|
Ethernet technology refers to a packaged based network that is most suitable for LAN (local area network) Environments and includes LAN products of the IEEE 802.3ae. 10 Gigabit Ethernet standards which is an up gradation on all other earlier versions, works on an optical fibre and operates in full duplex mode (Held, 1996). The IEEE 802.3ae is an upgraded version of the IEEE 802.3 base version developed as early as 1980. Ethernet technology originated nearly 30 years ago as a least expensive high-speed LAN based option (Ferrero, 1996).
Ethernet defines the CSMA/CD protocol and is currently used
in major LAN-based settings. Ethernet has the advantages of low
implementation cost, simplicity and ease of installation and
maintenance, reliability and compatibility with LAN based networks.
Almost all Internet traffic generates and terminates with an Ethernet
connection, which has been adapted for higher speeds and the volume of
traffic over the Internet. Ethernet technology is rapidly changing and
the new 10 Gigabit technologies maintains certain older characteristics
like package format and at the same time adapts itself to faster
connections, higher speeds and more efficient handling of internet
traffic. The newer versions of Ethernet like the Gigabit Ethernet and
10 Gigabit technologies have wide-ranging applicability in LAN, WAN,
MAN and SWAN (Axelson, 2003). We will discuss these points in greater
detail in the essay beginning our discussion with a brief description
of Ethernet technology, its history, modes of operations, Gigabit
architecture and applications and the rise of the 10 Gigabit technology
in Ethernet applications. In our final analysis we will analyse the
uses of Gigabit technology and the advantages involved. We would also
discuss the market demands as far as setting up the new technology in
business and enterprises is concerned and conclude with a final short
summary giving the prospects of Ethernet and Gigabit Ethernet in this
information savvy New Age.
Ethernet Technology:
Ethernet tends to connect computers using hardware that could be used
on different linking machines and workstations. It differs from the
Internet because the Internet connects remotely located computers with
a telephone or cable line or a software system. Ethernet uses software
but connects in LAN on the basis of hardware and the Ethernet patent is
referred to as the Multipoint data communication system with collision
detection. Ethernet was the first technology that could connect through
LAN hundreds of computers located in the same region or building
(Spurgeon, 2000). Ethernet technology can be used on optical fibre and
twisted cables and it operates on a half duplex and in recent versions,
full duplex mode. There are three major varieties of Ethernet and these
are Gigabit Ethernet, 10Base T Ethernet and Fast Ethernet that supports
a data transfer rate of 100mbps. Ethernet uses a star or bus topology
and supports a data transfer rate of 10 mbps. The CSMA/CD protocols are
used to handle simultaneous demands and workloads. The newest Gigabit
version supports a transfer rate of 1000 megabits per second and thus
has a speed of a 1000mbps. Ethernet is the most widely used LAN
standards. Ethernet specifications are used in IEEE 802.3 standard
software and specify physical software layers within the system.
Ethernet is used in nearly 90% of LAN workstations and LAN-connected
PCs in the world (Lee and Lee 2002). Although Ethernet is the most
popular physical layer of LAN technology that is in use, the other
common LAN types include the Token ring, Fibre Distributed Data
Interface (FDDI), Local Talk, Asynchronous Transfer Mode or ATM, Fast
Ethernet, a more advanced variety of Ethernet and Gigabit technology.
Ethernet helps in giving a high-speed connection along with the ease
and convenience of installation and maintenance. Ethernet applications
and installation are comparatively inexpensive and it has wide
applicability and compatibility with LAN and other widely used networks
but LAN is the most popular compatible networking. Ethernet is
acceptable in the computer market and supports all popular network
protocols making Ethernet the ideal networking technology for all
computers used at present. The Ethernet standard is set at IEEE 802.3,
deriving its name from the Institute of Electrical and Electronic
Engineers (IEEE), which defines this, standard. An adherence to the
standards can ensure the efficient communication and links between the
networking equipments and the networking protocols. The IEEE standard
802.3 defines the rules for configuring an Ethernet network and also
specifies how the elements in an Ethernet network can actually interact
with each other.
Brief History about Ethernet:
Ethernet originated nearly 30 years ago when Xerox corporation used an
experimental coaxial cable network with an original transfer rate of
just 3 mbps using a CSMA/CD (carrier sense multiple access/collision
detect protocol that could be used in then existing LAN and that could
support some heavy traffic. This experimentation and development of the
first Ethernet technology received some attention and around 1976-1980,
DEC (Digital Equipment Corporation) and Intel joined Xerox Corporation
to develop a LAN architecture that supported a data rate of 10 mbps.
This was the Ethernet version 1.0 specification and the original
version of IEEE 802.3 was also based on this initial Ethernet version
specification. The draft of IEEE 802.3 version received its approval in
1983 and was specified as an official standard by the IEEE in 1985.
Ethernet and IEEE 802.3 and all versions of these are ultimately
compatible, derived from and improved versions of this basic version of
original Ethernet technology. A number of supplemental versions and
technologically upgraded and advanced versions have been used to
support higher data transfer rates. Thus Ethernet technology saw a
gradual development with a gradual LAN technology development in the
1970s to the formation of IEEE subcommittee and approval of Ethernet
version IEEE 802.3 later declared as standard in the 1980s. This was
upgraded further with the development of Fast Ethernet and an
associated speed of 100 mbps in 1995 with the final 1000 mbps Gigabit
standards approved in the middle of June 1998.
Ethernet Explained:
A general description of what is Ethernet and how it works can be
given as Ethernet seems to be abased on the idea of a group of people
within the same location and on the same network sending messages
through a radio system and using a common wire or channel and this has
been referred to as ether, a supposed medium that 19th century
physicists believed helped in the propagation of light. Each member in
the system seems to have a 48-bit key MAC address assigned to the
network-interface card and this makes sure that all systems within the
Ethernet network have their own distinct addresses. As Ethernet is
found and available and compatible nearly everywhere with wide ranging
applicability, manufacturers tend to build the functionality of an
Ethernet card directly into PC motherboards, which are printed circuit
boards used in personal computers, also known as main boards.
In the initial stages of its development and discovery, Ethernet used a
shared coaxial cable, which was found in the building attached to every
connected machine brought within the system. All computers in a
building were connected to an attachment unit interface (AUI)
transceiver and this in turn connected to another cable. A simple
passive wire was used for small Ethernets and was considered not quite
applicable to large extended networks as any damage to the wire could
actually make the Ethernet segment dysfunctional. Within the Ethernet
technology, the systems are run in such a way that all communications
between PCs, happen along the same wire, and information transmitted by
one computer is received by all computers in the network even if the
information was meant to go only to one computer or specific
destination. The interface card of the network filters out all
information that is not originally addressed to the entire network and
the CPU is interrupted when applicable packets are received. This
information sharing by all computers is considered one of the weaker
points of the shared medium of Ethernet since any node within the
entire network connected by the Ethernet can dig out information that
seems to have been originally meant for a single location. Thus all
traffic within the single wire used for the technology, can be
intercepted and manipulated. A single cable thus has its ups and downs
because not only is there an information security risk with such a
networking, the bandwidth is also shared so after a power failure or a
restart, traffic can crawl to a point that it can cause a slowing down
of information transmission and data transfer due to too much crowding
within a single wire.
