Chapter 3 Switching and Forwarding

1
Chapter 3 - Switching and Forwarding
2
Problems

• 2500m limit on length of network
• Contention, collisions
• Limited number of hosts
• Communicate between hosts that do not have a
  direct connection

3
Scalable Networks

• Switch Forwards packets from input port to
  output port
• port selected based on destination address in
  packet header.

4
Advantages

• Can build networks that cover large geographic
  area
• Can build networks that support large numbers of
  hosts
• Can add new hosts without affecting performance
  of existing hosts

5
Source Routing

• Address contains sequence of ports on path from
  source to destination.

2 1 1
Where does it go?
6
Source Routing

• Source host must know entire topology
• Variable size header causes switches to be slow
• Approaches include rotation, stripping, pointers
• Impractical for large networks

7
Rotation, stripping and pointer
8
Virtual Circuit Switching

• Explicit connection setup (and tear-down) phase
• Subsequent packets follow same circuit
• Analogy phone call
• Sometimes called connection-oriented model
• Each switch maintains a VC table.
• Each circuit can be identified and handled
  differently (buffers, QoS)

9
Virtual Circuit Switching
10
Virtual Circuits

• ATM
• X.25
• Allocates buffers before connecting
• Frame Relay

11
Datagrams

• No connection setup phase
• Each packet forwarded independently
• Analogy postal system
• Sometimes called connectionless model
• Each switch maintains a forwarding (routing)
  table

12
Datagrams
13
Virtual Circuit versus Datagram

• 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.

14
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.
• Since packets are treated independently, 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.

15
Congestion vs Contention

• Contention occurs when multiple packets are
  queued on same output port
• Congestion when switch runs out of buffers and
  must drop a packet
• Virtual Circuits avoid congestion by negotiating
  on setup
• Datagrams use buffers more efficiently

16
Performance

• Switches can be built from a general-purpose
  workstations will consider special-purpose
  hardware later.

17
Performance

• Aggregate bandwidth
• 1/2 of the I/O bus bandwidth
• capacity is shared among all hosts connected to
  switch
• example 800Mbps bus can support 8 T3 ports
• Packets-per-second
• must be able to switch small packets
• 15,000 packets-per-second is an achievable number
• example 64-byte packets implies 7.69Mbps

18
Bridges and Extended LANs
19
Overview

• LANs have physical limitations (e.g., 2500m
  Ethernet)
• Connect two or more LANs with a bridge
• accept and forward strategy
• level 2 connection (does not add packet header)
• Collection of LANs connected by bridges called an
  exteneded LAN

20
Problems

• Heterogeneity
• Different Host types
• Different protocols
• Different hardware medium (FDDI, Ethernet)
• Scale
• Internet has doubled in size each year for the
  last 20 years

21
Bridges vs Switches

• A Bridge is a link-level node that forwards
  frames
• Connect two or more multi-access networks
• A Switch is a multi-input, multi-output device,
  which transfers packets from an input to one or
  more outputs.
• So, a Bridge meets the definition of a switch,
  but some switches dont use spanning tree and
  arent bridges

22
Seven Layer Model
Application Presentation Session Transport
Some Switches add network header for routing
information)
Routing address framing errors electrical signal
s
Network Data Link Physical
Bridge (Uses spanning tree to learn routing
information
23
Learning Bridges

• Do not forward when unnecessary
• Maintain forwarding table
• Learn table entries based on source address
• Table is an optimization need not be complete
• Always forward broadcast frames

Host A B C X Y Z
Port 1 1 1 2 2 2
24
Algorithm

• If the address is in the tables then
• Forward the packet onto the necessary port.
• If the address is not in the tables, then
• Forward the packet onto every port except for the
  port that the packet was received on, just to
  make sure the destination gets the message.
• Add an entry in your internal tables linking the
  Source Address of the packet to whatever port the
  packet was received from.

25
Bridging Functionality
26
Bridging Loops
27
Spanning Tree Algorithm

• Extended LANs sometimes have loops

28
Example
B1
3
4
G
H
0
7
8
B6
B4
1
6
I
J
29
Spanning Tree

• Bridges run a distributed spanning tree algorithm
  
• Select which bridges actively forward frames
• Developed by Radia Perlman at DEC
• Now IEEE 802.1 specification

30
Algorithm Overview

• Each bridge has a unique id (e.g., B1, B2, B3)
• Select bridge with smallest id as root
• Select bridge on each LAN that is closest to the
  root as that LAN's designated bridge (use id to
  break ties)
• Each bridge forwards frames over each LAN for
  which it is the designated bridge

31
Example
32
Algorithm Detail

• 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
• When learn not root, stop generating
  configuration message
• in steady state, only root generates
  configuration messages

33
Continued...

• When learn not designated bridge, stop forwarding
  configuration messages
• in steady state, only designated bridges forward
  configuration messages
• Root bridge continues to send configuration
  messages periodically
• If any given bridge does not receive
  configuration message after a period of time,
  starts generating configuration messages claiming
  to be the root

