The real beginning was the ALOHA system
constructed to allow radio communication between machines
scattered over the Hawaiian Islands . Later, carrier
sensing was added, and Xerox PARC built a 2.94-Mbps
CSMA(Carrier Sense Multiple Access)/CD system to connect
over 100 personal work-stations on a 1-km cable (Metcalfe
and Boggs, 1976). This system was called Ethernet
after the luminiferous ether , through
which electromagnetic radiations propagate. (When the
Nineteenth Century British physicist James Clerk Maxwell
discovered that electromagnetic radiation could be described
by a wave equation, scientists assumed that space must
be filled with some ethereal medium in which the radiation
was propagating. Only after the famous Michelson- Morley
experiment in 1887, did physicists discover that electromagnetic
radiation could propagate in vacuum).
The Xerox Ethernet was so successful that Xerox, DEC,
and Intel drew up a standard for a 10 - Mbps Ethernet.
This standard formed the basis for 802.3. The published
802.3 standard differs from the Ethernet specification
in that it describes a whole family of 1-persistent
CSMA/CD systems, running at speeds from 1 t0 10-Mbps
on various media. Also, the one header field differs
between the two (the 802.3 length field is used for
packet type in Ethernet). The initial standard also
gives the parameters for a 10 Mbps base-band system
using 70-ohm coaxial cable. Parameter sets for other
media and speeds came later.
Many people (incorrectly) use the name "Ethernet" in
a generic sense to refer to all CSMA/CD protocols, even
though it really refers to a specific product that almost
implements 802.3. We will use the terms "802.3" and
"CSMA/CD" except when specifically referring to the
Ethernet product in the next few paragraphs.
802.3 Cabling
Since the name "Ethernet"
refers to the cable (the ether), let us start our discussion
there. Four types of cabling are commonly used, as shown
in table that below. Historically, 10Base5
cabling, popularly called thick Ethernet,
came first. It resembles a yellow garden hose, with
markings every 2.5 meters to show where the taps go.
(The 802.3 standard does not actually require the
cable to be yellow, but it does suggest it.)
Connections to it are generally made using vampire
taps, in which a pin is carefully forced halfway
into the coaxial cable's core. The notation 10Base5
means that it operates at 10 Mbps, uses base-band signaling
can support segments of up to 500 meters.
Name
Cable
Max.
segment
Nodes/seg.
Advantages
10Base5
Thick
coax
500
m
100
Good
for backbones
10Base2
Thin
coax
200
m
30
Cheapest
system
10Base-T
Twisted
pair
100
m
1024
Easy
maintenance
10Base
-F
Fiber
optics
2000
m
1024
Best
between building
The most common kinds of
base-band 802.3 LANs
Historically, the second cable type was 10Base2
or thin Ethernet, which,
contrast to the garden-hose-like thick Ethernet- bends
easily. Connections are made using industry standard
BNC connectors to form T junctions, rather than using
vampire taps. These are easier to use and are more reliable.
Thin Ethernet is much cheaper and easier to install,
but it can run for only 200 meters and can handle only
30 machines per cable segment.
Detecting cable breaks, bad taps, or loose connectors
can be a major problem with both media. For this reason,
techniques have been developed to track them down. Basically,
a pulse of known shape is injected into the cable. If
the pulse hits an obstacle or the end of the cable,
an echo will be generated and sent back . By carefully
timing the interval between sending the pulse and receiving
the echo, it is possible to localize the origin of the
echo. This technique is called Tim domain reflectometry.
The problems associated with finding cable
breaks have driven system towards a different kind of
wiring pattern, in which all stations have a cable running
to a central hub . Usually, these wires
are telephone company twisted pairs, since most office
buildings are already wired this way, and there are
normally plenty of spare pairs available. This scheme
is called 10Base-T.
These three wiring schemes are shown below.A
transceiver is clamped securely around the
cable so that its tap makes contact with the inner core.
The transceiver contains the electronics that handle
carrier detection and collision detection. When a collision
is detected, the transceiver also puts a special invalid
signal on the cable to ensure that all other transceivers
also realize that a collision has occurred.
With 10Base5, a transceiver cable connects
the transceiver to an interface board in the computer.The
transceiver cable may be up to 50 meters long and contains
five individually shielded twisted pairs. Two of the
pairs are for data in and data out, respectively. Two
more are for control signals in and out. The fifth pair,
which is not always used, allows the computer to power
the transceiver electronics. Some transceivers allow
up to eight nearby computers to be attached to them,
to reduce the number of transceivers needed.