The computers with an Ethernet connection networking share the channel
using the CSMA/CD technology that stands for Carrier Sense Multiple
Access with Collision Detection. The CSMA was first developed and used
for ALOHAnet in Hawaii and is comparable to token Ring and master
controlled networks we have discussed here. An algorithm is generally
followed when the computers transmits data and there can be several
commands from start, transmitting, end transmission, wire busy, to wire
idle and transmission stopped or maximum transmission attempts
exceeded. In this way, a medium, the Ethernet is chosen through
computers try to transmit information. When there is too much traffic,
there is an apparent cessation of these attempts, which renews again
after a random period of time. This aid in avoiding collision and an
exponential increase of back-off times are used when there is more than
one failed attempt of transmission.
Depending on the type of medium used Ethernet segments can have
restricted size. A 10BASE5 coax cable can only have a maximum length of
500 metres. An Ethernet repeater can be used to have a bigger length
cable and this device uses the signal from one Ethernet cable and
repeats it into another cable. This way Ethernet repeaters can be used
to connect up to five different segments with a maximum three of these
devices used as attached devices. Since many segments are connected,
cable breakages can be handled more efficiently. Thus when any one of
the Ethernet segments breaks off, most of the devices are unable to
communicate as connections are broken, however Ethernet repeaters allow
the workings of different segments separately which can continue to
work despite the cable breakage in one part of the network (Quinn,
1997).
Ethernet segments can usually be terminated with a resistor at both its
ends within the network and as for the equipments used, each end of the
coaxial cable must have 50 ohm resistor and head sink and this is
called the terminator affixed to an N or BNC connector. When resistors
are not used, it is usually perceived as a cable breakage and
consequences will be such that the alternating current signal
transmitted will be reflected and as a reflected signal is
indistinguishable from a collision, similar results as in collision
will take place with complete cut-off of all communication between the
units. When Ethernet repeaters are used electrically different segments
can continue to function even in isolation as separate units and this
helps in regenerating and retiming the signals (Field, et al 2002).
Most repeaters have an auto-partition function and this means they are
capable of removing and isolating a segment from service when there are
too many collisions, breakages, and traffic or when the collisions last
to long. This is done to prevent the other segments from getting
affected by such collisions or breakages and once there is a renewal of
activities that are smooth and without collisions, the Ethernet
repeaters reconnect the unattached segment to the original network.
The advantages of using cables in the star Ethernet system has been
used effectively by net workers who create Ethernet repeaters with
multiple ports known as Ethernet hubs or fan outs that could be
connected to other hubs or coax backbones. The earlier hubs were also
known as multiport transceivers and DEC's DELNI is one such example of
an Ethernet fan out. Multiport transceivers allow the sharing of a
single transceiver by multiple hosts having AUI connections. This also
allows the working of a standalone single Ethernet segment that does
not essentially require a coax cable. Companies such as the DEC and
Synoptic sold many multiport transceivers that could connect many
10BASE-2 thin coaxial segments. However, coaxial Ethernet segments have
been made obsolete by the development of unshielded twisted pair cables
(UTP), which began with StarLAN and continued with 10BASE-T. The new
developed of the unshielded twisted pair variety allowed Cat-3/ Cat-5
cables and RJ45 telephone connectors to connect the fan outs to the
ends. This further helped to replace the coaxial and AUI cables.
Unshielded twisted pair Ethernet takes the termination problem into
consideration and every segment is taken separately so that the
termination could be built as a hardware component without requiring a
special and separate external resistor.
However despite the development of the star topology, Ethernet networks
use a half-duplex transmission access method and also use the CSMA/CD
with minimal cooperation from the hub that deals with packet
collisions. Every packet transmitted is sent to every port and node of
the hub so what really remains are bandwidth and security problems due
to this open information sharing, so to speak. Since the chance of
collisions is proportional to the number of transmitters and the data
that have to be transmitted, implying that the larger the amount of
data sent, the higher are the chances of packet collisions.
Modes of Operation:
The Ethernet network elements have interconnecting media and network
nodes which basically falls into two major classes known as the Data
terminal equipment (DTE) which are the PCs, workstations, servers, and
are usually the destination or source of data frames often referred to
as end stations; the other category of network node is the Data
communication equipment (DCE) that are the standalone intermediate
devices such as the repeaters, interface cards, modems, switchers or
routers (Hancock, 1988). The DCEs are intermediate network devices that
receive and forward data frames across the entire network.
The traditional Ethernet and IEEE 802.3 variety works in a half-duplex
mode. For the other mode of operation as in Full-duplex CSMA/CD is not
used. Auto-negotiation is another mode of operation for the Ethernet.
Half-Duplex Mode of Transmission/ CSMA/CD Access Method:
The Half-Duplex mode of transmission was developed in the original
version of IEEE 802.3 Ethernet application and the CSMA/CD is
considered as a means by which two or more stations could share a
common media in an environment that is switch less and does not require
arbitration, or assigned time slots to indicate when workstations are
ready to transmit information. This means that individual Ethernet MAC
is capable of determining when it can send a frame.
The CSMA/CD access rules are given in the full form of the CSMA/CD
protocol as it involves Carrier sense multiple access and Collision
Detect. Carrier sense implies that each station continuously keeps
alert for traffic on the medium to determine when gaps between frame
transmissions are seen. ‘Multiple access’ refers to stations that may
begin transmission any time they detect that there is no traffic and
the network is relatively quiet. Collision detect acronym implies that
when there are two or more stations in the CSMA/CD network and if they
begin transmitting data at the same instance the streams of bits from
each of the workstations will collide with each other and both the
transmissions which collide will then become unreadable (Halabi and
McPherson, 2000). Each transmitting station should be able to determine
and detect that a collision has taken place before the station has
finished sending its frame. Each transmitting station then must stop
transmitting any further data as soon as a collision is detected and
then this workstation must show a state of abeyance in its activity for
a random length of time determined by a back-off algorithm before
attempting to retransmit the frame all over again.
In certain situations when two distant stations on the network need to
send a frame and the second station does not even begin transmitting
until just before frame of the first station arrives. In that case, the
collision is detected immediately by the second station yet the first
station does not detect it until the corrupted frame signal goes all
the way to the station. The time required to detect a collision can
reach a twice the time required for the signal propagation between two
distant stations on the network situated at farthest points (Quinn,
1997). This suggests that the maximum collision diameter and minimum
frame length are directly related to the slot time and longer minimum
frame lengths translate to longer slot times and larger collision
diameters where the shorter the minimum frame lengths are they
correspond to shorter slot times and smaller collision diameters.
