Saturday, June 29, 2013
ATM NETWORKS
INTRODUCTION
In
recent years, computer systems have achieved an almost explosive increase in
performance, such that developments in the field of data communications have
been unable to keep up. Where the processor performance and memory capacity of
PCs have grown over the past ten years by a factor of a hundred, transfer
speeds in wide area networks have risen 'only' by a factor of ten, and in LANs
they have been stayed much the same.
ATM is a standard recognized throughout the world,
which provides for the first time a method for universal information exchange,
independent of the end-system and the type of information (data, audio, video).
The architecture of ATM (53 byte cells) supports the
design of massive parallel communication architectures and enables the
implementation of networks with transfer rates in the gigabit range. With the
high-speed networks, it is possible to send huge quantities of data generated
by the latest applications (video mail, interactive TV, virtual reality, etc)
ATM is suitable for local area networks as well as
wide area networks. The ability of ATM to emulate traditional LAN and WAN
architectures will ensure a smooth transition from today's computer network
infrastructure to ATM-based high-speed technology.
The intense development efforts being made all major
manufacturers of data communication systems in the area of ATM/B-ISDN are
evidence of the strategic importance to the industry of this new transmission
technology. ATM is increasingly being adopted as the central strategic
technology for data communication throughout the industry today.
ATM opens the possibility of designing networks with
transmission speeds up to the physical limits. ATM networks with bandwidths
approaching the bandwidth of light (30 Terahertz) are imaginable and ATM
switches with processing speeds of up to 1 Tbit/s have already been
demonstrated in research laboratories. In ATM we may have on our doorstep the
ultimate transfer mechanism in data communications.
ATM - ASYNCHRONOUS TRANSFER MODE:
Asynchronous Transfer Mode is a communication standard
that uses a high-speed form of packet switching network as the transmission
media ATM was developed as a part of the Broad band Integrated Services Digital
Network (BISDN). ATM is intended to utilize the synchronous optical network
(SONET).
Conventional
electronic switching (ESS) machines currently utilize a central processor to
establish switching paths and route traffic though a network. ATM switches,
however, will include self-routing procedures where individual cells containing
subscriber data will route their own way through the ATM switching network in
real time using their own address instead of relying on an external process to
establish the switching path (a cell is a short, fixed length packet of data)
PRINCIPLE OF ATM:
ATM - originally designed for WAN communications, but
quickly adapted for LANs as well, ends this historical separation and forms a
universal platform for data communication, In both ATM LAN AND ATM WAN networks
the data transport is achieved via connection-oriented communication paths,
which are set up though high-speed switching systems. These ATM switches
perform the cell routing from the input ports of the switch to the destination
port in real time and in parallel for the ports.
For data transport ATM uses packets with a fixed
length of 53 bytes, the so-called cells. These cells can be processed
significantly faster and more efficiently in switching systems than data
packets of variable length. Because of the cell structure, a massive parallel
architecture of ATM switching systems is possible. Since all cells have the
same length, all data units which wait at the input ports of a switch for
transportation at a given time, can be routed simultaneously to their
destination port.
ATM can handle all of today's data services
(telephone, data, video-broadcast and interactive) in an efficient way.
WHY ATM?
Reasons for ATM
Ø
Increased bandwidth and real-time
responsiveness.
Ø
Advantages over Ethernet networks.
Ø
Limitations of Token Ring and FDDI
networks.
Ø
High performance of ATM networks.
Ø
ATM is a world wide recognized
standard, with which a universal information exchange can be realized for the
first time, independent of the type of end system and service (data, video,
audio).
Ø
ATM is suitable for LANs as well as
WANs.
Ø
ATM is able to handle all existing
information services simultaneously and efficiently.
Ø
Since ATM is scalable and therefore
available in all speed classes.
ATM- THE EVOLUTION OF A UNIFIED
PLATFORM FOR DATA COMMUNICATION.
It will be possible in the future, using broad band
ISDN networks, to transmit the four basic services in
telecommunications-speech, picture, data and video- quickly and cheaply via a
single network infrastructure. In 1988, ATM was selected by ITU as the
transport mechanism for B-ISDN networks.
TRANSMISSION PROCEDURE.
