Packet Switching

Outline Switching and ForwardingBridges and LAN Switches Cell Switching (ATM)Switching Hardware

Problem: Not all networks are directly connected

• Limitations of the directly connected networks:– How many hosts can be attached.– How large of geographic area a single network can

serve.• A switch is used to enable packets (a limit-size block of

data) to travel from one host to another.• The jobs of a switch are:

– Forward packets– Handle contention– Solve the congestion (Chapter 6)

• Two technologies are focused in this chapter:– LAN switching– Asynchronous transfer mode (ATM)

Switching and Forwarding

Outline Store-and-Forward SwitchesBridges and Extended LANs Cell SwitchingSegmentation and Reassembly

Switching and Forwarding

• A switch is a multi-input, multi-output device, which transfers packets from an input to one or more outputs.

• A switch establishes the star topology:– Large networks can be built by interconnecting a

number of switches.– We can build networks of large geographic scope.– Adding a new host to the network does not necessarily

mean the hosts will get worse performance. Switched network is considered more scalable.

Scalable Networks • Switch is the main function of the network layer.

– forwards packets from input port to output port– port selected based on address in packet header

– Approaches: datagram/connectionless, virtual circuit/connection-oriented, and source routing

• Advantages – cover large geographic area (tolerate latency)– support large numbers of hosts (scalable bandwidth)

Inputports

T3T3

STS-1

T3T3STS-1

Switch

Outputports

Switching and Forwarding

A switch provides a star topology.

Datagram Switching• No connection setup phase• Each packet forwarded independently • Sometimes called connectionless model

0

13

2

0

1 3

2

0

13

2

Switch 3Host B

Switch 2

Host A

Switch 1

Host C

Host D

Host EHost F

Host G

Host H

• Analogy: postal system

• Each switch maintains a forwarding (routing) table

Datagram Model

• There is no round trip time delay waiting for connection setup; a host can send data as soon as it is ready.

• Source host has no way of knowing if the network is capable of delivering a packet or if the destination host is even up or running.

• Each packet is forwarded independently.• A switch or link failure might not have any serious

effect on communication if it is possible to route around link and node failures.

• Since every packet must carry the full address of the destination, the overhead per packet is higher than for the connection-oriented model.

Datagram Model

Destination Port

A

B

C

D

E

F

G

H

3

0

3

3

2

1

0

0

Forwarding table for switch 2

Virtual Circuit Model• The virtual circuit model requires a virtual connection

from the source host to the destination to be set up before the connection.

• It is a two-stage process: connection setup and data transfer.

• Two approaches to establish connection state: permanent virtual circuit (PVC) by a network administrator and switched virtual circuit (SVC) by signalling.

• A entry in PVC contains:– An incoming interface for the incoming packets– A virtual circuit identifier (VCI)– An outgoing interface– A VCI for the outgoing packets

Virtual Circuit Model

VC

Table

Incoming Interface

Incoming VCI

Outgoing Interface

Outgoing VCI

Switch 1

Switch 2

Switch 3

2

3

0

5

11

7

1

0

3

11

7

4

Virtual circuit table entries for three switches

Virtual Circuit Switching• Explicit connection setup (and tear-down) phase• Subsequence packets follow same circuit• Sometimes called connection-oriented model

0

13

2

01 3

2

0

13

25 11

4

7

Switch 3

Host B

Switch 2

Host A

Switch 1

• Analogy: phone call

• Each switch maintains a VC table

Virtual Circuit Model• Typically wait full RTT for connection setup before sending

first data packet.

• While the connection request contains the full address for destination, each data packet contains only a small identifier, making the per-packet header overhead small.

• If a switch or a link in a connection fails, the connection is broken and a new one needs to be established.

• Connection setup provides an opportunity to reserve resources.

Virtual Circuit Model

• In a datagram network, each packet competes with other packet. In the virtual model, different quality of service (QoS) can be provided. QoS means some performance-related guarantee.

• Examples of virtual circuit technologies:– X.25 - packet-switching technology which was designed

for transmitting analog data such as voice conversations. – Frame Relay – construct virtual private network

(VPNs). – asynchronous transfer mode (ATM)

Source Routing

• All the information about network topology for switching is provided by the source host.

• Possible ways to implement source routing:– Place a number to each output of each switch in the

header.

– Put an ordered list of switch ports in the header and

rotate this list as Figure 3.7. • Source routing can be used in both datagram and

virtual networks. The Internet Protocol includes a source route option.

• Source routing suffers from a scaling problem.