There was growing understanding and need to reduce the impact of
collision recovery and the need for network diameters to be large
enough to accommodate networks, which are large, sized. The general
consensus was to choose the maximum network diameter of around 2500
metres, and to set the minimum frame length to ensure that all
potentially damaging collisions are reported quickly. This system
worked for 10mbps speed of Ethernet but for higher transfer rate
Ethernet connections and developers such as the Fast Ethernet, a
backward compatibility with earlier Ethernet networks were necessary
and this involved the inclusion of the existing IEEE 802.3 frame format
and error detection procedures as also all the networking software and
applications which could run on the 10mbps networks (Quinn, 1997). For
all transmission rates, the time required to transmit a frame is
inversely related the transmission rate and at 100 mbps, a minimum
length frame could be transmitted at one tenth of the defined and
original slot time and so any collision that might occur at this time
would go undetected. The maximum network diameters used and specified
for 10 mbps could not be used for 1000mbps networks. Fast Ethernet
networks and connections help in reducing the maximum network diameter
by more than 200 meters. This problem is also seen in the Gigabit
Ethernet, as there is a decrease in network diameters by a factor of 10
to more than 20 meters for 1000 mbps operations. This is however a
potential hindrance although as a solution the same maximum collision
domain diameters were used to increase the minimum frame size adding an
extension field to frames shorter than the minimum lengths.
Full-Duplex Mode:
The full-duplex mode allows the simultaneous two way transmission of
data along the same or one link. The full-Duplex mode is an MAC
capability that allows the two-way transmission of information over
point-to-point links. This Full-duplex transmission is functionally
much simpler than half duplex transmissions as no collisions, media
contentions, schedule retransmissions and extension bits on short
frames ends are involved. Since there is a reduction is the procedural
complications, more time is available for transmission and there is
also a doubling of link bandwidth and each width is capable of
supporting full-rate, simultaneous two way transmission of data frames
and each transmission begins as soon as frames are ready to send. The
main restriction is inter-frame gap between two successive frames that
has to be of a minimum specified length. Usually all frames conform to
the Ethernet standard frame formats.
Ethernet Types:
Apart from the broad varieties of Ethernet types that transmit data at
10mbps (Ethernet), 100mbps (Fast Ethernet) and 1000 mbps (Gigabit
Ethernet), 10 Gigabit Ethernet and the earlier varieties such as
StarLAN, there can be variations in the Ethernet frame types as well.
We give a brief description of Ethernet types below.
The earliest and first varieties of Ethernet was the Xerox Ethernet
which was the original 3m bit per second Ethernet implementation and
had versions 1 and 2(Hall, 2003).The framing format version 2 is still
in use.
The other earlier Ethernet application 10BROAD36 is now no longer used
although it was one of the earlier standards supporting Ethernet over
long ranges. The broadband modulation techniques are similar as found
in cable modems and systems operated on coaxial cables.
StarLAN or 1BASE5 as the first Ethernet implementation on twisted
paired wiring and operated at 1mbps speed. This was gradually replaced
by the other versions.
Among the more recent versions of the 10 mbps Ethernet is the 10BASE5
that uses thick net coaxial cables also called thick wire or yellow
cable is the original implementation of the 10mbps Ethernet.
Transceivers could be connected using a vampire tap and connecting the
core and the screen with N connectors. There is a cable that could be
used to connect the transceiver to the AUI or Attachment unit
Interface. This type of Ethernet can have 5 network segments with 4
repeaters, with three of the segments that could be connected to the
network. The bus topology is used here and the maximum segment length
is 500 meters with the overall length at 2500 metres. The minimum
length between nodes is stipulated at 2.5 metres with the maximum
number of nodes per segment at 100. This system is obsolete as of now.
10BASE2 uses the thin net coaxial cable, also called the Thin wire or
Cheaper net, a BNC connector and bus topology with a terminator at the
end of each cable. The cable specified for its purposes is RG-58 A/U or
RG-58C/U with a 50 ohms resistance. the 5-4-3 rule is applicable here
meaning like the 10BASE5 Ethernet there are 5 network segments, 4
repeaters and 3 of these could be connected to computers. 185-200
metres is the maximum length for each segment and each machine uses a
T-adaptor to connect with a BNC connector. Although signal quality is
considerably reduced with each barrel connector, barrel connectors
could be added to link the smaller cable pieces in the network. Length
between nodes is given at a minimum of 0.5 meters. This is one of the
more widely used Ethernet applications.
The StarLAN which was the first Ethernet twisted wire implementation
later evolved into 10BASETwhich comprises of 4 wires and two twisted
pairs with a Cat-3 or Cat-5 cable of up to 100 meters in length. At the
middle of the connection is a hub or switch, which has a port for each
node. The 10BASET uses star topology and as we mentioned 2 pairs of
unshielded twisted wires. This category of the Ethernet is not subject
to the 5-4-3 rule and can use 3, 4 or 5 cables with the best
performance given by category 5 cable. The maximum segment length here
is 100 meters. Maximum number of connected segments can be nearly a
thousand. The minimum length between nodes is 2.5 meters. Whereas only
1 node can be present in every segment using the star topology, this
system uses RJ-45 connectors.
10BASEF is the generic name for 10 Mbits/s Ethernet standards using
fibre optic cable extending up to 2 kms in length. The number of
network nodes can be 1024 with a maximum segment length of 200 meters.
Specialized connectors for fibre optic cables are used here. This has
three main varieties: the 10BASEFL is an updated and more advanced
version of the FOIRL standard. FOIRL is the Fibre-optic inter-repeater
link and the original standard for Ethernet on fibre optic. This
Ethernet type is used to connect computers in a LAN setting a task not
done mainly due to the incurring costs. This is the most widely used of
the 10BASEF Ethernet type. The 10BASEFB that has never been used serves
as a backbone between hubs. The 10BASEFP is a star network that does
not require a repeater for its connections connect a number of
computers with hubs and switches and gets cable distances up to 500
metres.
The 100BASET is also known as the Fast Ethernet and is a generic term
for any of the three standard varieties if Ethernet that transfers data
at 100mbits/s over twisted cables up to 100 meters long. These three
varieties include 100-BASETX, 100BASE-T4 and 100-BASET2. Fast Ethernet
uses RJ-45 connectors and star topology. The CSMA/CD media access is
used here. The minimum length between nodes is specified at 2.5 meters.
With 1024 maximum number of connected segments possible, the IEEE802.3
specification is used. The 100BASETX is a star shaped configuration
similar to 10BASE T and uses two pairs of wires and Cat-5 cable to
achieve the 100mbit/s speed. 100BaseTX requires category 5 two pair
cables and the maximum distance of these is given at 100 meters. The
100BaseT4 requires a category 3 cable with 4 pair and maximum distance
of these cables is 100 meters. This system uses a Cat-3 cabling and
uses all four pairs in the cable limited to half-duplex access methods.
As now the popular Ethernet cabling is the cat-5 cabling, his is now
considered obsolete. The final variety 100BASEFX is a 100mbit/s
Ethernet type that uses a multimode fibre. The maximum length specified
here is 400 metres for half-duplex connections and 2 kms for full
duplex connections to ensure that all collisions are detected.
100BASEFX can thus use fibre optic to transmit up to 2000 meters and
for this purpose it requires two strands of fibre optic cable.
100VG LAN is also another Ethernet variety that uses star topology, a
series of interlinked hubs and RJ-45 connectors. In addition to
Ethernet packets this implementation also supports the Token ring
packets and has an IEEE 802.12 specification. It requires 4 pairs of
category 3 cable wires and the maximum distance covered is 100 metres
(Held, 1996). However with a category 5 cable 150 metres could be
reached. Additionally the fibre optic can be used to transmit data up
to 2000 metres.
