Switching Data arriving to an input port of a switch have to be - - PowerPoint PPT Presentation

SLIDE 1

Switching

An Engineering Approach to Computer Networking An Engineering Approach to Computer Networking

What is it all about?

 How do we move traffic from one part of the network to another?

How do we move traffic from one part of the network to another?

 Connect end-systems to switches, and switches to each other

Connect end-systems to switches, and switches to each other

 Data arriving to an input port of a switch have to be moved to

Data arriving to an input port of a switch have to be moved to

• ne or more of the output ports
• ne or more of the output ports

Types of switching elements

 Telephone switches

Telephone switches

 switch samples

switch samples

 Datagram routers

Datagram routers

 switch datagrams

switch datagrams

 ATM switches

ATM switches

 switch ATM cells

switch ATM cells

Classification

 Packet vs. circuit switches

Packet vs. circuit switches

 packets have headers and samples don

packets have headers and samples donʼt

 Connectionless vs. connection oriented

Connectionless vs. connection oriented

 connection oriented switches need a call setup

connection oriented switches need a call setup

 setup is handled in

setup is handled in control plane control plane by switch controller

 connectionless switches deal with self-contained datagrams

Connectionless (router) Connection-oriented (switching system) Packet switch Internet router ATM switching system Circuit switch Telephone switching system

Other switching element functions

 Participate in routing algorithms

Participate in routing algorithms

 to build routing tables

to build routing tables

 Resolve contention for output trunks

Resolve contention for output trunks

 scheduling

scheduling

 Admission control

Admission control

 to guarantee resources to certain streams

to guarantee resources to certain streams

 We

Weʼll discuss these later ll discuss these later

 Here we focus on pure data movement

Here we focus on pure data movement

Requirements

 Capacity of switch is the maximum rate at which it can move

Capacity of switch is the maximum rate at which it can move information, assuming all data paths are simultaneously active information, assuming all data paths are simultaneously active

 Primary goal:

Primary goal: maximize capacity maximize capacity

 subject to cost and reliability constraints

subject to cost and reliability constraints

 Circuit switch must reject call if can

Circuit switch must reject call if canʼt find a path for samples t find a path for samples from input to output from input to output

 goal:

goal: minimize call blocking minimize call blocking

 Packet switch must reject a packet if it can

Packet switch must reject a packet if it canʼt find a buffer to store t find a buffer to store it awaiting access to output trunk it awaiting access to output trunk

 goal:

goal: minimize packet loss minimize packet loss

 Don

Donʼt reorder t reorder packets packets

SLIDE 2

A generic switch Outline

 Circuit switching

Circuit switching

 Packet switching

Packet switching

 Switch generations

Switch generations

 Switch fabrics

Switch fabrics

 Buffer placement

Buffer placement

 Multicast switches

Multicast switches

Circuit switching

 Moving 8-bit samples from an input port to an output port

Moving 8-bit samples from an input port to an output port

 Recall that samples have no headers

Recall that samples have no headers

 Destination of sample depends on

Destination of sample depends on time time at which it arrives at the at which it arrives at the switch switch

 actually, relative order within a

actually, relative order within a frame frame

 We

Weʼll first study something simpler than a switch: a multiplexor ll first study something simpler than a switch: a multiplexor

Multiplexors and demultiplexors

 Most trunks time division multiplex voice samples

Most trunks time division multiplex voice samples

 At a central office, trunk is demultiplexed and distributed to

At a central office, trunk is demultiplexed and distributed to active circuits active circuits

 Synchronous multiplexor

Synchronous multiplexor

 N input lines

N input lines

 Output runs N times as fast as input

Output runs N times as fast as input

More on multiplexing

 Demultiplexor

Demultiplexor

 one input line and N outputs that run N times slower

• ne input line and N outputs that run N times slower

 samples are placed in output buffer in round robin order

samples are placed in output buffer in round robin order

 Neither multiplexor nor demultiplexor needs addressing

Neither multiplexor nor demultiplexor needs addressing information (why?) information (why?)