Source Routing

0

13

2

0

1 3

2

0

13

2

0

13

2

3 0 1 3 01

30 1

Switch 3

Host B

Switch 2

Host A

Switch 1

Implementation and Performance

• A general-purpose workstation with a number of network interfaces

• A specialized switching device

Bridges and Extended LANs

• LANs have physical limitations (e.g., 2500m)

• Connect two or more LANs with a bridge– accept and forward strategy

• An Ethernet bridge can carry as 10n Mbps, where n is the number of port.

A

Bridge

B C

X Y Z

Port 1

Port 2

Learning Bridges • Do not forward when unnecessary• Maintain forwarding table

Host Port A 1 B 1 C 1 X 2 Y 2 Z 2

• Learn table entries based on source address• Table is an optimization; need not to be complete• Always forward broadcast frames

A

Bridge

B C

X Y Z

Port 1

Port 2

Spanning Tree Algorithm • Problem: loops in the previous design

• Bridges run a distributed spanning tree algorithm – select which bridges actively forward– developed by Radia Perlman– now IEEE 802.1 specification

B3

A

C

E

DB2

B5

B

B7 K

F

H

B4

J

B1

B6

G

I

Algorithm Overview • Each bridge has unique id (e.g., B1, B2, B3)• Select bridge with smallest id as root• Select bridge on each LAN closest to root as

designated bridge (use id to break ties)

B3

A

C

E

DB2

B5

B

B7 K

F

H

B4

J

B1

B6

G

I

• Each bridge forwards frames over each LAN for which it is the designated bridge

Algorithm Details

• Bridges exchange configuration messages– id for bridge sending the message

– id for what the sending bridge believes to be root bridge

– distance (hops) from sending bridge to root bridge

• Each bridge records current best configuration message for each port

• Initially, each bridge believes it is the root

Algorithm Detail (cont)• When learn not root, stop generating config messages

– in steady state, only root generates configuration messages

• When learn not designated bridge, stop forwarding config messages– in steady state, only designated bridges forward config messages

• Root continues to periodically send config messages• If any bridge does not receive config message after a period

of time, it starts generating config messages claiming to be the root

Broadcast and Multicast

• Forward all broadcast/multicast frames– current practice

• Each host in a multicast group must periodically send a frame with the address for the group in the source field of the frame header.

Limitations of Bridges

• Do not scale– The spanning tree algorithm does not scale

– Broadcast does not scale. It is not necessary to broadcast messages to all hosts in a large environment.

• Do not accommodate heterogeneity

• Caution: beware of transparency. Bridges might drop frames.

Cell Switching (ATM)

• Architecture Features

– Similarities between ATM and packet switching

– Transfer of data in discrete chunks

• Multiple logical connections over single physical interface

• In ATM flow on each logical connection is in fixed sized packets called cells

• Minimal error and flow control

– Reduced overhead

• Data rates (physical layer) 25.6Mbps to 622.08Mbps

Cell Switching (ATM)

• Connection-oriented packet-switched network• Used in both WAN and LAN settings• Signaling (connection setup) Protocol: Q.2931

– An ITU-T specification defining user-to-network interface signaling for Broadband ISDN.

– Discover a suitable route– Responsible for allocating resources at the switches

• The QoS capabilities of ATM are one of its greatest strengths.

Cell Switching (ATM)

• Two Addressing schemes– Public ATM networks use 8-octet format (E.164

standard)– Computers attached to private ATM network use 20-

octet Network Service Access Point (NSAP) address (ATM Forum)

• Packets are called cells – Fixed length 53 bytes– 5-byte header + 48-byte payload

• Commonly transmitted over SONET– other physical layers possible

Cell Switching (ATM)

• ATM media - Commonly transmitted over SONET– DS-1/T1

– NxDS-1

– DS-3

– Multi-mode fiber (155Mbps)

– SONET/SDH

– (622 Mbps)

ATM Network

workstation

LAN Switch

Router

UNIATM Switch

12

Variable vs. Fixed-Length Packets

• No Optimal Length– if small: high header-to-data overhead

– if large: low utilization for small messages

• Fixed-Length Easier to Switch in Hardware– simpler

– enables parallelism

Big vs Small Packets

• Small Improves Queue behavior – finer-grained pre-emption point for scheduling link

• maximum packet = 4KB• link speed = 100Mbps• transmission time = 4096 x 8/100 = 327.68us• high priority packet may sit in the queue 327.68us• in contrast, 53 x 8/100 = 4.24us for ATM