We next turn to Gigabit Ethernet Technology, which being a new and
improved version of Ethernet altogether requires a separate section.
Gigabit Ethernet Technology:
Also known in its abbreviated form GbE, the Gigabit Ethernet technology
describes the implementation of Ethernet networking and transmitting at
a speed of one or more than one Gigabit per second. Gigabit Ethernet is
supported with the use of optical fibre and twisted pair cables and the
physical layer standards in this category includes 1000BASET. 1 Gbps is
used over a category 5 cable with copper cabling and 1000BASE SX is
used to attain for short to medium distances over fibre. The first
Gigabit Ethernet standard was set at the IEEE 802.3 standard in 1998.
Its relatively recent launch suggests that the Gigabit Ethernet is the
latest version of the Ethernet, which is the most popular and widely
used computing network worldwide (Kadambi, 1998). The Gigabit speed of
1000 mbps of raw bandwidth is 100 times faster than the original
Ethernet version of 10mbps and its greatest advantages lies in the fact
that it is compatible with existing Ethernets and uses the existing
CSMA/CD and MAC protocols. Gigabit Ethernet competes directly with ATM
as far as market competition is concerned. It is deployed in high
capacity backbone network links and for small installations Gigabit
speed is not yet necessary. Gigabit Ethernet has been used in desktop
technology in apple computers, Power MacG5, Apple's power notebook and
is also being built into Pentium boards. One of its desktop features
includes professional video editing. Gigabit Ethernet has been
outsmarted by the 10gigabit Ethernet technology, which is the fastest
Ethernet standard that became fully operational in 2002.
Introduction to Gigabit Ethernet
Our discussion on Gigabit Ethernet began by providing a brief
description of the Gigabit technologies and the main characteristic of
this upgraded Ethernet variety. The transceiver used for Gigabit
Ethernet is the GBIC also known as the Gigabit Interface Converter. The
GBIC measures 8.5 mm by 13.4 mm and has a depth of about 50mm. A hot
swap standard electrical interface of a one Gigabit Ethernet port can
support any physical media including copper to 100 km of single mode
fibre. The standard Gigabit Ethernet system operates at 1000 Mbps speed
of transfer of information (Norris, 2002). The 802.3z is the standard
that describes the specifications for fibre optics for the 1000BASE-X
Gigabit Ethernet system. The 802.3ab standard describes specifications
for the category 1000BASE-T twisted pair Gigabit Ethernet system. This
rate the major two varieties of Gigabit Ethernet technologies used. The
10 Gigabit Ethernet is a further up gradation and has speeds reaching
10000 Mbps.
According to Frazier, the chair of the Gigabit Task Force developing
the Gigabit Ethernet technology has been challenging and hard work
mainly because this new technology aimed at developing a standard that
scales the operation of Ethernet networks to 1000 Mbps while retaining
the known characteristics that are compatible with Ethernet and that
have made Ethernet the dominant-local area network (LAN) technology.
The Draft D3.1 of the P802.3z was approved in a letter ballot of the
IEEE 802.3 working group and this was only possible when an approval
rate of 75% was obtained from the working group. The protocol layers in
802.3 standards are the areas that have been developed and modified in
802.3z version (Riley and Brever, 1998). The 100BASE SX and 1000BASE LX
fibre optic transceiver specifications have already been highlighted
here.
The 1000BASE-SX specifications for short wavelength laser
transceivers uses 62.5 micron fibre and supports multimode fibre optic
links of up to 260 meters. The 50 micron fibre can be used to support
multimode fibre optic links of up to 550 metres. The 1000BASE LX
supports installations at longer distances and uses higher cost
components with 62.5 micron fibre used for 440 meters and 50 micron
fibre on 550 meters. On a single mode fibre, up to 3 kms of fibre optic
links are possible. One of the latest technologies 1000BASE-CX that
supports copper cabling links of 25 metres, is included in the 802.3z
specification for transceiver technology.
Frasier goes on to describe that the Gigabit Ethernet encompasses the
new full-duplex Media Access Control (MAC) and the Carrier Sense
Multiple Access with Collision Detection (CSMA/CD) MAC. The full duplex
operation takes advantage of the contention free access and flexible
topologies and the 802.3z used the CSMA/CD MAC to work at an extended
range of 1000Mbps. This was done with the aid of a technique known as
'carrier extension' which was added to the CSMA/CD to overcome certain
limitations of the algorithm that was based on the rule that the round
trip messaging time between two workstations could not be higher than
that required to transmit the smallest frame. To improve the throughput
of Gigabit CSMA/CD LANs, frame bursting was introduced as an optional
feature.
In the early stages of deployment of the Gigabit Ethernet it is mainly
being used to interconnect high performance switches, routers and
servers in LAN backbones. The full duplex operating mode is suited for
this type for application environment and is favoured over the CSMA/CD
half duplex mode. This also ensures that high performance desktop
computers can take advantage of high peak bandwidths of repeating hubs
that are cost effective. Frasier's claim also points to the fact that
gigabit technology may be more compatible with full duplex rather than
half duplex modes of Ethernet operation. The 10 Gigabit Ethernet is a
full-duplex technology and the half-duplex access mode cannot be used
for its purposes.
The Gigabit Media Independent Interface (GMII) allows MAC and PCS
implementations to interoperate and provides the starting point for a
future possibility of attaching 1000BASE - T PHY to the 802.3z MAC.
Gigabit Architecture:
In its architectural features, the Gigabit Ethernet is to an extent
similar to the original Ethernet architectural model. At the Layer 1,
an Ethernet physical layer device PHY connects the optical and copper
media to the MAC layer with the help of a connective technology. In
Ethernet architecture this physical layer is further divided into three
sub layers, Physical Medium Dependent (PMD), Physical Medium Attachment
(PMA) and Physical Coding Sublayer (PCS). PMDs or physical medium
dependents are optical transceivers and other such similar devices that
provide physical connection and signalling to the medium. The PCS
includes a serializer or a multiplexor and consist of coding. The IEEE
802.3ae defines two PHY types namely the WAN PHY and the LAN PHY.
Although both these physical layer devices have similar functionalities
and features, the WAN PHY has an extended feature that enables
connectivity with a variety of networks including the SONET-STS 192c
and the SHD VC-4-64c.The architectural model given specifically for the
Gigabit Ethernet includes the Media Access Control parameters (MAC)
specifications of the IEEE, the physical layers as we have discussed,
the repeaters and the management parameters for successful operations
of Ethernet at speeds of 1000 Mbps.
Advantages of Gigabit Ethernet:
One of the main advantages of the Gigabit Ethernet is the speed of data
transfer, which is several times faster than the traditional Ethernet
standards. The Gigabit variety 10GBASE-T has chips capable of
auto-negotiating to lower speeds and eventually it will utilize just
one chip available for NICs. This one chip usage to perform several
functions is one the other advantages of this type of Ethernet. With a
Gigabit Ethernet, the bandwidth is able to move around a large amount
of data. With the inclusion of a Gigabit switch hub or a twisted RJ 45
cable, data could be transferred faster within the LAN (local area
network) that connects the different computers. The advantages of using
a gigabit connection is that in the transfer of large files differences
are conspicuously noticed although in case of smaller files, the
transfer rate might not be as fast as expected. Since the Gigabit
Ethernet is an application that is a up gradation of the original
Ethernet technology, it increases established Ethernet speeds by a
factor of 10. Yet the Gigabit technology quickly won approval, as no
new technological knowledge was necessary for its workings and
administration. As the Gigabit Ethernet uses the Category-5 cabling,
this means that no system up gradation or migration to new standards
were necessary. Gigabit Ethernet used existing cabling system and was
perfectly compatible with the existing physical networking between
sites.