In asynchronous time division multiplexing, the data
streams to be transmitted are converted into information units of fixed or
variable length and transferred asynchronously. The allocation of the units of
information to the different transmission channels is carried out using
numerical channel identifiers attached to each data packet. For this
reason,asynchronous time division multiplexing is some times referred to as
label multiplexing. If variable- length data packets are used for transmission,
this is known as Packet switching. If fixed length data packets are used then
it is known as Cell switching.
An ATM Cell contains all of the network information
needed to relay individual cells from node to node over a pre-established ATM
connection. The figure shows the ATM cell structure, which are 53 bytes long,
which includes 5-bytes header field and a 48-byte information field.
The information field consists of user data. The header field
is for networking purposes and contains all of the address and control
information necessary for address and flow control.
ATM header field
The figure shows the
structure of the 5-byte ATM header field which includes the following: generic
flow control field, virtual path identifier, virtual channel identifier, payload
type identifier, cell loss priority and header error control.
| Generic Flow control: | |||
The GFC field uses the first 4 bits of the first byte of the header field . The GFC controls the flow of traffic across the user network interface and into the network .
Virtual path identifier and Virtual channel identifier uses the 24 bits immediately following the GFC are used for the ATM address .
|
Pay
Load Type (PT):
The First 3 bits of the second half of byte 4 specify
the type of message in cell . With 3 bits there are 8 different types of
payloads possible. However types 0 to 3 are used for identifying the type of
user data and types 4 and 5 indicate
management information, types 6 and 7 are reserved for future use.
Cell
loss Priority (CLP):
The last bit of byte
4 is used to indicate whether a cell is eligible to be discarded by the
network during congested traffic periods . the Clp bit is set by the user or
cleared by the user.If set the network may be discard the cell during heavy times of heavy use.
Header
Error Control:
The last byte of the header field is for error control
and is used to detect and correct single bit errors that occur in the header
field only, the HEC does not serve the entire cell check character .the value
placed in the HEC is computed from the 4 previous bytes of the header field .the HEC provides some protection
against the delivery of cells to the wrong destination address.
ATM
Information Field:
The 48 -byte information field is reserved for user
data. Insertion of data into the information field of a cell is a function the
upper half of layer two of the ISO-OSI
seven layer protocol hierarchy. this layer is specifically called the ATM Adaptation layer(AAL). The AAL gives
the versatality necessary to facilitate. In a single format ,a wide variety of
different types of services ranging from continuous process signals, such as
voice transmission ,to message carrying highly fragmented bursts of data such
as those produced by the local area networks. The AAL divides information into
48-byte segments and places them into a series of segments.
It is hoped that incorporating high -speed ATM systems into data communications networks
will encourage the installation of more optical fiber systems in the very
future. the transmission rates agreed upon by a consortium of over 120 firms known as the ATM Forum,
are 45Mbps,100Mbps,and 155 Mbps, which
are the same rates are used by SONET. The AATM Forum was founded in the year
1991 by Adaptive CISCO , Current ATM
transmission rate go up to 622 Mbps.
ATM
MULTIPLEXING PROCEDURE:
ATM is a transmission procedure based on
asynchronous time division multiplexing using fixed data packets .these data
packets are known as cells and have a length of 53 bytes .all the nodes in the
net work are connected Via one or more ATM switches which route the cells to
their destinations. the stations on the
network do not share the common transffer medium ,as is the case of in local
networks ,but hand over their cells to
at the ATM switches without the need of media access algorithms. the total
transmission bandwidth available is allocated by the ATM switch as required .
The fixed length of 53bytes for a cell
is the result of compromise between the demands of analog speech data transfer
and digital data transmission .In digital transfer of analog speech signals the
speech is sampled 8000 times a second
and each sampled value is
transmitted as an 8-bit code.
ADVANTAGES OF ATM:
Ø Efficient
utilization of the total transmission bandwidth:
ATM networks using asynchronous time division
multiplexing with a fixed packet length ,can allocate the available bandwidth
in a flexible way, with each user provided with more or less capacity according
to changing requirements. this means that the entire bandwidth is divided among
the network nodes active at any time .this enables ATM networks to implements services with very
variable bandwidth requirements such application with highly varying bit rates.
Real-time
applications
Fixed bit rate and time critical applications with a high
degree of efficiency .ATM networks are suitable for all types of data traffic
,whether videoconferencing ,phone calls or file transfers.
