Frame Relay, ATM, Transport Layery p yDeepak JohnSJCET-PalaSJCET Pala

Frame Relay

Frame Relayi k i h d A l h h h i l d d li k is a packet switched WAN protocol that operates at the physical and data linklayers of the OSI reference model.

As fibre optic was introduced, the quality of circuits improved and there was no As fibre optic was introduced, the quality of circuits improved and there was noneed for error control.

was developed in response to a high speed, high performance and greater efficienti itransmission.

It puts data in variable-size units called "frames" and provide minimal internalcheckingchecking

support data transfer rates at T-1 (1.544 Mb/s) T-3 (45 Mb/s) speeds.

enabling end stations to dynamically share the network medium and the availablebandwidth.

i h d l A f ll i h f ll i l Devices attached to a Frame Relay WAN fall into the following two generalcategories:

1. Data terminal equipment (DTE)q p ( )

for a specific network and typically are located on the premises of a customer. Example of DTE devices are terminals, personal computers, routers, and bridges.p , p p , , g

2. Data circuit-terminating equipment (DCE)carrier-owned internetworking devices.Th i t id l ki d it hi i i t k hi hThe purpose is to provide clocking and switching services in a network, which

are the devices that actually transmit data through the WAN.

A hiArchitecture Frame Relay has 2 layers: physical and data link (LAPF, Link Access Procedure for

Frame Mode Bearer Services))

Physical Layer No specific protocol is defined for the physical layer in Frame Relay. Instead, it

is left to the implementer to use whatever is availableis left to the implementer to use whatever is available. Frame Relay supports any of the protocols recognized by ANSI.

Data Link Layer Link layer uses the services of the physical layer. It, in turn, provides the

f ll i ifollowing services : Flag recognition. Frame check sequence (FCS) generation and checking Frame check sequence (FCS) generation and checking.Recognition of invalid frames.Discard incorrect frames.Routing.Congestion control notification

Frame Relay Virtual Circuit Virtual circuits provide a bidirectional communication path from one DTE device

t th d i l id tifi d b b ll d d t li k tito another and are uniquely identified by a number called data link connectionidentifier (DLCI).

When a virtual circuit is established by the network, a DLCI number is given to ay , gDTE in order to access the remote DTE.

Frame Relay virtual circuits fall into two categories: switched virtual circuits (SVCs) permanent virtual circuits (PVCs).

1. Switched virtual circuits (SVCs) temporary connections, A new virtual circuit connection will be established each

time a DTE wants to make a connection with another DTEtime a DTE wants to make a connection with another DTE. A communication session across a SVC consists of the following four

operational states(Call setup ,Data transfer ,Idle and Call termination )2. Permanent virtual circuits (PVCs) permanently established connections by the network provider that are used for

frequent and consistent data transfers between DTE devices across the Framefrequent and consistent data transfers between DTE devices across the FrameRelay network.

Always operate in one of the following two operational states(Idle and DataTransfer)

P t l A hit tProtocol Architecture

LAPF CoreLAPF Frame

The address area which is 2 bytes in length is comprised of 10 bits The address area, which is 2 bytes in length, is comprised of 10 bitsrepresenting the actual circuit identifier and 6 bits of fields related tocongestion management.

DLCI fi ld 10 bit DLCI fi ld t th dd f th f d DLCI field: 10-bit DLCI field represents the address of the frame andcorresponds to a PVC.

Command/response (C/R): Designates whether the frame is a command orresponse.

Extended address (EA): used for expanding the number of possible addressesExtended address (EA): used for expanding the number of possible addresses.Forward explicit congestion notification (FECN):can be set by any switch to

indicate that traffic is congested. This bit informs the destination that congestiong ghas occurred.

Backward explicit congestion notification (BECN):is set (in frames that travelin the other direction) to indicate a congestion problem in the networkin the other direction) to indicate a congestion problem in the network.

