port of MDI-X, and pairs 4 -5 and 7 -8 They are bidirectional and used for both reception and transmission, depending on the need.


Connection of nodes by specification 100Base-T4