Ø Scalability and
modularity:
ATM
networks can be implemented on any one of number of transfer media. until now
the network standards have been strictly defined right down to the physical
level (Ethernet, FDDI, and so on).for ATM networks this is not the case .there
is thus no explicit specification as to which physical medium ATM cells should
be transffered over , or at what speed in addition ,ATM _based networks can be
to accomadate new users without the bandwidth available to existing users being
restricted as a result. It is simply a matter of adding more connection modules to the ATM switch
serving the the users. so that ATM can be used be used ias the transmission
mechanism in practically all areas of
data communications .this makes ATM equally suitable for local and wide
area traffic.
ATM
in wide area networks :
ATM cells can be carried both by
existing
1.544/2.048-Mbits/s,34/45-Mbits/s or 140-Mbits/s links and by latest
SDH networks developed and
standardized only in the last few years .In
1988 the agreement was reached in
the ITU on the SDH transmission procedure(SDH,155,622 and 2400- Mbits/s) as a
worldwide unified standard in wide-area communications and its implementation
throughout the world was recommended.
This means that the SDH will certainly
prove to be the front runner as a transfer medium for ATM in the long term
. however as the transition phase will
involve using conventional transmission links for several years the transfer of ATM cells has
also been standardized for these interfaces.
ATM i n
LAN Applications:
For LAN ATM infrastructures,interfaces have been
defined by the ATM Forum with transfer speeds of 25 Mbits/s52 Mbits/s and
155Mbits/s via unshielded and shielded copper twisted -pair lines .
Existing FDDI infrastructures can be upgraded to
ATM-based LANS using what are known as
TAXI chipsets. the field of application
for ATM in the local area are primarily sophisticated multimedia LANS and backbones connecting up conventional
LANS. This means that existing network nodes can be connected directly to ATM
topologies without any change to existing network nodes can be connected
directly to ATM topologies without any change to existing software applications.
The
architecture of B-ISDN network:
Introduction:
The
logical B-ISDN network architecture comprises 4 independent communication
layers based on the OSI reference model (ITU
X.200).The 4 layers of the B-ISDN protocol are linked together via 3 planes:
·
The user
plane
·
The control plane
·
The management plane
The control
plane :
The
control plane is responsible for setting up,
releasing and monitoring data connections . ATM is a connection oriented
transfer mechanism : this
means that every connection with in the ATM layer must first be allocated a
unique numerical identifier via the control planes signaling procedures . This
number is either virtual path identifier or the virtual channel identifier
depending on the hierarchy of the connection.
The management plane :
The management plane has 2 functions:
·
plane management
·
layer management
Plane management coordinates the functions and
procedures of the management plane with
those of the other 2 planes. Layer
management is responsible for functions such as meta-signaling and the OAM
information flow. Meta-signaling is a separate information channel to
control various signaling procedures
. OAM
( operation &
maintenance ) information is used
to monitor network performance and for error management at ATM level .
Functions of the layers in
the B-ISDN reference model :
The 4 layer of the B-ISDN model are
·
The physical layer
·
The ATM layer
·
The ATM adaptation layer
·
The user layer
The physical layer :
The physical layer consists of two sub
layers transmission convergence sub
layer and physical medium . The Transmission convergence (TC) is in charge of
embedding the cells of the ATM layer in
the transmission frames of the transport
medium in use . two of the most important functions of TC
are cell delineation and HEC generation . The standards for B-ISDN are
set up in such a way that practically any physical medium can be used as an appropriate transmission adapter has
been specified .
As
a general rule the maximum
achievable BW for copper cables
over short distances can be said between
300 and 400 MHz . The
corresponding figures for optical fibers will be in the Terahertz range . High bandwidths(BW) are simply for
less expensive to carry over long distances using optical media than by
electrical means.
The ATM layer:
The
main task of the ATM layer is to transport the
data passed down to it by the adaptation layer(AAL) to
its intended destination .this makes the
ATM layer the transport mechanism in B-ISDN
networks. the information in the ATM layer are 53-byte cells, each of
which includes in its cell header a
numerical identifier allocating it to a specific connection .these cell streams
are divided in to two logical hierarchies :Virtual channels and virtual paths.
Each cell can be assigned to be a specific path or channel by reference to
their numerical path (VPI,VCI) contained in its header .
VPI/VCI Conversion:
If
cells are routed via ATM switches or cross connects ,the VCCI and VPI values
applying up to that point need to converted in to new VPIs or VCIs specifying
the cells new destination .
