Multiple Access

An Engineering Approach to Computer Networking

What is it all about?
■ Consider an audioconference where
◆ if one person speaks, all can hear
◆ if more than one person speaks at the same time, both voices are
garbled
■ How should participants coordinate actions so that
◆ the number of messages exchanged per second is maximized
◆ time spent waiting for a chance to speak is minimized
■ This is the multiple access problem
Some simple solutions
■ Use a moderator
◆ a speaker must wait for moderator to call on him or her, even if no
one else wants to speak
◆ what if the moderator’s connection breaks?
■ Distributed solution
◆ speak if no one else is speaking
◆ but if two speakers are waiting for a third to finish, guarantee
collision
■ Designing good schemes is surprisingly hard!
Outline
■ Contexts for the problem
■ Choices and constraints
■ Performance metrics
■ Base technologies
■ Centralized schemes
■ Distributed schemes
Contexts for the multiple access problem
■ Broadcast transmission medium
◆ message from any transmitter is received by all receivers
■ Colliding messages are garbled
■ Goal
◆ maximize message throughput
◆ minimize mean waiting time
■ Shows up in five main contexts
Contexts
Contexts
Solving the problem
■ First, choose a base technology
◆ to isolate traffic from different stations
◆ can be in time domain or frequency domain
■ Then, choose how to allocate a limited number of transmission
resources to a larger set of contending users
Outline
■ Contexts for the problem
■ Choices and constraints
■ Performance metrics
■ Base technologies
■ Centralized schemes
■ Distributed schemes
Choices
■ Centralized vs. distributed design
◆ is there a moderator or not?
◆ in a centralized solution one of the stations is a master and the
others are slaves
✦ master->slave = downlink

✦ slave->master = uplink

◆ in a distributed solution, all stations are peers

■ Circuit-mode vs. packet-mode
◆ do stations send steady streams or bursts of packets?
◆ with streams, doesn’t make sense to contend for every packet
◆ allocate resources to streams
◆ with packets, makes sense to contend for every packet to avoid
wasting bandwidth
Constraints
■ Spectrum scarcity
◆ radio spectrum is hard to come by
◆ only a few frequencies available for long-distance communication
◆ multiple access schemes must be careful not to waste bandwidth
■ Radio link properties
◆ radio links are error prone
✦ fading

✦ multipath interference

◆ hidden terminals
✦ transmitter heard only by a subset of receivers

◆ capture
✦ on collision, station with higher power overpowers the other

✦ lower powered station may never get a chance to be heard

The parameter ‘a’
■ The number of packets sent by a source before the farthest
station receives the first bit
Outline
■ Contexts for the problem
■ Choices and constraints
■ Performance metrics
■ Base technologies
■ Centralized schemes
■ Distributed schemes
Performance metrics
■ Normalized throughput
◆ fraction of link capacity used to carry non-retransmitted packets
◆ example
✦ with no collisions, 1000 packets/sec

✦ with a particular scheme and workload, 250 packets/sec

✦ => goodput = 0.25

■ Mean delay
◆ amount of time a station has to wait before it successfully transmits
a packet
✦ depends on the load and the characteristics of the medium
Performance metrics
■ Stability
◆ with heavy load, is all the time spent on resolving contentions?
◆ => unstable
◆ with a stable algorithm, throughput does not decrease with offered
load
◆ if infinite number of uncontrolled stations share a link, then
instability is guaranteed
◆ but if sources reduce load when overload is detected, can achieve
stability
■ Fairness
◆ no single definition
◆ ‘no-starvation’: source eventually gets a chance to send
◆ max-min fair share: will study later
Outline
■ Contexts for the problem
■ Choices and constraints
■ Performance metrics
■ Base technologies
■ Centralized schemes
■ Distributed schemes
Base technologies
■ Isolates data from different sources
■ Three basic choices
◆ Frequency division multiple access (FDMA)
◆ Time division multiple access (TDMA)
◆ Code division multiple access (CDMA)
FDMA
■ Simplest
■ Best suited for analog links
■ Each station has its own frequency band, separated by guard
bands
■ Receivers tune to the right frequency
■ Number of frequencies is limited
◆ reduce transmitter power; reuse frequencies in non-adjacent cells
◆ example: voice channel = 30 KHz
◆ 833 channels in 25 MHz band
◆ with hexagonal cells, partition into 118 channels each
◆ but with N cells in a city, can get 118N calls => win if N > 7
TDMA
■ All stations transmit data on same frequency, but at different
times
■ Needs time synchronization
■ Pros
◆ users can be given different amounts of bandwidth
◆ mobiles can use idle times to determine best base station
◆ can switch off power when not transmitting
■ Cons
◆ synchronization overhead
◆ greater problems with multipath interference on wireless links
CDMA
■ Users separated both by time and frequency
■ Send at a different frequency at each time slot (frequency
hopping)
■ Or, convert a single bit to a code (direct sequence)
◆ receiver can decipher bit by inverse process
■ Pros
◆ hard to spy
◆ immune from narrowband noise
◆ no need for all stations to synchronize
◆ no hard limit on capacity of a cell
◆ all cells can use all frequencies
CDMA
■ Cons
◆ implementation complexity
◆ need for power control
✦ to avoid capture