Discard eligibility (DE): indicates the priority level of the frame. In emergencysituations, switches may have to discard frames to relieve bottlenecks and keepy pthe network from collapsing due to overload.

core functions of LAPF are used for frame Relay: Frame delimiting and transparency Frame mux and demux using addressing field Ensure frame is neither too long nor short Detection of transmission errors Detection of transmission errors Congestion control functions

LAPF-ControlThe user terminals(DTEs) implement full LAPF protocol, which is also called

LAPF-Control Protocol.The only difference b/w this protocol and LAPF-core is the inclusion of a control

field.Control protocol provides the functions of flow and error control that are missingControl protocol provides the functions of flow and error control that are missing

from core protocol

Congestion-Control Mechanisms Frame Relay reduces network overhead by implementing simple congestion-

tifi ti h i F R l i l t t ti tifi tinotification mechanisms. Frame Relay implements two congestion-notificationmechanisms: Forward-explicit congestion notification (FECN)p g ( ) Backward-explicit congestion notification (BECN)

FECN and BECN each is controlled by a single bit contained in the Frame Relayf h d Th F R l f h d l i Di d Eli ibiliframe header. The Frame Relay frame header also contains a Discard Eligibility(DE) bit, which is used to identify less important traffic that can be droppedduring periods of congestion.g p g

Four Cases of Congestion

Asynchronous Transfer yMode(ATM)

Introduction ATM is a concept similar to frame relay which take advantages of modern

digital facilities to provide faster packet switchingi i i d hi h d l d l i hi d i i is a connection-oriented, high-speed, low-delay switching and transmissiontechnology

allows multiple logical connections to be multiplexed over a single physicala ows u t p e og ca co ect o s to be u t p e ed ove a s g e p ys cainterface.

uses fixed sized packets called cells Developed to enable simultaneous Voice, Video, and Data traffic on the same

network with minimal error and flow control data rates of 25 6Mbps to 622 08Mbps data rates of 25.6Mbps to 622.08Mbps

Design Goals1. Use of high data rate transmission media (i.e fiber optic)2 I t bilit ith iti t h l i2. Interoperability with exiting technologies3. Implementation at reasonable cost4. Support for existing telecommunications hierarchiespp g5. Reliable and predictable6. Suitable for real-time and non real-time services

Cell Networks A cell network uses the cell as the basic unit of data exchange ATM carries information on cells ATM carries information on cells The length of each cell is 53 Bytes First 5 bytes are used as the cell header First 5 bytes are used as the cell header Next 48 bytes are used as the payload carrying the data

Fixed Length Cell Advantage Delay or latency is significantly reduced ATM is therefore suited for voice and video transmission ATM is therefore suited for voice and video transmission

Fixed length cells make it easier to switch data across multiple networks ATM networks are built based on switches and not routers

Fixed length cell is similar to container based road transportation

Multiplexing with cells The cells from the two lines are interleaved so that none suffers a long delay. Hi h d f th li k l d ith th ll i f th ll th t ll High speed of the links coupled with the small size of the cells means that cells

from each line arrive at their respective destinations in a continuous stream. A cell network can handle real-time transmissions, such as a phone call, without , p ,

the parties being aware of the segmentation

Asynchronous Time-Division Multiplexing ATM uses asynchronous time-division multiplexing to multiplex cells coming

from different channelsfrom different channels. It uses fixed-size slots, ie cells. ATM multiplexers fill a slot with a cell from any input channel that has a cell;

h l i if f h h l h ll dthe slot is empty if none of the channels has a cell to send

ArchitectureATM Devices ATM networks are built around two categories of devices ATM Switch ATM end-point

An ATM switch can be connected to either another ATM switch or and ATM An ATM switch can be connected to either another ATM switch or and ATMend-point.

ATM end point contain and ATM end-point adapterExamples of ATM end-points are Workstations,LAN switches, Routers etc

Two Types of Interfaces that interconnect ATM devices over point to point links:

User-Network Interface (UNI):connects an ATM end-system (client side) withan ATM switch (network site).

Network-Network Interface (NNI) : switches are connected through network-k i f (NNI )to-network interfaces (NNIs).