 Can cascade multiplexors

Can cascade multiplexors

 need a standard

need a standard

 example: DS hierarchy in the US and Japan

example: DS hierarchy in the US and Japan

Inverse multiplexing

 Takes a high bit-rate stream and scatters it across multiple

Takes a high bit-rate stream and scatters it across multiple trunks trunks

 At the other end, combines multiple streams

At the other end, combines multiple streams

 resequencing

resequencing to accommodate variation in delays to accommodate variation in delays

 Allows high-speed virtual links using existing technology

Allows high-speed virtual links using existing technology

SLIDE 3

A circuit switch

 A switch that can handle N calls has N logical inputs and N

A switch that can handle N calls has N logical inputs and N logical outputs logical outputs

 N up to 200,000

N up to 200,000

 In practice, input trunks are multiplexed

In practice, input trunks are multiplexed

 example: DS3 trunk carries 672 simultaneous calls

example: DS3 trunk carries 672 simultaneous calls

 Multiplexed trunks carry

Multiplexed trunks carry frames frames = set of samples = set of samples

 Goal: extract samples from frame, and depending on position in

Goal: extract samples from frame, and depending on position in frame, switch to output frame, switch to output

 each incoming sample has to get to the right output line and the

each incoming sample has to get to the right output line and the right slot in the output frame right slot in the output frame

 demultiplex

demultiplex, switch, multiplex , switch, multiplex

Call blocking

 Can

Canʼt find a path from input to output t find a path from input to output

 Internal blocking

Internal blocking

 slot in output frame exists, but no path

slot in output frame exists, but no path

 Output blocking

Output blocking

 no slot in output frame is available

no slot in output frame is available

 Output blocking is reduced in

Output blocking is reduced in transit transit switches switches

 need to put a sample in one of

need to put a sample in one of several several slots going to the desired next hop

Time division switching

 Key idea: when

Key idea: when demultiplexing demultiplexing, position in frame determines , position in frame determines

• utput trunk
• utput trunk

 Time division switching interchanges sample position within a

Time division switching interchanges sample position within a frame: time slot interchange (TSI) frame: time slot interchange (TSI)

How large a TSI can we build?

 Limit is time taken to read and write to memory

Limit is time taken to read and write to memory

 For 120,000 circuits

For 120,000 circuits

 need to read and write memory once every 125 microseconds

need to read and write memory once every 125 microseconds

 each operation takes around 0.5

each operation takes around 0.5 ns ns => impossible with current => impossible with current technology technology

 Need to look to other techniques

Need to look to other techniques

Space division switching

 Each sample takes a different path through the switch,

Each sample takes a different path through the switch, depending on its destination depending on its destination

Crossbar

 Simplest possible space-division switch

Simplest possible space-division switch

 Crosspoints

Crosspoints can be turned on or off

 For multiplexed inputs, need a switching schedule (why?)  Internally nonblocking  but need N2 crosspoints  time taken to set each crosspoint grows quadratically  vulnerable to single faults (why?)

SLIDE 4

Multistage crossbar

 In a crossbar during each switching time only one

In a crossbar during each switching time only one crosspoint crosspoint per per row or column is active row or column is active

 Can save crosspoints if a

Can save crosspoints if a crosspoint crosspoint can attach to more than can attach to more than

• ne input line (why?)
• ne input line (why?)

 This is done in a multistage crossbar

This is done in a multistage crossbar

 Need to rearrange connections every switching time

Need to rearrange connections every switching time

Multistage crossbar

 Can suffer internal blocking

Can suffer internal blocking

 unless sufficient number of second-level stages

unless sufficient number of second-level stages

 Number of crosspoints < N

Number of crosspoints < N2

 Finding a path from input to output requires a depth-first-search

Finding a path from input to output requires a depth-first-search

 Scales better than crossbar, but still not too well

Scales better than crossbar, but still not too well

 120,000 call switch needs ~250 million crosspoints

120,000 call switch needs ~250 million crosspoints

Time-space switching

 Precede each input trunk in a crossbar with a TSI

Precede each input trunk in a crossbar with a TSI

 Delay samples so that they arrive at the right time for the space

Delay samples so that they arrive at the right time for the space division switch division switchʼs schedule s schedule

Time-space-time (TST) switching

 Allowed to flip samples both on input and output trunk

Allowed to flip samples both on input and output trunk

 Gives more flexibility => lowers call blocking probability

Gives more flexibility => lowers call blocking probability

Outline

 Circuit switching

Circuit switching

 Packet switching

Packet switching

 Switch generations

Switch generations

 Switch fabrics

Switch fabrics

 Buffer placement

Buffer placement

 Multicast switches

Multicast switches

Packet switching

 In a circuit switch, path of a sample is determined at time of

In a circuit switch, path of a sample is determined at time of connection establishment connection establishment