There are many advantageous features of the Gigabit Ethernet. It is
nearly 100 times faster than traditional Ethernet and 10 times faster
than the Fast Ethernet. Its maintenance and implementation costs are
very low and decreasing even further quite steadily. Gigabit Ethernet
can ensure wide support for the expanding portfolio of LAN
applications, it has a full duplex access mode allowing a two-way
transfer of information and doubling the rate of data transfer
(Seifert, 1998). Gigabit systems have higher networking bandwidth and
superior scalability. It has an excellent deployment capability on
existing systems using the category-5 cabling and its main points are
Network reliability, Troubleshooting capability and Scalability.
Gigabit Ethernet is easy to install and enhances network efficiency and
speed. The familiarity with Ethernet technologies allows the
maintenance of Gigabit Ethernet to be quite hassle-free as new
technological skills are not required. Ethernet management tools SNMP
are easily available and can be used on the Gigabit systems. The
Gigabit and Fast Ethernet standards are scalable given the fact that
all these systems are compatible in a particular working LAN set-up.
GIGABIT APPLICATIONS :
GbE is the first large-scale application that uses lasers to launch
light signals in multimode optical fibre. Applications of high speed
Gigabit Ethernet are found in the installed bases of Ethernet hubs,
switches and routers (Perlman, 1999). The requirements for higher
bandwidth met by the gigabit Ethernet find its applications in storage,
back-ups and video. Gigabit applications can also be found in
high-performance cluster computing, server attachment and storage
interconnect. Cluster computing finds its applications in Ethernet
switches and are found in NAS (network attached storage) devices. The
bandwidth intensive applications of the Gigabit Ethernet extend to
Video on demand (VoD) and teleradiology. Among the important Gigabit
applications are 100 port Hybrid switches and a Gigabit switches with
ATM/WAN uplinks.
Gigabit Ethernet Standards are explained by a diagram below: Courtesy: Acer
GIGABIT Compared To ATM :
Gigabit and ATMs are both applicable in providing higher bandwidth. ATM
can help in the up gradation of 25mbps at the desktop to 622 Mbps at
the core. This is completely upgraded advanced technology and is
available easily. The bandwidth given by ATM is at 2.4 Gbps given
through the OC-48 and this became functional by the end of 1998.
Ethernet variations se 10 Mbps at the desktop and 100 Mbps at the core
and there are varieties available like the Fast Ether channel that has
helped scale the core bandwidth and enabled an up gradation to Gigabit
Ethernet. Characteristic so both ATM and Gigabit Ethernet is that they
allow a higher bandwidth and switching capacity. Gigabit Ethernet
allows compatibility with existing desktops and networking protocols.
So existing PCs, monitors, servers, mainframes, cabling plants are all
retained as the Gigabit Ethernet could be implemented within the old
and already existing infrastructure. This was important as millions
usually go into the set up of a networking infrastructure by large
companies. For a smooth and easy migration to the Gigabit standards,
existing LAN protocols had to be compatible with the new applications
and that was a main advantage and major challenge for the Gigabit work
force (Cheng et al 2005). ATM also faces a similar challenge. ATM
guarantees the quality of service (QoS)within the backbone and over the
WAN technology by using mechanisms such as the available bit rate (ABR)
, the constant bit rate (CBR), the variable bit rate (VBR) and the
unspecified bit rate (UBR). Ethernet also promises to provide a class
of service (CoS) and in this way gigabit and Ethernet both solve the
similar problems of applicability. ATM as with Gigabit has been
utilised recently to build backbones also at a low cost. The benefits
of ATM are seen in the wide area network and WAN interaction has been
very effective in scaling campus networks. Data types such as voice and
video have also been integrated into the system and services
integration in WAN systems have been instrumental in reductions of
costs of installation and maintenance as with Gigabit applications.
ATMs could be used on LANs and WANs in an equally efficient way whereas
the original Ethernet versions were more suited for LAN. Ethernet
systems have tried to include ATM features such as Resource reservation
protocol (RSVP) and RTSP (Real-time streaming transport protocol). On
the other hand ATM has tried to incorporate some of Ethernet
functionalities like LANE (LAN emulation) and IPOA (IP over ATM). ATM
has an edge over Ethernet as it is also installed in several
workstations and offers the QoS, quality of service. It is better
suited for video applications although its speed is only about 622 Mbps
whereas the Gigabit Ethernet has a speed of about 1000Mbps. One of the
main advantages of the gigabit Ethernet over ATM is that whereas ATM up
gradations may find previous applications and structures obsolete, up
gradation of Gigabit Ethernet can be done effortlessly without
implementing any changes in the networking protocols.
10 Gigabit Ethernet Technology:
The 10 gigabit Ethernet operation is as of yet the highest speed of
Ethernet operation and has been formally ratified by the IEEE in June
2002. 10 Gigabit Ethernet has been standardised as IEEE 802.3a and the
associated telecommunications technology offers a speed of data
transfer of ten billion bits per second. It is an advanced version of
the original Ethernet standard, which is the current technology, used
in LAN networks in most computers around the world. Ethernet technology
has the features of being least expensive and most efficient when it
comes to features of handling data movement and transfer. It also
provides a consistent end-to-end connection between networks using
optical fibre connections. 10-gigabit Ethernet technology can finally
replace connections using the ATM and this mainly because of the speed
obtained by the 10-gigabit implementation. Switches, routers,
multiplexors, can be used now with simpler 10 Gigabit Ethernet switches
and data rate can improve from a mere 2.5 Gbps to 10 Gbps. 10 Gigabit
Ethernet interconnects LANs (local area network), WANs (wide area
network) and metropolitan area networks (MAN) and is thus quite an
improvement over the original LAN only Ethernet application. 10 gigabit
Ethernet also uses the IEEE 802.3 Ethernet Media Access Control (MAC)
and its frame format and also the size remains the same. A full-duplex
transmission is used here so data transfer rates are doubled here
supporting distances of about 300 metres. With single mode fibre
optics, distances of up to 40 kms can be supported and smaller networks
can be integrated into larger Gigabit networks.
The 10 Gigabit Ethernet is thus an extension of the IEEE 802.3ae
standard protocols and has a data speed of 10 Gbps but as we discussed,
goes a step ahead of Ethernet in using WAN compatible applications.
Along with these specifications, the Gigabit Ethernet provides an
increase in bandwidth while maintaining maximum compatibility with
installed 802.3 standard bases and infrastructure networking systems.