– near cut-through behavior • two 4KB packets arrive at same time• link idle for 327.68us while both arrive• at end of 327.68us, still have 8KB to transmit • in contrast, ATM can transmit first cell after 4.24us• at end of 327.68us, just over 4KB left in queue

Big vs. Small (cont)

• Small Improves Latency (for voice) – voice digitally encoded at 64KBps (8-bit samples at

8KHz)

– need full cell’s worth of samples before sending cell

– example: 1000-byte cells implies 125ms per cell (too long)

– smaller latency implies no need for echo cancellers

• ATM Compromise: 48 bytes = (32+64)/2

Cell Format• User-Network Interface (UNI)

– host-to-switch format (telephone companies and customers)– GFC: Generic Flow Control (still being defined)– VCI: Virtual Circuit Identifier– VPI: Virtual Path Identifier– Type: management, congestion control, AAL5 (later)– CLPL Cell Loss Priority – HEC: Header Error Check (CRC-8)

• Network-Network Interface (NNI)– switch-to-switch format (phone companies)– GFC becomes part of VPI field

GFC HEC (CRC-8)

4 16 3 18

VPI VCI CLPType Payload

384 (48 bytes)8

ATM Architecture

Presentation

Session

Network

Data Link

PhysicalTransmission-convergencephysical medium dependent

Transport

Application

ATM Layer

ATM Adaptation Layer(AAL)

1(CBR)

2(VBR)

3/4(SMDS)

5(Data)

SAAL

Upper Layer Protocols

ATM Adaptation layer

• Supports multiple-application operations

• Type of user payload is identified

• Maps higher layer information into ATM cell payload.

• Handle transmission errors

• Segmentation and re-assembly

• Handle lost and misinserted cells

• Flow control and timing

Transmission-convergencephysical medium dependent

ATM Layer

CS

1(CBR)

2(VBR)

3/4(SMDS)

5(Data)

SAAL

Upper Layer Protocols

SAR

ATM Adaptation Sub Layers

• Convergence Sublayer (CS)– Functions needed to support specific

applications using AAL

– AAL user attaches at SAP

• Segmentation and Reassembly(SAR)– Responsible for creating 48 byte payload for

ATM cells.

– Also unpacks cell payload data received from ATM layer for delivery up to CS sublayer

AAL Protocols and PDU

AAL Applications

• Support for information transfer protocol not based on ATM– PCM (voice)

• Assemble bits into cells

• Re-assemble into constant flow

– IP• Map IP packets onto ATM cells

• Fragment IP packets

• Use LAPF over ATM to retain all IP infrastructure

Supported Application Types

• Circuit emulation

• VBR voice and video

• General data service

• IP over ATM

• Multiprotocol encapsulation over ATM (MPOA)– IPX, AppleTalk, DECNET)

• LAN emulation

ATM Layer

• Responsible for ATM cell transmissions

• Maps network layer address to ATM address

Transmission-convergencephysical medium dependent

ATM Layer

Upper Layer Protocols

ATM Adaptation Layer

Physical Layer

Divided into two sublayers:

• Transmission Convergence– Synchronization of

transmission & reception

– Cell delineation

– Error control

• Physical Medium Dependent (PMD)– Specifies physical medium

used

physical medium dependent

ATM Layer

Upper Layer Protocols

ATM Adaptation Layer

Transmission-convergence

Segmentation and Reassembly • ATM Adaptation Layer (AAL)

– AAL 1 and 2 designed for applications that need guaranteed rate (e.g., voice, video)

– AAL 3/4 designed for packet data– AAL 5 is an alternative standard for packet data

AAL

ATM

AAL

ATM

… …

AAL 3/4

• Convergence Sublayer Protocol Data Unit (CS-PDU)

– CPI: commerce part indicator (version field)

– Btag/Etag:beginning and ending tag

– BAsize: hint on amount of buffer space to allocate

– Length: size of whole PDU

CPI Btag BASize Pad 0 Etag Len

8 16 0– 24 8 8 16< 64 KB8

User data

Cell Format

– Type• BOM: beginning of message

• COM: continuation of message

• EOM end of message

– SEQ: sequence of number

– MID: message id

– Length: number of bytes of PDU in this cell

ATM header Length CRC-10

40 2 4

SEQ MIDType Payload

352 (44 bytes)10 6 10

AAL5

• CS-PDU Format

– pad so trailer always falls at end of ATM cell

– Length: size of PDU (data only)

– CRC-32 (detects missing or misordered cells)

• Cell Format– end-of-PDU bit in Type field of ATM header

CRC-32

< 64 KB 0– 47 bytes 16 16

ReservedPad Len

32

Data