If
ATM layer receives n information unit
from the AAL layer above it must generate an appropriate ATM header .It is a central task of the ATM
layer to convert network addresses in the higher levels into the corresponding
VPIs and VCI values. the number
subscribers are accessing the physical medium ,the GFC field in an ATM cell can be used to control cell transfer.
The ATM Adaptation
layer(AAL):
The
job of the AAL layer is to 'segment' the
data streams from the higher applications layer into 48_byte units of
information and to reassemble the original data streams from ATM from cells
.the functions of the ATM layer depend
on the characteristics of the governing applications -that is ,the AAL layer is
service -dependent .it consists of two sub-layers : Convergence Layer (CS) and
the Segmentation and reassemble of Sub-layer(SAR).
The different AAL types:
To
limit the number of different AAL implementations, four service classes have
been defined for the AAL Layer: AAL1, AAL2, AAL3/4 and AAL5. The definition of the various AAL types is
based on the following three parameters.
Ø Real time
requirements
Ø Bit rate
(constant or variable )
Ø Connection type
(connection oriented or Non-connection oriented)
AAL
TYPE 0:
AAL
type 0 denotes an absence of any AAL function, meaning that AAL0 is not
really an AAL type in the true sense of the term. The functions of the applications layer are
infact superfluous for any service if the transfer mechanism already based on
cells and can therefore be dispensed with.
AAL
TYPE 1:
The type 1 adaptation layer is used to transmit
applications with constant bit rate via the B-ISDN network. In addition AAL type 1 protocol can transfer
structured data in structured form. Lost
or erroneous data is not corrected or repeated.
As with all other AAL types the type 1 ATM adaptation layers consists of
segmentation and reassembles sub layer and convergence of layer.
AAL
TYPE 2:
The
adaptation layer for type 2 is designed
for the transmission of data streams with variable bit rates, there is a time
correlation between sender and receiver in the case of AAL type 1. The adaptation layer for AAL type 2 has not
yet been specified in detail.
AAL
TYPE 3/4:
The
adaptation layer type 3/4 specifies the connection oriented and non-connection
oriented transfer of data packets via B-ISDN network. The connection setup for this may be either
point to point or point to multi point.
This makes the AAL 3/4 protocol suitable.
AAL
TYPE 5:
The
AAL type 5 sub layer amounts to a
greatly simplifier implementation of AAL3/4.
TYPES
OF ATM CELL
As
well as dividing into UNI and NNI cells, ATM cells can be further allocated one
of four categories: Idle cells, Unassigned cells, Physical layer OAM cells and
VP/VC cells.
Idle cells:
Idle
cells allow the cell rate to be adjusted to the transfer medium bandwidth. If there are not enough cells to fill
bandwidth provided idle cells are transmitted.
This achieves synchronization with the transmission speed of the
physical medium. Idle cells are not
passed to the ATM layer.
Un-assigned cells:
Unassigned
cells are cells that have a VPI or VCI value but a blank data field.
Physical layer OAM cells:
For
direct cell transfer on the cell based physical layer, every 27th
cell is used to transfer OAM information concerning the physical layer. After receipt by the physical layer, these
cells are not passed on to the ATM layer.
VP/VC cells:
The
cells used for communication within virtual channels or paths can be subdivided
into six functional groups. Cells for
transmission of user data, cells for media signaling, cells for broadband
signaling, VC OAM cells SMDS/CBDS cells, ILMI(Interim Local Management Interface
Specification) cells.
ATM
SWITCHES AND CROSS-CONNECTS
ATM switching units -VC
switches and Crossconnects
-are the central element in any B-ISDN network. the fact that all ATM cells are
the same size is exploited by a massive parallel architecture . It is the
gigabit and terabit cell throughput rates made possible by this architecture
that enable high speed networks like B-ISDN to be implemented . the switching
speeds in ATM switches and
cross-connects exceed the transfer rates of the connected stations by many
times , and all the user channels that are to be connected can be fully processed..
Basic Functions Of ATM Switching
Units:
An ATM switching unit
has two basic jobs:
To identify and
analyse the channel and path identifiers
( VPI/VCI ) in the ATM cell.
v To Transport the ATM cell from one of the units input ports ti the output port that takes the
ATM cell to its intended destination.