The transceiver cable terminates on an interface
board inside the computer. The interface board contains
a controller chip that transmits frames to, receives
frames from and the transceiver. The controller is responsible
for assembling the data into the proper frame format,
as well as computing checksums on outgoing frames and
verifying them on incoming frames. Some controller chips
also manage a pool of buffers for incoming frames, a
queue of buffers to be transmitted, DMA transfers with
the host computers, and other aspects of network management.
With 10Base2, the connection to the cable is just a
passive BNC T-junction connector. The transceiver electronics
are on the controller board, and station always has
its own transceiver.
With l0Base-T, there is no cable at all, just the hub
is a box full of electronic where the adding or removing
a station is simpler and cabled can be detected easily.
The disadvantage of l0Base-T is that the maximum run
from the hub is only 100 meters, maybe 150 meters if
high-quality (category 5) twisted pairs are used. Also,
a large hub costs thousands of dollars. Still l0Base-T
is becoming steadily more popular due to the ease of
maintenance that it offers.
A fourth cabling option for 802.3 is l0Base-F, which
uses fiber optics. This alternative is expensive due
to the cost of the connectors and terminators, but it
has excellent noise immunity and is the method of choice
when running between buildings or widely separated hubs.
Different ways of wiring up
a building
Above illustration shows different ways
of wiring up a building. In above illustration, a single
cable is snaked from room to room, with each station
tapping onto it at the nearest point. In the illustration
below, a vertical spine runs from the basement to the
roof.
Different ways
of wiring up a building
Fast Ethernet
FDDI was supposed to be the next generation
LAN, but it never really caught on much beyond the backbone
market (where it continues to do fine). The station
management was too complicated, which led to complex
chips and high prices. The substantial cost of FDDI
chips made workstation manufacturers unwilling to make
FDDI the standard network, so volume production never
happened and FDDI never broke through to the mass market.
The lesson that should have been learned here was KISS
(Keep It Simple, Stupid).
In any event, the failure of FDDI to catch fire left
a gap for a garden-variety LAN at speeds above 10 Mbps.
Many installations needed more bandwidth and thus had
numerous 10-Mbps LANs connected by a maze of repeaters,
bridges routers, and gateways, although to the network
managers it sometimes felt that they were being held
together by bubble gum and chicken wire.
It was in this environment that IEEE reconvened the
802.3 committee in 1992 with instructions to come up
with a faster LAN. One proposal was to keep 802.3 exactly
as it was, but just make it to go faster. Another proposal
was to redo it totally, to give it lots of new features,
such as real-time traffic and digitized voice but just
keep the old name (for marketing reasons). After some
wrangling, the committee decided to keep 802.3 the way
it was, but faster. The people behind the losing proposal
did what any computer-industry people would have done
under these circumstances-they formed their own committee
and standardized their LAN anyway (eventually as 802.12).
The three primary reasons that the 802.3
committee decided to go with a souped-up 802.3 LAN were:
The need to be backward compatible with thousands
of existing LANs.
The fear that a new protocol might have unforeseen
problems.
The desire to get the job done before the technology
changed.
The work was done quickly (by standards
committees' norms), and the result, 802.3u, was officially
approved by IEEE in June 1995. Technically, 802.3u is
not a new standard, but an addendum to the existing
802.3 standard (to emphasize its backward compatibility).
Since everyone calls it fast Ethernet, rather than 802.3u,
we will do that, too.
The basic idea behind fast Ethernet was simple: keep
all the old packet formats, interfaces, and procedural
rules, but just reduce the bit time from 100 nsec to
10 nsec. Technically, it would have been possible to
copy l0Base5 or 10Base-2 and still detect collisions
on time by just reducing the maximum cable length by
a factor of ten. However, the advantages of 10Base-T
wiring were so overwhelming, that fast Ethernet is based
entirely on this design. Thus all fast Ethernet systems
use hubs; multidrop cables with vampire taps or BNC
connectors are not permitted.
Nevertheless, some choices still had to be made, the
most important of which was which wire types to support.
One contender was category 3 twisted pair. The argument
for it was that practically every office in the Western
world has at least four category 3 (or better) twisted
pairs running from it to a telephone wiring closet within
100 meters. Sometimes two such cables exist. Thus using
category 3 twisted pair would make it possible to wire
up desktop computers using fast Ethernet without having
to rewire the building, an enormous advantage for many
organizations.