The existing principles of networking, installation and management are
all retained making up gradation easy and hassle free. this helps in
protecting investment in infrastructure, and in research and learning
of Ethernet applications. The main Ethernet architecture is retained
here including the MAC (media access Control), maximum and minimum
frame size and frame formats. Along with this certain upgrading involve
the inclusion of several physical coding sub layers such as the 10GBASE
X, 10GBASE R and 10GBASE-W as also additional supporting material such
as the 10 Gigabit Media Independent Interface (XGMII), 10 Gigabit
Attachment Unit Interface (XAUI) and 10 gigabit Sixteen bit Interface
(XSBI).
Why Do We Need 10 Gigabit Ethernet?:
The 10 Gigabit Ethernet has a performance level of 10 times the
standard Gigabit Ethernet technology and with the addition of the 10
Gigabit Ethernet, LAN networks can reach greater distances and support
higher bandwidth applications. 10 Gigabit Ethernet applications can be
used as low cost and highly efficient system of Ethernet technology
that can be deployed and made compatible with any existing LAN
environments. This is a very attractive option to companies and
enterprises that have invested heavily on equipment, processes,
training and cabling using the original Ethernet technology and yet
getting a higher speed of data transfer with zero added costs of
maintenance of infrastructural changes. The 10 Gigabit Ethernet
protocol also gives flexibility in network designs with server,
switches and router connections (Saunders 1998). Multiple vendor
sourcing of standard base products of Ethernet provides
interoperability between systems and new products do not have to be
installed. 10 Gigabit Ethernet is ready to replace other smaller
proprietary technologies for server and storage area networks. It has
several advantages as it offers the higher necessary bandwidth, it
offers a cost saving server consolidation and it has a future
possibility of offering a planned growth of 10 Gigabit network support
features making 10 Gigabit Ethernet the most effective in future
implementation of networks.
What Is 10 Gigabit Ethernet?:
The 10 Gigabit Ethernet being the advanced and upgraded version of
Ethernet offers data transfer speed s 10 times that of Gigabit Ethernet
technology and can be used in LAN , local area network, WANs, wide area
networks and MAN, metropolitan area networks. This 10GBASE Gigabit
Ethernet system operates in Full-duplex mode only over fibre optic
medium. However there are several media over which it can work making
it easily usable in the LAN, WAN and the MAN varieties. The packet
format of the Ethernet is retained here although 10 Gigabit Ethernet
uses the MAC (Media access Control) protocol and does not require the
CSMA/CD (carrier sense multiple access with collision detection)
protocol, which is generally used in other Ethernet technologies
(Field, 2002). The 10 Gigabit Ethernet is capable of reaching new
distances with LAN and have higher bandwidth and the fastest speed in
Ethernet operations. These features make it unique although most of its
features, implementation techniques and maintenance principles are
similar to other Ethernet technologies, a fact which has made it highly
popular since its introduction.
10 Gigabit Ethernet Architecture:
The 10 Gigabit Ethernet protocol is a supplement and extension to
the 802.3 standard and is contained in the IEEE 802.3ae with MAC
specifications and an operating speed of 10 Gbps. The data rate is
10Gbps although slower data rates are also sometimes accommodated
through the WAN interface sub layer (WIS) and this allows the 10
Gigabit Ethernet equipment to be compatible with Synchronous Optical
Network (SONET) STS-192c transmission format. The 10GBASE-SR and
10GBASE-SW media types are used for shorter wavelength multimode fibre
covering a distance of up to 300 metres. For long distance data
communications the 10GBASESW media type connects to SONET equipment.
The 10 Gigabit Ethernet is similar to the original Ethernet model in
most ways with the only difference that the MAC rather than the CSMA/CD
protocol is used.
The architectural components of the 10 Gigabit Ethernet with specification IEEE 802.3ae is given as follows:
Benefits of 10 Gigabit Ethernet:
The 10 Gigabit Ethernet is usually found in the form of the Intel®
PRO/10GbE SR and Intel® PRO10GbE LR Server Adapters. 10 Gigabit
Ethernet has the power to deliver 10GbE performance reaching to a
distance of 10 kms between ports.
The benefits of a 10 Gigabit technology are many. With Internet traffic
tripling by the day and large files being transferred there is an
increased need to get information transferred faster and 10 Gigabit
Ethernet is the answer to maximising networking speed (Burg, 2001). Any
collaborative LAN, WAN or MAN environment that has many workstations to
share, create and manipulate huge files usually send a lot of traffic
over the system and these applications can include high performance
scientific computing, remote medical imaging and diagnostic, silicon
manufacturing tests, global weather monitoring and data, large database
mining and storage back-up, film animations and special effects and
normal desktop operations suited to user demands (Field, 2002).
The major benefits of the 10GbE apart from its very high speed of data
transfer are its low costs in implementation, ease of installation and
compatibility with older existing systems in infrastructural and
networking requirements. There is a low maintenance cost and an
efficient method to move around information and large data files within
a very short period of time.
Applications of 10 Gigabit Ethernet:
The several applications of the 10 Gigabit Ethernet technology can be
seen in Fabric Connect, in Local Area Networks, in Wide Area Networks
and in Metropolitan Area Networks. The Key application areas of the 10
GbE are given here taking practical examples of the 10 Gigabit Ethernet
applications as developed by Intel:
Fabric Interconnect: Fabric interconnect has small user bases
compared with the Ethernet. 10GbE offers the performance capabilities
to provide a cost effective fabric interconnect for storage and server
area networks an area dominated by proprietary solutions. These are
server area networks including the Infiniband, Servernet, Wultkit,
Myranet, and Quadrics technologies that offer excellent bandwidth and
latency performances. With the exception of Infiniband , all other
varieties are proprietary networks the deployment and maintenance of
which are difficult due to the small number of professionals familiar
with this technology. There are also higher costs for server adapters
and switches. The major disadvantage of these proprietary technologies
is that they are not interoperable without the installation of
appropriate routers and switches (Ferrero, 1999). The new Intel
PRO/10GbE SR server adapter with a 300m reach in multimode fibre is
considered to offer a cost -effective alternative that does not require
the specialized maintenance and management skills required for
proprietary interconnects.
Local Area Networks (LAN): For high performance LAN environments, the
10 Gigabit Ethernet technology is the most deployed technology. Like
the Gigabit Ethernet, the 10Gigabit Ethernet supports both single mode
and multimode fibre optic media. 10 Gigabit applications in LAN can
support higher bandwidths and reach greater distances. The 10 Gigabit
Ethernet within LAN settings can reach a distance of up to 40 kms so
companies are allowed to choose their networking settings so that the
data centre and servers can be as far as 40 kms away from their
campuses. So multiple campus locations within a 40 kms range is
possible creating a possibility of data sharing within a company over
widely separated workstations. Switch to switch, switch to server,
between buildings and data centres are all supported by a 10 Gigabit
Ethernet LAN environment. Within the widely used Intel variety, the
Intel PRO/10GbE SR Server Adapter is considered as effective in
creating high performance, high-throughput LAN backbones (Intel, 2004).
The 10GbE offers a cost effective solution by replacing multiple
adapters and links with a single PRO/10 GbE SR Server Adapter. For a
larger campus needs, the 10 km reach of the Intel Server adapter allows
for servers to be placed at cost effective locations reaching multiple
campus buildings at the same time.