There are two main
types of ATM switching units :
v VP Switching or crossconnects.
v VC Switches.
Cross -Connects(VP Switches):
ATM
crosss-connetcs terminate incoming paths and transfers them -along with all the
channels in the path-to another
,outgoing path the individual channels are unaffected by this .
VC Switches:
VC switches terminate both
incoming paths(VPs)and incoming channels (VCs) and re-route them to other
outgoing paths and channels .Switching of virtual channels thus always implies
a re-routing of paths ,as the path in
the channel is being transported must always be terminated when the channel is terminated.
VC switches, however ,can also the switching
unit unaffected.
The
Topology of ATM Switching units:
The actual transport of ATM
cells within an ATM switching units is carried out via the switching fabric,
the heart of the unit. the task of the switching fabric is to provide dynamic transmission paths between
the input ports and the out[put ports requested at any given time in such away
that the fewest possible external and internal conflicts occur. An internal
conflict occurs if to ATM cells in a multi-stage switching network
are competing for the same output port
at the dame switching stage. if a blockage of this type occurs at an
output controller-that is, at the output to a switching network-it is known as
external conflict.
Switching
Elements:
Switching
fabrics are made up of small cell-routing units known as switching elements.
Even a single switching element can be
used as a switching fabric. The switching elements themselves consists of an
interconnection network providing the transmission paths for the ATM cells
.there are two basic types of interconnection network for switching elements.
1.
Matrix structure networks.
2.
Time division Multiplexing networks.
Matrix
structure:
In Switching
elements with a matrix structure ,the ATM cells are transported in parallel via
a network lattice(crossbar) connecting together the inputs and outputs of the
switching element. the transfer of all the cells arriving at the input
controllers at a given instant in the switching process is carried out
simultaneously and in synchronization with a local clock. the cycle time
between two switching instants is known as a slot. if two cells are competing
for the same output port at the same
instant , a blockage may occur. To avoid the loss of cells through
blockages it is necessary to include buffer memory at the input and output
ports and at the points where the transmission paths cross.
The matrix Switching element of dimensions N by N will not cause blockages for a
randomly distributed load if the speed--up factor is equal to N. If K is less
than N, there will need to be additional buffer memory available at the input
ports to ensure there are no cell losses.
Switching
Elements Based on Time Division Multiplexing:
1. Bus Switching
elements:
In switching elements based
on bus topology ,the interconnection
network is implemented using a 16-bit or
32-bit high-speed Bus. To enable the ATM cells to be transfered without colliding , the transfer capacity of
the bus must be at least equal to the sum of the transfer Capacities of the
input ports .As the Transfer capacities of the bus is several times higher than
the rate of the incoming cells,the input controllers have no difficulty in forwarding the cells immediately .In order to
adjust their transfer rates to the output port rate Bus switching systems
therefore need output port buffering .
2. Ring Switching Elements:
In
this the input and output controllers
are connected
via
a ring. Compared with Bus topology ,a ring offers theadvantage
that
a time slot can be used by more than one input controller
within each rotation , althougth
extra overhead is required to
control
this mechanism. Although it enable an effective load of
more
than 100% of the ring capacity to be achieved .
3. Central Memory
Switching Elements:
In
central memory switching the cells are written by The input controllers to a common
area of memory from where they are read by the output controllers. As the
buffers for all the output ports Share the same area of memory, this can result
in significant savings of memory Space. Because of their efficient use of
memory switching elements with central memory topology are used
particularly in the large switching
units with a large number of input and output ports.
Switching
Networks
The switching
Structure itself is made up of
Switching
networks , which link up the individual switching
elements. give
the central importance of the architecture of
switching
networks to the performance of switching fabrics,
numerous research
projects have been carried out in this area over
recent
years ,the aim of all this research was to achieve the highest
possible
throughput rates at the lowest level of integrated
circuits
.the following are the different network
tolpologies .
1.The Shuffle Exchange
Network.
2.The Extended
Switching Matrix Network.
3. Banyan Networks.
4. Benes Networks.
5. Parallel Banyan Networks.
6. Distribution Networks.
Cell
Routing In Switching Networks
There are two
methods of routing cells along the different transmission paths
inside switching networks :
1.Self-Routing .
2.Table-Controlled
Routing.