 No need for a sample header--position in frame is enough

No need for a sample header--position in frame is enough

 In a packet switch, packets carry a destination field

In a packet switch, packets carry a destination field

 Need to look up destination port on-the-fly

Need to look up destination port on-the-fly

 Datagram

Datagram

 lookup based on entire destination address

lookup based on entire destination address

 Cell

Cell

 lookup based on VCI

lookup based on VCI

 Other than that, very similar

Other than that, very similar

SLIDE 5

Repeaters, bridges, routers, and gateways

 Repeaters: at physical level

Repeaters: at physical level

 Bridges: at datalink level (based on MAC addresses) (L2)

Bridges: at datalink level (based on MAC addresses) (L2)

 discover attached stations by listening

discover attached stations by listening

 Routers: at network level (L3)

Routers: at network level (L3)

 participate in routing protocols

participate in routing protocols

 Application level gateways: at application level (L7)

Application level gateways: at application level (L7)

 treat entire network as a single hop

treat entire network as a single hop

 e.g mail gateways and

e.g mail gateways and transcoders transcoders

 Gain functionality at the expense of forwarding speed

Gain functionality at the expense of forwarding speed

 for best performance, push functionality as low as possible

for best performance, push functionality as low as possible

Port mappers

 Look up output port based on destination address

Look up output port based on destination address

 Easy for VCI: just use a table

Easy for VCI: just use a table

 Harder for datagrams:

Harder for datagrams:

 need to find

need to find longest prefix match longest prefix match

 e.g. packet with address 128.32.1.20

e.g. packet with address 128.32.1.20

 entries: (128.32.*, 3), (128.32.1.*, 4), (128.32.1.20, 2)

entries: (128.32.*, 3), (128.32.1.*, 4), (128.32.1.20, 2)

 A standard solution:

A standard solution: trie trie

Tries

 Two ways to improve performance

Two ways to improve performance

 cache recently used addresses in a CAM

cache recently used addresses in a CAM

 move common entries up to a higher level (match longer strings)

move common entries up to a higher level (match longer strings)

Blocking in packet switches

 Can have both internal and output blocking

Can have both internal and output blocking

 Internal

Internal

 no path to output

no path to output

 Output

Output

 trunk unavailable

trunk unavailable

 Unlike a circuit switch, cannot predict if packets will block (why?)

Unlike a circuit switch, cannot predict if packets will block (why?)

 If packet is blocked, must either buffer or drop it

If packet is blocked, must either buffer or drop it

Dealing with blocking

 Overprovisioning

Overprovisioning

 internal links much faster than inputs

internal links much faster than inputs

 Buffers

Buffers

 at input or output

at input or output

 Backpressure

Backpressure

 if switch fabric doesn

if switch fabric doesnʼt have buffers, prevent packet from entering t have buffers, prevent packet from entering until path is available until path is available

 Parallel switch fabrics

Parallel switch fabrics

 increases effective switching capacity

increases effective switching capacity

Outline

 Circuit switching

Circuit switching

 Packet switching

Packet switching

 Switch generations

Switch generations

 Switch fabrics

Switch fabrics

 Buffer placement

Buffer placement

 Multicast switches

Multicast switches

SLIDE 6

Three generations of packet switches

 Different trade-

Different trade-offs

• ffs between cost and performance

between cost and performance

 Represent evolution in switching capacity, rather than in

Represent evolution in switching capacity, rather than in technology technology

 With same technology, a later generation switch achieves greater

With same technology, a later generation switch achieves greater capacity, but at greater cost capacity, but at greater cost

 All three generations are represented in current products

All three generations are represented in current products

First generation switch

 Most Ethernet switches and cheap packet routers

Most Ethernet switches and cheap packet routers

 Bottleneck can be CPU, host-

Bottleneck can be CPU, host-adaptor adaptor or I/O bus, depending

• r I/O bus, depending

Example

 First generation router built with 133 MHz Pentium

First generation router built with 133 MHz Pentium

 Mean packet size 500 bytes

Mean packet size 500 bytes

 Interrupt takes 10 microseconds, word access take 50

Interrupt takes 10 microseconds, word access take 50 ns ns

 Per-packet processing time takes 200 instructions = 1.504 µs

Per-packet processing time takes 200 instructions = 1.504 µs

 Copy loop

Copy loop

register <- memory[read_ register <- memory[read_ptr ptr] memory [write_ memory [write_ptr ptr] <- register ] <- register read_ read_ptr ptr <- read_ <- read_ptr ptr + 4 + 4 write_ write_ptr ptr <- write_ <- write_ptr ptr + 4 + 4 counter <- counter -1 counter <- counter -1 if (counter not 0) branch to top of loop if (counter not 0) branch to top of loop