Metropolitan Area Networks (MAN): The 10 gigabit Ethernet application
is already being deployed as a backbone technology for dark fibre
metropolitan networks. With the 10 Gigabit Ethernet interfaces, optical
transceivers and single mode fibre optics, network and Internet
services providers are able to connect 40 kms of a single metropolitan
area and create a citywide network. 10 Gigabit Ethernet allows a cost
effective high-speed infrastructure to be built for Network attached
storage (NAS) and storage area networks (SAN) and have superior data
carrying capacity at latencies similar to storage networking
technologies like High performance Parallel interface, Fibre Channel or
ATM-OC3. The applications developed by Intel for this includes the
Intel PRO/10GbE LR Server Adapter and its 10 km reach that allows
multiple LAN communities in a metropolitan area to be serviced and
connected by a single server farm. Furthermore, with multiple 10GbE
adapters, MANs can have a 20 kms diameter.
Wide Area Networks (WAN): The 10 Gigabit Ethernet allows Internet
service providers (ISPs) and network service providers (NSPs) to create
very high-speed links between carrier class switches and routers and
optical equipment attached to the SONET/SDH. With WAN PHY, the 10
Gigabit Ethernet allows a construction of WANs that connect
geographically distributed LANs found between campuses over existing
SONET/SDH networks. 10 Gigabit Ethernet links between a service provide
switch and a line termination equipment can be less than 300 metres.
10GbE technology is capable of producing high performance points of
presence (POP) in WANs and has a potential for global reach connecting
networks beyond just within a metropolis. The throughput contributions
of 10GbE for this application has been successfully demonstrated in an
Intel study with a WAN extension from Sunnyvale to Chicago to
Switzerland as detailed in the Los Alamos National Lab case study.
Market Review and Conclusion
The 10 Gigabit Ethernet technology, the latest developed technology
is highly customer and market friendly. The 10 Gigabit Ethernet is
compatible with existing Ethernet supportive environments and is
interoperable with other networks. This new technology has a
comparatively low maintenance cost and low cost of ownership as well.
According to markets analysts, new networking technologies supporting
higher bandwidth and high speeds of data transfer as in case of the 10
GbE, indicate a future direction of a rapidly technologically
sophisticated world. For a high speed LAN technology, 'mainstream'
would refer to a volume-specific market for servers and desktops.
Ethernet has been the most widely deployed networking technologies and
its popularity continues into the 21st century as well being further
empowered by the Gigabit Ethernet technology. The fast Ethernet
connection operates at 10-100 million bits over second whereas the
Gigabit/ 10 Gigabit Ethernet operates at 1000-10000million bits per
second (Burg, 2001). Mainstream adoption plants use existing cable
plant and category-5 unshielded twisted pair (UTP) cable. The
incorporation of the Gigabit Medium Independent Interface (GMII) is one
of the primary features for LAN controller chips. The costs for
deployment are similar to that of Ethernet application and fast
Ethernet applications in the mid-1990s.
Although some market analysts suggest that the migration from slow to
Fast Ethernet and to Gigabit Ethernet has not been smooth. The market
is slow in catching up to technological progress and this is mainly
because of a suggested slow economy where IT budgets are tight and
taxes are high. However predictions are such that Gigabit and 10GbE
will make up 65% of the $15.2 billion Ethernet market in 2007. Prices
are already going down to encourage companies to upgrade their system
to a Gigabit or 10 Gigabit protocol. The market is showing a positive
trend already. According to 2004 InStat Market Research Report
Information reported a growth of the overall worldwide Ethernet switch
market in 2003 with total port shipments rising by 16% from 166.3
million ports shipped in 2002 to 193 million ports shipped in 2003. Due
to declining ASPs over the year, manufacturer's revenue declined by
11.8% from $13 billion in 2002 to $11.4 billion in 2003. The global
economy however has been looking up and according to InStat reports the
market growth in 2003 was driven by a sharp increase in the total
number of endpoints connected to the LAN. The traffic flow through LAN
has increased in diversity and many features such as network
convergence, heightened security awareness a threat of cyber terrorisms
and cyber attacks have become issues of major concern. However there is
an increasing need to replace aging and outdated equipments after three
years of strict IT budgets and a steady drop in ASPs across all
Ethernet switch products. According to industry pundits, these factors
along with the need for faster and more efficient connections drive
continued market growth and will continue to impact sales in the next
five years and the total shipments are expected to rise to 502.8million
ports by 2008 (InStat, 2004). These market trends are causing an
underlying technology shift in the Ethernet switch market and there is
an overall increase in deployments of Ethernet switches and Gigabit
Ethernet switches in company networks worldwide and LANs are generally
perceived to be intelligent, offering increased bandwidth and higher
efficiency a growing demand for an increasingly global economy and a
faster world reaching speeds of a billion bits of information transfer
per second.
The Gigabit Ethernet Alliance supports the implementation and
development of the Gigabit Ethernet and is largely responsible for its
patronage and popularity giving GbE a marketing and managerial edge
over ATM and other competing technologies.
Summary:
We began this essay with an analysis of Ethernet technologies, tracing
its history and development by the Xerox Corporation. The IEEE 802.3
was earliest standard Ethernet protocol that saw several up gradations
and variations since the 1980s when it was first developed by three
companies - Xerox, DEC and Intel. We then discussed the working of the
Ethernet technology across LAN when multiple workstations within a
building or specified location can access information. This was
explained using the functions of the underlying technology of the
traditional Ethernet - the half-duplex mode using CSMA with collision
detection. As we discussed further, we suggested that more advanced
versions of the Ethernet that could transmit data much faster are now
being used by several enterprises. The Gigabit Ethernet is capable of
transmitting information at the rate of a 1000Mbps and the more recent
version the 10Gigabit Ethernet has a data transfer rate at 10000 Mbps,
which is the fastest among the Ethernet operations. Along with an up
gradation in speed and efficiency we suggested that the Gigabit and
10Gigabit technologies are cost effective and easy to install, and
maintain. We finally tailed off our discussion with the applications of
10 Gigabit in LAN, WAN and MAN set ups implying its wide applicability
and connectivity across distant workstations. The final analysis was on
the market trends and we suggested that although the market is being
slow to catch up with a Gigabit technology due to low IT budgets, the
increasing needs of the global economy predicts a steady rise in
Ethernet and Gigabit Ethernet applications worldwide. This factor
coupled with decreasing prices of the Ethernet technology and an
increasing demand to transfer information across the globe faster, we
can conclusively predict that Ethernet technology is here to stay.