Virtual ConnectionConnection between two endpoints is accomplished through1. Transmission Paths (TPs):is the physical connection between an endpoint and a

switch or between two switches. A transmission path is divided into severalvirtual paths.p

2. Virtual Paths (VPs): provides a connection or a set of connections between twoswitches.

3. Virtual Circuits (VCs):SVC or PVC

A virtual connection is defined by a pair of numbers: VPI and VCI ATM assigns each Virtual Connection a 24-bit identifier

i l h d ifi ( ) ifi h h h f ll h h h1. Virtual Path Identifier (VPI), specifies the path the VC follows through thenetwork.8 bits long.

2. Virtual Channel Identifier (VCI), specifies a single VC within the path.16 bits( ), p g plong.

Cell networks are based on Virtual Connection and all cells belonging to asingle message follow the same virtual circuit

Example of VPs and VCsExample of VPs and VCs

ATM Protocol Layers

ATM Adaptation Layer ATM Adaptation Layer

ATM Endpoint ATM EndpointATM Switch

ATM Layer

Physical LayerATM LayerPhysical Layer

ATM LayerPhysical Layer

Physical Mediumy

Physical Layer It describes the physical transmission media. W hi ld d d hi ld d t i t d i i l bl d fib ti We can use shielded and unshielded twisted pair, coaxial cable, and fiber-optic

cable.

ATM Layer The ATM layer is responsible for establishing connections and passing cells

through the ATM networkthrough the ATM network. It provides routing, traffic management, switching, and multiplexing services

ATM cell 5-byte ATM cell header .the header can be in either UNI or NNI format

d di th I t fdepending on the Interface. 48-byte payload

General Format

ATM Cell Header—UNI Format

ATM Cell Header—NNI Format

General Flaw Control(GFC): Provides local functions, such as flow control fromend point equipment to the ATM switch.

P l d T (PT) I di t h th th ll t i d t t l d t Payload Type(PT): Indicates whether the cell contains user data or control data.

Cell Loss Priority(CLP): Indicates whether the cell should be removed if itencounters errors as it moves through the network.

d C l( C) C i C li d d Ch k (C C) h Header Error Control(HEC): Contains Cyclic Redundancy Check (CRC) on thecell header.

Virt al Path Identifier (VPI): Identifies semi permanent connections bet een Virtual Path Identifier (VPI): Identifies semi-permanent connections betweenATM end points.

Virtual Channel Identifier (VCI): Have only local significance on the link Virtual Channel Identifier (VCI): Have only local significance on the linkbetween ATM nodes.

ATM Adaptation Layer(AAL) It converts the submitted information into streams of 48-octet segments and

t t th i th l d fi ld f lti l ATM lltransports these in the payload field of multiple ATM cells.Similarly, on receipt of the stream of cells it converts the 48-octet information

field into required form for delivery to the particular higher protocol layer.q y p g p yAAL exists only in end systems, not in switches. AAL Services

dl i i Handle transmission errors Segmentation/reassembly (SAR) Handle lost and misinserted cell conditionsa d e ost a d s se ted ce co d t o s Flow control and timing control

AAL is classified into four(The classification was made with respect to the,following parameters: Timing relationship between sender and receiver Timing relationship between sender and receiver Related Not related

Bit rate Constant bit rate Variable bit rate

Connection mode Connection oriented Connection-oriented Connectionless

AAL is divided into two sub layers:i. Convergence Sub layer :manages the flow of data to and from

SAR blSAR sublayer.ii. Segmentation and reassembly sub layer:Packages data from CS

into cells and unpacks at other endp

Class A Class B Class C Class DServiceTimingbetween

Bit Rate Constant Variable

Related Not Relatedbetween

Source andDestination

Connectionless

ConnectionMode

ExamplesConstant Bit R t Vid

Connection Oriented Connectionless

Variable BitConnection-oriented Connectionless

Data TransferExamples

of Services

Rate Videoand Audio

Variable BitRate Videoand Audio

DataTransfer

AALTYPE AAL 1 AAL 2

AAL 3/4AAL 5

AAL 5AAL 3/4TYPE AAL 1 AAL 2 AAL 5 AAL 3/4

AAL 1 (Constant Bit Rate)The CS layer divides the bit stream into 47-byte segments and passes them to

the SAR sub layer belowthe SAR sub layer belowThe SAR sub layer adds 1 byte of header and passes the 48-byte segment to the