34
Example
A
(3,3,0)
B
B3
(5,5,0)
(7,7,0)
(3,3,0)
(5,5,0)
C
B5
(2,2,0)
(5,5,0)
D
B7
K
B2
(7,7,0)
(2,2,0)
(7,7,0)
E
F
(1,1,0)
(1,1,0)
(1,1,0)
B1
(1,1,0)
G
H
(6,6,0)
(4,4,0)
(4,4,0)
B6
B4
(6,6,0)
(4,4,0)
I
J
35
Next Phase
A
(3,2,1)
B
B3
(5,1,1)
(7,1,1)
(5,1,1)
C
B5
(2,1,1)
D
B7
K
B2
(7,1,1)
E
F
(1,1,0)
(1,1,0)
(1,1,0)
B1
(1,1,0)
G
H
(4,1,1)
B6
B4
(6,1,1)
(4,1,1)
I
J
36
Next Phase
A
B
B3
(5,1,1)
(5,1,1)
C
B5
(2,1,1)
(7,1,1)
D
B7
K
B2
E
F
(1,1,0)
(1,1,0)
(1,1,0)
B1
(1,1,0)
G
H
(4,1,1)
B6
B4
(4,1,1)
I
J
37
Broadcast and Multicast

• Forward all broadcast/multicast frames (current
  practice)
• Learn when no group members downstream
• Accomplished by having each member of group G
  send a frame to bridge multicast address with G
  in source field

38
Limitations of Bridges

• Do not scale
• spanning tree algorithm does not scale
• broadcast does not scale
• Do not accommodate heterogeneity
• Since dont have own header, cant join
  dissimilar networks, address formats
• Caution beware of transparency
• Congestion can cause dropped frames
• Additional delay
• frame reordering

39
Asynchronous Transfer Mode (ATM)
40
Overview

• Connection-oriented packet-switched network
• Used in both WAN and LAN settings
• Signaling (connection setup) Protocol Q.2931
• Specified by ATM Forum
• Packets are called cells 5-byte header 48-byte
  payload
• Commonly transmitted over SONET (but not
  necessarily)

41
Cells

• Variable versus Fixed-Length
• no optimal fixed-length
• if small high header-to-data overhead
• if large low utilization for small messages
• fixed-length are easier to switch in hardware
• simpler
• enables parallelism

Transmission Cable
5 Bytes
48 Bytes
Data Payload
Header
42
Small size improves queue behavior

• finer-grained pre-emption point for scheduling
  link
• maximum packet 4KB
• link speed 100Mbps
• transmission time 4096 x 8 / 100 327.68 ? s
• high priority packet may sit in the queue 327.68
  ? s
• in contrast, 53 x 8 / 100 4.24 ?s for ATM

43
near cut-through behavior

• two 4KB packets arrive at same time
• link idle for 327.68?s while both arrive
• at end of 327.68 ? s, still have 8KB to transmit
• in contrast, can transmit first cell after 4.24 ?
  s
• at end of 327.68 ? s, just over 4KB left in queue

44
Carrying Voice in Cells

• 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
  cancellation
• Settled on compromise of 48 bytes (3264)/2

45
Cell Format

• User-Network Interface (UNI)
• host-to-switch format
• GFC Generic Flow Control (still being defined)
• VCI Virtual Circuit Identifier
• VPI Virtual Path Identifier
• Type management, congestion control, AAL5
  (later)
• CLP Cell Loss Priority
• HEC Header Error Check (CRC-8)

46
Network-Network Interface (NNI)

• switch-to-switch format
• GFC becomes part of VPI field

header
payload
Fixed length packet cell
47
What is ATM?
Conventional LAN
Conventional Telecom
ATM
Traffic Type Transmission Unit Switching Connec
tion Type Delivery Access Rate Media
Data Variable Packet Packet Connectionless Bes
t Effort Shared Protocol Dependent
Voice Fixed Frame Circuit Connection-oriented
Guaranteed Dedicated Channel Dependent
Data, Voice, Video Fixed Cell Cell Connection-o
riented Defined Classes Dedicated Application
Dependent
48
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

49
AAL 3/4

• Convergence Sublayer Protocol Data Unit (CS-PDU)
• CPI common part indicator (version field)
• Btag/Etag beginning and ending tag
• BAsize hint on amount of buffer space to
  allocate
• Length size of whole PDU

50
Cell Format

• Type
• BOM beginning of message
• COM continuation of message
• EOM end of message
• SEQ sequence number
• MID message id
• Length number of bytes of PDU in this cell

51
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

52
VPI/VCI

• Host treat as 24-bit circuit identifier
• if cheap one-per application use for
  demultiplexing
• if expensive multiplex several applications onto
  one VCI
• Network aggregate multiple circuits into one path

53
Switching Hardware
54
Overview

• Terminology n x m switch has n inputs and m
  outputs
• Design Goals
• throughput (depends on traffic model)
• scalability (a function of n)
• Ports and Fabrics
• ports
• circuit management (e.g., map VCIs, route
  datagrams)
• buffering (input and/or output)
• fabric
• as simple as possible
• sometimes do buffering (internal)

55
Switching Fabric
56
Buffering

• Wherever contention is possible
• input port (contend for fabric)
• internal (contend for output port)
• output port (contend for link)
• Head-of-Line Blocking
• input buffering

57
Crossbar Switches
58
Knockout Switch

• Example of Crossbar
• Concentrator select one of n packets
• Complexity n2

59
Knockout Switch

• Output Buffers

60
Self-Routing Fabrics

• Banyan Network
• constructed from simple 2 x 2 switching elements
• self-routing header attached to each packet
• elements arranged to route based on this header
• no collisions if input packets sorted into
  ascending order
• complexity n log2 n

61
Banyan Network
010
010
62
Contention
1010
1000
63
Shared Media Switches