Glossary and Abbreviations List
ABR -Available bit rate
ATM - Asynchronous Transfer Mode is a switching technique in which information is organised in cells
AUI - Attachment Unit Interface
Bandwidth - the capacity of information a communication channel is capable of carrying
BNC connector -Bayonet Neill Concelman connector, also called a British
Naval Connector or Bayonet Nut Connector, a type of connector used with
coaxial cables and used with the 10Base-2 Ethernet system
Cat-3 - category 3 cable
Cat-5 - category 5 cable
CBR - Constant bit rate
Channel - smallest subdivision of a circuit
CoS - Class of service
CSMA/CD - Carrier Sense Multiple Access with Collision Detection
DTE - Data Terminal Equipment
DCE - Data Communication Equipment
DEC - Digital Equipment Corporation
FDDI - Fibre Distributed Data Interface
FOIRL - Fibre-optic inter-repeater link
Full Duplex Mode - Transmission of data in two directions simultaneously
GEA - Gigabit Ethernet Alliance
GbE - Gigabit Ethernet
Gbps - Gigabits per second or billions of bits per second
GMII - Gigabit Media Independent Interface
Half Duplex Mode - Data is transmitted at one direction at a time
IEEE - Institute of Electrical and Electronic Engineers
IPOA - IP over ATM
ISPs- Internet service providers
LAN - Local Area Network
LANE - LAN emulation
MAN- Metropolitan Area Network
MAC - Media Access Control
Mbps - Megabits per second or millions of bits per second
MIBs - Management Information Bases
Multiplexor - Device for combining many channels that have to carried in one fibre
NAS - Network Attached Storage
NIC – Network Interface Card
NSPs - Network service providers
PCS - Physical Coding Sublayer
PMD - Physical Medium Dependent
PMA - Physical Medium Attachment
POP -Points of Presence
QoS - Quality of service
RSVP - Resource reservation protocol
RTSP - Real-time streaming transport protocol
SAN - Storage Area Networks
SONET -Synchronous Optical Network
StarLAN - Developed by AT&T, this is based on IEEE802.3 implementation, and is a CSMA/CD LAN
STS- Synchronous Transport Signal
SDH -Synchronous Digital Hierarchy
SNMP -Simple Network Management Protocol
SWAN- ServerNet Wide Area network
TDM - Time Division Multiplexing
UBR - Unspecified bit rate
UTP - Unshielded Twisted Pair
VoD - Video on demand
VBR - Variable bit rate
WAN- Wide Area Network
WIS - WAN interface sub layer
Bibliography
Meeting the broadband access infrastructure demands: The promise of Gigabit Ethernet
Technological Forecasting and Social Change, Volume 72, Issue 1, January 2005, Pages 1-10
Joe Z. Cheng, Hsiao-Cheng Yu and W. David Sincoskie
Network traffic behaviour in switched Ethernet systems
Performance Evaluation, Volume 58, Issues 2-3, November 2004, Pages 243-260
Tony Field, Uli Harder and Peter Harrison
Wireless and Internet communications technologies for monitoring and control
Control Engineering Practice, Volume 12, Issue 6, June 2004, Pages 781-791
Haydn A. Thompson
An assessment of Gigabit Ethernet technology and its applications at the NASA Glenn Research Center: a case study
Journal of Engineering and Technology Management, Volume 20, Issue 3, September 2003, Pages 245-272
Catherine Murphy Bakes, Chan M. Kim and Calvin T. Ramos
Ethernet networks: Ethernet Networks, Design, Implementation, Operation
and Management; Gilbert Held (Ed.); Wiley, New York, 2003, ISBN
0-470-84476-0 Computer Communications, Volume 26, Issue 12, 21 July
2003, Page 1426
Hartmut Weik
Performance evaluation of switched Ethernet for real-time industrial communications
Computer Standards & Interfaces, Volume 24, Issue 5, November 2002, Pages 411-423
Kyung Chang Lee and Suk Lee
Ethernet : The Definitive Guide
by Charles E. Spurgeon
O'Reilly; 1 edition (February 1, 2000)
Gigabit Ethernet : Technology and Applications for High-Speed LANs
Rich Seifert
Addison-Wesley Professional; 1st edition (April 21, 1998)
Embedded Ethernet and Internet Complete
by Jan Axelson
Lakeview Research (June 1, 2003)
Switched, Fast, and Gigabit Ethernet (3rd Edition)
by Sean Riley, Robert Breyer
Sams; 3 edition (December 30, 1998)
Internet Application Protocols: The Definitive Guide
by Eric Hall
O'Reilly & Associates; Bk&CD Rom edition (September, 2003)
The Industrial Ethernet Networking Guide
by Donald J. Sterling, Steven P. Wissler
Thomson Delmar Learning; 1 edition (October 11, 2002)
Interconnections: Bridges, Routers, Switches, and Internetworking Protocols (2nd Edition)
by Radia Perlman
Addison-Wesley Professional; 2nd edition (September 14, 1999
Internet Routing Architectures (2nd Edition)
by Sam Halabi , Danny McPherson
Cisco Press; 2nd edition (January 15, 2000)
Networks (2nd Edition)
by Timothy Ramteke
Prentice Hall; 2nd edition (July 28, 2000)
Hancock, Bill.
Designing and implementing Ethernet networks / Bill Hancock.
2nd ed.
Wellesley, Mass. : QED Information Sciences, c1988.
Saunders, Stephen.
Data communications gigabit Ethernet handbook / Stephen Saunders.
Gigabit Ethernet handbook
New York ; London : McGraw-Hill, c1998.
Held, Gilbert, 1943-
Ethernet networks : design, implementation, operation, management.
2nd ed.
New York ; Chichester : Wiley, c1996.
Ferrero, Alexis.
The eternal Ethernet / Alexis Ferrero.
2nd ed.
Harlow : Addison-Wesley, 1999.
The Ethernet sourcebook / editor, Robyn Shotwell ; with an introduction by Robert M. Metcalfe.
New York ; Oxford : North-Holland, c1985.
Computer systems Local networks Ethernet
Quinn, Liam B., 1957-
Fast Ethernet / Liam B. Quinn, Richard G. Russell.
New York ; Chichester : Wiley, c1997.
Ferrero, Alexis.
The evolving Ethernet / Alexis Ferrero ; translated by Stephen S. Wilson.
Harlow : Addison-Wesley, 1996.
( Data communications and networks series)
Kadambi, Jayant.
Gigabit Ethernet : migrating to high-bandwidth LANs / Jayant Kadambi, Mohan Kalkunte, Ian Crayford.
Upper Saddle River, N.J. : Prentice Hall PTR ; London : Prentice-Hall International (UK), c1998.
Norris, Mark.
Gigabit ethernet technology and applications / Mark Norris.
Boston, Mass. ; London : Artech House, 2002.
Cunningham, David G.
Gigabit Ethernet networking / David G. Cunningham & William G. Lane.
Indianapolis : MacMillan Technical Pub., c1999.
Handbook of networking and connectivity / edited by Gary R. McClain.
Boston, Mass. : AP Professional, c1994.
Nunemacher, Greg, 1952-
LAN primer : an introduction to Local Area Networks / Greg Nunemacher.
New York, [N.Y.] : Prentice Hall, 1990.
Field, Tony.
Network traffic behaviour in switched Ethernet systems / Tony Field, Uli Harder & Peter Harrison.
London : Imperial College of Science, Technology and Medicine, Department of Computing, 2002.
Breyer, Robert.
Switched, fast, and gigabit Ethernet / Robert Breyer and Sean Riley.
3rd ed.
[U.S.A.] ; [Great Britain] : MacMillan Technical Pub, c1999.
Hein, Mathias.
Switching technology in the local network : from LAN to switched LAN to virtual LAN / Mathias Hein and David Griffiths.
London : International Thomson Computer, c1997.
Burg, Urs von.
The triumph of Ethernet : technological communities and the battle for the LAN standard / Urs von Burg.
Stanford, Calif. ; [Great Britain] : Stanford University Press, 2001.
|