ATM layer.The 1 byte header is divided into two 4 bit fieldsThe 1 byte header is divided into two 4-bit fieldsSequence number (SN)Sequence number protection (SNP)

ATM Adaptation Layers: AAL2 AAL2 was intended to support a variable-data-rate bit stream. It is used for low-bit-rate traffic and short-frame traffic such as audio

(compressed or uncompressed), video, or fax allows the multiplexing of short frames into one cell allows the multiplexing of short frames into one cell. It widely used in wireless applications

AAL 3/4 AAL3 was intended to support connection-oriented data services and AAL4 to

support connectionless servicessupport connectionless services Later they have been combined into a single format called AAL3/4 the convergence sub layer (CS) creates a protocol data unit (PDU) by adding a

b i i h d t th f l th fi ld t il d i bl l th dbeginning header to the frame ,a length field as a trailer and a variable-length pad the segmentation and reassembly (SAR) sub layer fragments the PDU and append

a header to it . Then, the SAR sub layer appends a CRC-10 trailer to each PDU fragment for

error control The completed SAR PDU becomes the Payload field of an ATM cell The completed SAR PDU becomes the Payload field of an ATM cell

AAL 5 is the primary AAL for data and supports both connection-oriented and

ti l d tconnectionless data. also known as the Simple and Efficient Adaptation Layer (SEAL) The SAR sub layer simply accepts the CS-PDU and segments it into 48-octet The SAR sub layer simply accepts the CS PDU and segments it into 48 octet

SAR-PDUs without adding any additional fields. the CS sublayer appends a variable-length pad and an 8-byte trailer to a frame.

The trailer includes the length of the frame and a 32-bit cyclic redundancy check(CRC)

the SAR sub layer segments the CS-PDU into 48-byte blocks the SAR sub layer segments the CS PDU into 48 byte blocks . the ATM layer places each block into the Payload field of an ATM cell

Transport Layer

Introduction The transport layer is concerned with the provision of host-to-host user

ti f th li bl d t ff ti t f f d tconnections for the reliable and cost effective transfer of user data It Isolates upper layers from the network layer The transport layer is responsible for process-to-process delivery of a packet The transport layer is responsible for process to process delivery of a packet. At the transport layer, we need a transport layer address, called a port number,

to choose among multiple processes running on the destination host.

Transport Services provide logical communication between application processes running on

different hosts.Th t t f t t i Th ti i t d t t There are two types of transport service. The connection-oriented transportservice and connection-less transport service.

transport protocols are used for providing transport services .transport protocolsp p p g p p prun in end systems sender side: breaks messages into segments, passes to network layer receiver side: reassembles segments into messages, passes to higher layer

more than one transport protocol available to apps Internet: TCP and UDP Internet: TCP and UDP

Elements Of Transport Protocols Elements

1. Addressing2. Connection Establishment3 Connection Release3. Connection Release4. Flow Control and Buffering5. Multiplexing5. Multiplexing6. Crash Recovery

1. Addressing When an application process wishes to set up a connection to a remote application

it t if hi h t t tprocess, it must specify which one to connect to The method normally used is to define transport addresses is by using connection

requestsq The network layer address identifies a host. The transport layer address identifies

a user process – a service – running on a host In the Internet, these endpoints are called ports or TSAP (Transport Services

Access Points). The endpoints in the network layer (i e network layer addresses) are called The endpoints in the network layer (i.e., network layer addresses) are called

NSAPs (Network Service Access Points)

TSAPs, NSAPs, and transport connections

TSAPs, NSAPs, and transport connections

IP addresses versus port numbers

2 C i i2. Connection Establishment Just send REQUEST, wait for ACCEPTED. The problem occurs when the network can lose and duplicate packets The problem occurs when the network can lose and duplicate packets. Main problem is delayed duplicates