The main disadvantage of category 3 twisted pair is
its inability to carry 200 megabaud signals (100 Mbps
with Manchester encoding) 100 meters, the maximum computer-to-hub
distance specified for l0Base-T (see above table). In
contrast, category 5 twisted pair wiring can handle
100 meters easily, and fiber can go much further. The
compromise chosen was to allow all three possibilities,
as shown in table as below, but to pep up the category
3 solution to give it the additional carrying capacity
needed.
Name
Cable
Max.
segment
Advantages
100Base-T4
Twisted
pair
100
m
Uses
category 3 UTP
100Base-TX
Twisted
pair
100
m
Full
duplex at 100 Mbps
100Base-FX
Fiber
optics
2000
m
Full
duplex at 100 Mbps; long runs
Fast Ethernet cabling
The category 3 UTP scheme, called 100Base-T4,
uses a signaling speed of 25 MHz, only 25 percent faster
than standard 802.3's 20 MHz (remember that Manchester
encoding, as shown below, requires two clock periods
for each of the 10 million bits each second). To achieve
the necessary bandwidth, 100Base-T4 requires four twisted
pairs. Since standard telephone wiring for decades has
had four twisted pairs per cable, most offices are able
to handle this. Of course, it means giving up your office
telephone, but that is surely a small price to pay for
faster email.
Manchester Encoding
Of the four twisted pairs, one is always
to the hub, one is always from the hub, and other two
are switchable to the current transmission direction.
To get the necessary bandwidth, Manchester encoding
is not used, but with modern clocks and such short distances,
it is no longer needed. In addition, ternary signals
are sent, so that during a single clock period the wire
can contain a 0, a 1, or a 2. With three twisted pairs
going in the forward direction and ternary signaling,
any one of 27 possible symbols can be transmitted, making
it possible to send 4 bits with some redundancy. Transmitting
4 bits in each of the 25 million clock cycles per second
gives the necessary 100 Mbps. In addition, there is
always a 33.3 Mbps reverse channel using the remaining
twisted pair. This scheme, known as IB 8B6T, (8 bits
map to 6 trits) is not likely to win any prizes for
elegance, but it works with the existing wiring plant.
For category 5 wiring, the design, 100Base-TX, is simpler
because the wires
can handle clock rates up to 125 MHz and beyond. Only
two twisted pairs per
station are used, one to the hub and one from it. Rather
than just use straight
binary coding, a scheme called 4B5B is used at 125 MHz.
Every group of five
clock periods is used to send 4 bits in order to give
some redundancy, provide
enough transitions to allow easy clock synchronization,
create unique patterns for
frame delimiting, and be compatible with FDDI in the
physical layer. Consequently, 100Base-TX is a full-duplex
system; stations can transmit at 100 Mbps
and receive at 100 Mbps at the same time. In addition,
you can have two telephones in your office for real
communication in case the computer is fully occupied
with web surfing.
The last option, 100Base-FX, uses two strands of multimode
fiber, one for each direction, so it, too, is full duplex
with 100 Mbps in each direction. In addition, the distance
between a station and the hub can be up to 2 km.
Two kinds of hubs are possible with 100Base-T4 and 100Base-TX,
collectively known as 100Base-T. In a shared hub, all
the incoming lines (or at least all the lines arriving
at one plug-in card) are logically connected, forming
a single collision domain. All the standard rules, including
the binary back off algorithm, apply, so the system
works just like old-fashioned 802.3. In particular,
only one station at a time can be transmitting.
In a switched hub, each incoming frame is buffered on
a plug-in line card. Although this feature makes the
hub and cards more expensive, it also means that
all stations can transmit (and receive) at the same
time, greatly improving the total
bandwidth of the system, often by an order of magnitude
or more. Buffered
frames are passed over a high-speed backplane from the
source card to the destination card. The backplane has
not been standardized, nor does it need to be since
it is entirely hidden deep inside the switch. If past
experience is any guide, switch vendors will compete
vigorously to produce ever faster backplanes in order
to improve system throughput. Because 100Base-FX cables
are too long for the normal Ethernet collision algorithm,
they must be connected to buffered, switched hubs, so
each one is a collision domain unto itself.
As a final note, virtually all switches can handle a
mix of 10-Mbps and 100-
Mbps stations, to make upgrading easier. As a site acquires
more and more 100-
Mbps workstations, all it has to do is buy the necessary
number of new line cards
and insert them into the switch.
More information about Fast Ethernet can be found in
(Johnson, 1996). For a
comparison of high-speed local area networks, in particular,
FDDI, fast Ethernet
ATM, and VG-AnyLAN, see (Cronin et al.; 1994).