 4 instructions + 2 memory accesses = 130.08

4 instructions + 2 memory accesses = 130.08 ns ns

 Copying packet takes 500/4 *130.08 = 16.26 µs; interrupt 10 µs

Copying packet takes 500/4 *130.08 = 16.26 µs; interrupt 10 µs

 Total time = 27.764 µs => speed is 144.1

Total time = 27.764 µs => speed is 144.1 Mbps Mbps

 Amortized interrupt cost balanced by routing protocol cost

Amortized interrupt cost balanced by routing protocol cost

Second generation switch

 Port mapping intelligence in line cards

Port mapping intelligence in line cards

 ATM switch guarantees hit in lookup cache

ATM switch guarantees hit in lookup cache

 Ipsilon

Ipsilon IP switching IP switching

 assume underlying ATM network

assume underlying ATM network

 by default, assemble packets

by default, assemble packets

 if detect a flow, ask upstream to send on a particular VCI, and

if detect a flow, ask upstream to send on a particular VCI, and install entry in port install entry in port mapper mapper => implicit signaling => implicit signaling

Third generation switches

 Bottleneck in second generation switch is the bus (or ring)

Bottleneck in second generation switch is the bus (or ring)

 Third generation switch provides parallel paths (fabric)

Third generation switch provides parallel paths (fabric)

Third generation (contd.)

 Features

Features

 self-routing fabric

self-routing fabric

 output buffer is a point of contention

• utput buffer is a point of contention

 unless we

unless we arbitrate arbitrate access to fabric access to fabric

 potential for unlimited scaling, as long as we can resolve contention

potential for unlimited scaling, as long as we can resolve contention for output buffer for output buffer

SLIDE 7

Outline

 Circuit switching

Circuit switching

 Packet switching

Packet switching

 Switch generations

Switch generations

 Switch fabrics

Switch fabrics

 Buffer placement

Buffer placement

 Multicast switches

Multicast switches

Switch fabrics

 Transfer data from input to output, ignoring scheduling and

Transfer data from input to output, ignoring scheduling and buffering buffering

 Usually consist of links and

Usually consist of links and switching elements switching elements

Crossbar

 Simplest switch fabric

Simplest switch fabric

 think of it as 2N buses in parallel

think of it as 2N buses in parallel

 Used here for

Used here for packet packet routing: routing: crosspoint crosspoint is left open long is left open long enough to transfer a packet from an input to an output enough to transfer a packet from an input to an output

 For fixed-size packets and known arrival pattern, can compute

For fixed-size packets and known arrival pattern, can compute schedule in advance schedule in advance

 Otherwise, need to compute a schedule on-the-fly (what does

Otherwise, need to compute a schedule on-the-fly (what does the schedule depend on?) the schedule depend on?)

Buffered crossbar

 What happens if packets at two inputs both want to go to same

What happens if packets at two inputs both want to go to same

• utput?
• utput?

 Can defer one at an input buffer

Can defer one at an input buffer

 Or, buffer crosspoints

Or, buffer crosspoints

Broadcast

 Packets are tagged with output port #

Packets are tagged with output port #

 Each output matches tags

Each output matches tags

 Need to match N addresses in parallel at each output

Need to match N addresses in parallel at each output

 Useful only for small switches, or as a stage in a large switch

Useful only for small switches, or as a stage in a large switch

Switch fabric element

 Can build complicated fabrics from a simple element

Can build complicated fabrics from a simple element

 Routing rule: if 0, send packet to upper output, else to lower

Routing rule: if 0, send packet to upper output, else to lower

• utput
• utput

 If both packets to same output, buffer or drop

If both packets to same output, buffer or drop

SLIDE 8

Features of fabrics built with switching elements

 NxN

NxN switch with switch with bxb bxb elements has elements with elements has elements with elements per stage elements per stage

 Fabric is

Fabric is self routing self routing

 Recursive

Recursive

 Can be synchronous or asynchronous

Can be synchronous or asynchronous

 Regular and suitable for VLSI implementation

Regular and suitable for VLSI implementation

! "

log bN

! " N b /

Banyan

 Simplest self-routing recursive fabric

Simplest self-routing recursive fabric

 (why does it work?)