Solutions for delayed duplicates1. Using throw-away transport addresses

I hi h h i dd i d d In this approach, each time a new transport address is needed, When a connection is released, the address is discarded and never used again.2 Give each connection a connection identifier2. Give each connection a connection identifier Each connection is associated with a connection identifier. Whenever a

connection request comes ,transport entity update a table with connection. After each connection is released, each transport entity could update a table

listing obsolete connections. Wh ti t i it ld b h k d i t th t bl Whenever a connection request comes in, it could be checked against the table,

to see if it belonged to a previously-released connection

3.Setting Packet lifetime Packet lifetime can be restricted to a known maximum using one of the

following techniques:following techniques:Restricted subnet design(Any method that prevents packets from looping)Putting a hop counter in each packet(hop counter incremented every timePutting a hop counter in each packet(hop counter incremented every time

the packet is forwarded)Time stamping each packet(Each packet caries the time it was created, with

i di d k ld h i i )routers agreeing to discard any packets older than a given time)

Three-way handshake protocol Used for connection establishment Each packet is responded to in sequence Each packet is responded to in sequence Duplicates must be rejected Three protocol scenarios for three way hand shake) l ia) Normal operation

b) Old CONNECTION REQUEST appearing out of nowhere.c) Duplicate CONNECTION REQUEST and duplicate ACK.) p Q p

Three way handshake

(a)Normal operation. (b) Old duplicate CONNECTION REQUEST appearing out of nowhere. (c) Duplicate CONNECTION REQUEST(a) and duplicate ACK.

3. Connection RELEASE There are two styles of terminating aconnection:

asymmetric release symmetric release

Asymmetric releaseAsymmetric release only 1 peer closes the connection.is abrupt

and may cause data lossCR: Connection RequestDR: Disconnect Request

Asymmetric releaseAsymmetric releaseAbrupt disconnection with loss of data

The two-army problem.

The blue army has 4 troops (2 on either side of valley) while the white army has 3troops. If both blue armies charge at the same time they can vanquish the whitetroops. If both blue armies charge at the same time they can vanquish the whitearmy. If only one of the blue armies charges it will succumb (3 white troops against2 blue troops). This means: the blue armies have to synchronize their attack. But inorder to synchronize they need to send a messenger through the valley; of courseorder to synchronize they need to send a messenger through the valley; of coursethe messenger can get caught by the white army (‘lost packet’).

Approach #1: The blue army #1 sends a messenger to tell blue army #2 to attack@ 1400.

Problem: The blue army #1 does not know if the messenger managed to conveymessage or if he was caught. Thus blue army #1 will not attack.

Approach #2: The blue army #2 sends back a messenger to acknowledge to blueApproach #2: The blue army #2 sends back a messenger to acknowledge to bluearmy #1 that it got the message.

Problem: The blue army #2 does not know if acknowledge-messenger reachedblue army #1. Thus blue army #2 will not attack.

This play can be continued.

S i lSymmetric release each direction is released independently of the other one.Four protocol scenarios for releasing a connection:Four protocol scenarios for releasing a connection:

(a) Normal case of a three-way handshake. (b) final ACK lost

(c) Response lost. (d)Response lost and subsequent DRs lost.

4. Flow control and Buffering The sender process may send at much higher speed than the receiver process can

handle the data thus causing overflow (= packet loss)handle the data thus causing overflow ( packet loss). Transport layer segments the data stream

If most Segments are nearly the same size, it is natural to organize the buffers asg y , ga pool of identically-sized buffers, with one Segment per buffer

If there is wide variation in Segment size, a pool of fixed-sized buffers presentsproblemsproblems.

(a) Chained fixed-size buffers. (b) Chained variable-sized buffers. (c) One large circular buffer per connection.

If the buffer size is chosen equal to the largest possible Segment, space will bewasted whenever a short Segment arrives.If h b ff i i h l h h i S i l i l If the buffer size is chosen less than the maximum Segment size, multiplebuffers will be needed for long Segments, with the attendant complexity.

Another approach to the buffer size problem is to use variable-sized buffers.ot e app oac to t e bu e s e p ob e s to use va ab e s ed bu e s. The advantage here is better memory utilization, at the price of more

complicated buffer management. A third possibility is to dedicate a single large circular buffer per connection This system is simple and elegant and does not depend on segment sizes, but

makes good use of memory only when the connections are heavily loadedmakes good use of memory only when the connections are heavily loaded.