Self-Routing :
In Self-Routing
an additional header -
specific to the
element -is added to the front of cells ,containing coding for the transmission
path along which the cell is to be sent .If the Switching network is
constructed from n stages ,this header will contain n sub-fields giving the
path selection at each of the nodes in the switching network .Because of the
additional header ,the internal
processing speed must be increased in proportion o the length of the Self-Routing header is five bytes .
Table
-Controlled Routing:
In table controlled routing the length
of the ATM cell is unchanged . before
each switching element ,the channel or Path identifier( VPI/VCI) of the cell translated into a switching
-specific value indicating the destination for the cell at this switching
element .the values allocated to the cells are defined in the connection set-up
or path selection phase and stored in tables .Self -Routing algorithm is better
suited to large multistage architectures than the table routing method.
Traffic
Control And Congestion Control In ATM Networks
A network element in an ATM network is
described as
congested (overloaded)if it is no longer able to maintain the agreed performance
parameters for an existing connection .two things may be responsible for this
1. unpredictable
statistical variation in the
traffic flow.
2. Errors within the network.
The task of ATM congestion control is to take various
measures to minimize the extent and
duration of congestion episodes .the traffic control function is designed to
achieve by optimizing the usage of existing network capacity.
Functions
And Mechanisms:
The following functions are provided forThe
implementation of traffic monitoring and
congestion control in ATM networks.
Traffic Control:
Management of network capacity.
Access Controls (connection
administration control).
Usage Parameters Control
(UPC).
Selective discarding of
cell.
Traffic Shaping.
Sending Congestion messages
to remote station.
Congestion
Control:
Traffic shaping.
Sending congestion messages
to remote station.
Management of Network Capacity:
The management of
network capacity is implemented by means of path management this allows the
switching requirements for the setting up of path connections to be reduced by reserving in paths.
The
end -to-end transmission quality for a given channel connection is directly
dependent on the quality of the series of paths in which the channel is located
.if various channel connections are routed via the same path they will have
similar performance and quality parameters, such as cell loss rate and cell
transfer delay. Channels with similar
quality parameters should therefore be
routed over the same ATM path by the traffic
control. If the overall transmission rate of all the channels exceeds
the capacity of the path ,the cell loss can be distributed over all he channels
by means of statistical time division multiplexing.
Connection Admission Control
:
Access
controls are carried out by the during the connection setup phase. This involves checking the traffic contract requested by user
through the source traffic descriptor and the QOS class to see whether it is
possible to set up a connection.
Usage Parameter control And Network Parameter
Control:
Usage
parameter control and network Parameter control are similar functions of two different interfaces. UPC is
carried out at the user-network interface ,NPC at the inter-network interfaces.
The task of the
parameter control function
Is To check that
the traffic profile negotiated for the
particular case and correctness of the
path or channel identifiers(VPI,VCI) are being maintained this involves first
checking the validity of the values of the
VCI/VPI, then measuring the traffic volumes generated by the appropriate channels and
paths. A check can be made on whether the traffic parameters negotiated for the
connection concerned are being adhered to. if not the parameter control function can take one
of the action as Cell marking or Cell discarding.
Traffic Shaping:
Traffic
Shaping allows a Cell stream o be profiled so that if conforms to a specific
traffic characteristic .traffic shaping
can be implemented by either the network or the user .
Sending Congestion Messages
to Remote station:
A
network element in a congestion situation can send a congestion message to its
remote station asking for a reduction in he transmission rate . however ,no
precise mechanism for this has yet been defined.
ATM ON INTERNET
ATM
will become an increasingly important technology not only in corporate
environments but also with in the realm of the Internet .Graphics-and
video-intensive applications necessciates higher speeds .by current standards
,high -speed refers to networks that operate at 155Mbps.A typical ATM switch
can between 16 and 64 ATM devices .although a single ATM switch has finite
capacity,switches can be interconnect4d to form a larger network.
IP
address Binding In An ATM network as in technologies, ATM assigns to each
attached computer a physical address that must used when establishing a virtual
circuit .On one hand because an ATM
physical address is larger than an IP address. Thus IP can not use
static address binding for ATM networks ATM hardware does not support
broadcast. Thus IP cannot use conventional ARP to bind addresses on ATM
networks.
conclusions:
ATM networks are
suitable for all types of data traffic ,
videoconferencing ,phone calls , file transfers and it is useful for
internet applications.