(why does it work?)

 What if two packets both want to go to the same output?

What if two packets both want to go to the same output?

 output blocking

• utput blocking

Blocking

 Can avoid with a buffered

Can avoid with a buffered banyan banyan switch switch

 but this is too expensive

but this is too expensive

 hard to achieve zero loss even with buffers

hard to achieve zero loss even with buffers

 Instead, can check if path is available before sending packet

Instead, can check if path is available before sending packet

 three-phase scheme

three-phase scheme

 send requests

send requests

 inform winners

inform winners

 send packets

send packets

 Or, use several

Or, use several banyan banyan fabrics in parallel fabrics in parallel

 intentionally misroute and tag one of a colliding pair

intentionally misroute and tag one of a colliding pair

 divert tagged packets to a second

divert tagged packets to a second banyan banyan, and so on to k stages , and so on to k stages

 expensive

expensive

 can reorder packets

can reorder packets

 output buffers have to run k times faster than input

• utput buffers have to run k times faster than input

Sorting

 Can avoid blocking by choosing order in which packets appear

Can avoid blocking by choosing order in which packets appear at input ports at input ports

 If we can

If we can

 present packets at inputs sorted by output

present packets at inputs sorted by output

 remove duplicates

remove duplicates

 remove gaps

remove gaps

 precede

precede banyan banyan with a perfect shuffle stage with a perfect shuffle stage

 then no internal blocking

then no internal blocking

 For example, [X, 010, 010, X, 011, X, X, X] -(sort)->

For example, [X, 010, 010, X, 011, X, X, X] -(sort)-> [010, 011, 011, X, X, X, X, X] -(remove [010, 011, 011, X, X, X, X, X] -(remove dups dups)-> )-> [010, 011, X, X, X, X, X, X] -(shuffle)-> [010, 011, X, X, X, X, X, X] -(shuffle)-> [010, X, 011, X, X, X, X, X] [010, X, 011, X, X, X, X, X]

 Need sort, shuffle, and trap networks

Need sort, shuffle, and trap networks

Sorting

 Build sorters from merge networks

Build sorters from merge networks

 Assume we can merge two sorted lists

Assume we can merge two sorted lists

 Sort pairwise, merge,

Sort pairwise, merge, recurse recurse

Merging

SLIDE 9

Putting it together- Batcher Banyan

 What about trapped duplicates?

What about trapped duplicates?

 recirculate

recirculate to beginning to beginning

 or run output of trap to multiple

• r run output of trap to multiple banyans

banyans (dilation dilation)

Effect of packet size on switching fabrics

 A major motivation for small fixed packet size in ATM is ease of

A major motivation for small fixed packet size in ATM is ease of building large parallel fabrics building large parallel fabrics

 In general, smaller size => more per-packet overhead, but more

In general, smaller size => more per-packet overhead, but more preemption points/sec preemption points/sec

 At high speeds, overhead dominates!

At high speeds, overhead dominates!

 Fixed size packets helps build synchronous switch

Fixed size packets helps build synchronous switch

 But we could fragment at entry and reassemble at exit

But we could fragment at entry and reassemble at exit

 Or build an asynchronous fabric

Or build an asynchronous fabric

 Thus, variable size doesn

Thus, variable size doesnʼt hurt too much t hurt too much

 Maybe Internet routers can be almost as cost-effective as ATM

Maybe Internet routers can be almost as cost-effective as ATM switches switches

Outline

 Circuit switching

Circuit switching

 Packet switching

Packet switching

 Switch generations

Switch generations

 Switch fabrics

Switch fabrics

 Buffer placement

Buffer placement

 Multicast switches

Multicast switches

Buffering

 All packet switches need buffers to match input rate to service

All packet switches need buffers to match input rate to service rate rate

 or cause heavy packet loses

• r cause heavy packet loses

 Where should we place buffers?

Where should we place buffers?

 input

input

 in the fabric

in the fabric

 output

• utput

 shared

shared

Input buffering (input queueing)

 No speedup in buffers or trunks (unlike output queued switch)

No speedup in buffers or trunks (unlike output queued switch)

 Needs arbiter

Needs arbiter

 Problem:

Problem: head of line blocking head of line blocking

 with randomly distributed packets, utilization at most 58.6%

with randomly distributed packets, utilization at most 58.6%

 worse with

worse with hot spots hot spots

Dealing with HOL blocking

 Per-output queues at inputs

Per-output queues at inputs

 Arbiter must choose one of the input ports for each output port

Arbiter must choose one of the input ports for each output port

 How to select?