5. Multiplexing & De-multiplexing In the transport layer the need for

multiplexing can arise in amultiplexing can arise in anumber of ways.

There are two types ofl i l imultiplexing

i. Upwardii. Downward

i. Upward multiplexing Traffic from a “data stream” is distributed over several transport connections

(TSAPs)(TSAPs). For Eg, if only one network address is available on a host, all transport

connections on that machine have to use it.

ii. Downward Multiplexing Many “data streams” share the same transport connection using multiple NSAPs,

possibly over multiple network interfaces (load balancing)possibly over multiple network interfaces (load balancing).

6. Crash Recovery A crash of one host (server) during the transmission leads to a connection loss

which results in data loss Solution for this the client retransmits onlywhich results in data loss. Solution for this, the client retransmits onlyunacknowledged packets.

Does not work in all cases

A crash at layer N can only be handled at layer N+1 (a system crash is a crash atevery layer).

Th It i l ft t th li ti l t h dl h f th t h tThus: It is left to the application layer to handle crashes of the remote host(client or server).

Generally applications detect that the remote host has died and then simplyy pp p yrestart the connection and retransmit everything.

Each client can be in one of two statesi. S1: 1 unacknowledged packet outstanding ii S0:No unacknowledged packet outstanding

Three events are possibleat the server

i sending an ack (A)ii. S0:No unacknowledged packet outstanding The server can be programmed in one of two waysi. First ACK, then write

i. sending an ack (A),ii. writing to the output

process (W),,ii. First write, then ACK The client can be programmed in one of four ways

iii. crashing (C)

i. always retransmit the last segment,ii. never retransmit the last segment,

i l i S0iii. retransmit only in state S0,iv. retransmit only in state S1.

A Simple Transport Protocoli i i i ll ( li i ) Transport Service Primitives allows transport users (e.g., application programs) to

access transport service. five primitives: CONNECT, LISTEN, DISCONNECT, SEND, and RECEIVE. five primitives: CONNECT, LISTEN, DISCONNECT, SEND, and RECEIVE.

The parameters for the service primitives and library procedures are as follows: connum = LISTEN(local) connum = CONNECT(local, remote) status = SEND(connum, buffer, bytes) status = RECEIVE(connum buffer bytes) status = RECEIVE(connum, buffer, bytes) status = DISCONNECT(connum)

The LISTEN primitive announces the caller's willingness to accept connection p g prequests directed at the indicated TSAP.

The CONNECT primitive takes two parameters, a local TSAP (i.e., transportaddress), local, and a remote TSAP, remote, and tries to establish a transportconnection between the two If it succeeds it returns in connum a nonnegativeconnection between the two.If it succeeds, it returns in connum a nonnegativeotherwise a negative number

The SEND primitive transmits the contents of the buffer as a message on theindicated transport connection, in several units if needed. Possible errors, returnedin status, are no connection, illegal buffer address.

The RECEIVE primitive indicates the caller's desire to accept data The size of the The RECEIVE primitive indicates the caller s desire to accept data. The size of theincoming message is placed in bytes. If the remote process has released theconnection or the buffer address is illegal, status is set to an error code indicatingh f h blthe nature of the problem

The DISCONNECT primitive terminates a transport connection. The parameterconnum tells which one.P ibl b l t th i t lid Possible errors are connum belongs to another process or connum is not a validconnection identifier.

The transport layer makes use of the network service primitives to send andp y preceive TPDUs.

The hardware and/or software within the transport layer that does the work isll d th t t titcalled the transport entity.

Network packet Meaning

Call request Sent to establish a connection

Call accepted Response to Call Request

Clear Request Sent to release connectionClear Request Sent to release connection

Clear confirmation Response to Clear request

Data Used to transport data

Credit Control packet to manage window

Transport entity: packet types

Transport entity: states of a connection

State Meaningg

Idle Connection not established

Waiting CONNECT done; Call Request sent

Queued Call Request arrived; no LISTEN yet

Established connection has been established.Sending Waiting for permission to send a packet

Receiving RECEIVE has been done

Disconnecting DISCONNECT done locally

Graphical Representation