How to select?

 Parallel Iterated Matching

Parallel Iterated Matching

 inputs tell arbiter which outputs they are interested in

inputs tell arbiter which outputs they are interested in

 output selects one of the inputs

• utput selects one of the inputs

 some inputs may get more than one

some inputs may get more than one grant grant, others may get none , others may get none

 if >1 grant, input picks one at random, and tells output

if >1 grant, input picks one at random, and tells output

 losing inputs and outputs try again

losing inputs and outputs try again



Used in DEC Used in DEC Autonet Autonet 2 switch 2 switch

SLIDE 10

Output queueing

 Don

Donʼt suffer from head-of-line blocking t suffer from head-of-line blocking

 But output buffers need to run much faster than trunk speed

But output buffers need to run much faster than trunk speed (why?) (why?)

 Can reduce some of the cost by using the

Can reduce some of the cost by using the knockout knockout principle principle

 unlikely that all N inputs will have packets for the same output

unlikely that all N inputs will have packets for the same output

 drop extra packets, fairly distributing losses among inputs

drop extra packets, fairly distributing losses among inputs

Shared memory

 Route only the header to output port

Route only the header to output port

 Bottleneck is time taken to read and write

Bottleneck is time taken to read and write multiported multiported memory memory

 Doesn

Doesnʼt scale to large switches t scale to large switches

 But can form an element in a multistage switch

But can form an element in a multistage switch

Datapath: clever shared memory design

 Reduces read/write cost by doing wide reads and writes

Reduces read/write cost by doing wide reads and writes

 1.2

1.2 Gbps Gbps switch for $50 parts cost switch for $50 parts cost

Buffered fabric

 Buffers in each switch element

Buffers in each switch element

 Pros

Pros

 Speed up is only as much as fan-in

Speed up is only as much as fan-in

 Hardware

Hardware backpressure backpressure reduces buffer requirements reduces buffer requirements

 Cons

Cons

 costly (unless using single-chip switches)

costly (unless using single-chip switches)

 scheduling is hard

scheduling is hard

Hybrid solutions

 Buffers at more than one point

Buffers at more than one point

 Becomes hard to analyze and manage

Becomes hard to analyze and manage

 But common in practice

But common in practice

Outline

 Circuit switching

Circuit switching

 Packet switching

Packet switching

 Switch generations

Switch generations

 Switch fabrics

Switch fabrics

 Buffer placement

Buffer placement

 Multicast switches

Multicast switches

SLIDE 11

Multicasting

 Useful to do this in hardware

Useful to do this in hardware

 Assume

Assume portmapper portmapper knows list of outputs knows list of outputs

 Incoming packet must be copied to these output ports

Incoming packet must be copied to these output ports

 Two

Two subproblems subproblems

 generating and distributing copies

generating and distributing copies

 VCI translation for the copies

VCI translation for the copies

Generating and distributing copies

 Either implicit or explicit

Either implicit or explicit

 Implicit

Implicit

 suitable for bus-based, ring-based, crossbar, or broadcast switches

suitable for bus-based, ring-based, crossbar, or broadcast switches

 multiple outputs enabled after placing packet on shared bus

multiple outputs enabled after placing packet on shared bus

 used in Paris and

used in Paris and Datapath Datapath switches switches

 Explicit

Explicit

 need to copy a packet at switch elements

need to copy a packet at switch elements

 use a

use a copy copy network network

 place # of copies in tag

place # of copies in tag

 element copies to both outputs and decrements count on one of

element copies to both outputs and decrements count on one of them them

 collect copies at outputs

collect copies at outputs

 Both schemes increase blocking probability

Both schemes increase blocking probability

Header translation

 Normally, in-VCI to out-VCI translation can be done either at

Normally, in-VCI to out-VCI translation can be done either at input or output input or output

 With multicasting, translation easier at output port (why?)

With multicasting, translation easier at output port (why?)

 Use separate port mapping and translation tables

Use separate port mapping and translation tables

 Input maps a VCI to a set of output ports

Input maps a VCI to a set of output ports

 Output port swaps VCI

Output port swaps VCI

 Need to do two lookups per packet

Need to do two lookups per packet