◆ need for a large contiguous frequency band (for direct sequence)

◆ problems installing in the field
FDD and TDD
■ Two ways of converting a wireless medium to a duplex channel
■ In Frequency Division Duplex, uplink and downlink use different
frequencies
■ In Time Division Duplex, uplink and downlink use different time
slots
■ Can combine with FDMA/TDMA
■ Examples
◆ TDD/FDMA in second-generation cordless phones
◆ FDD/TDMA/FDMA in digital cellular phones
Outline
■ Contexts for the problem
■ Choices and constraints
■ Performance metrics
■ Base technologies
■ Centralized schemes
■ Distributed schemes
Centralized access schemes
■ One station is master, and the other are slaves
◆ slave can transmit only when master allows
■ Natural fit in some situations
◆ wireless LAN, where base station is the only station that can see
everyone
◆ cellular telephony, where base station is the only one capable of
high transmit power
Centralized access schemes
■ Pros
◆ simple
◆ master provides single point of coordination
■ Cons
◆ master is a single point of failure
✦ need a re-election protocol

✦ master is involved in every single transfer => added delay

Circuit mode
■ When station wants to transmit, it sends a message to master
using packet mode
■ Master allocates transmission resources to slave
■ Slave uses the resources until it is done
■ No contention during data transfer
■ Used primarily in cellular phone systems
◆ EAMPS: FDMA
◆ GSM/IS-54: TDMA
◆ IS-95: CDMA
Polling and probing
■ Centralized packet-mode multiple access schemes
■ Polling
◆ master asks each station in turn if it wants to send (roll-call polling)
◆ inefficient if only a few stations are active, overhead for polling
messages is high, or system has many terminals
■ Probing
◆ stations are numbered with consecutive logical addresses
◆ assume station can listen both to its own address and to a set of
multicast addresses
◆ master does a binary search to locate next active station
Reservation-based schemes
■ When ‘a’ is large, can’t use a distributed scheme for packet
mode (too many collisions)
◆ mainly for satellite links
■ Instead master coordinates access to link using reservations
■ Some time slots devoted to reservation messages
◆ can be smaller than data slots => minislots
■ Stations contend for a minislot (or own one)
■ Master decides winners and grants them access to link
■ Packet collisions are only for minislots, so overhead on
contention is reduced
Outline
■ Contexts for the problem
■ Choices and constraints
■ Performance metrics
■ Base technologies
■ Centralized schemes
■ Distributed schemes
Distributed schemes
■ Compared to a centralized scheme
◆ more reliable
◆ have lower message delays
◆ often allow higher network utilization
◆ but are more complicated
■ Almost all distributed schemes are packet mode (why?)
Decentralized polling
■ Just like centralized polling, except there is no master
■ Each station is assigned a slot that it uses
◆ if nothing to send, slot is wasted
■ Also, all stations must share a time base
Decentralized probing
■ Also called tree based multiple access
■ All stations in left subtree of root place packet on medium
■ If a collision, root <- root ->left_son, and try again
■ On success, everyone in root->right_son places a packet etc.
■ (If two nodes with successive logical addresses have a packet
to send, how many collisions will it take for one of them to win
access?)
■ Works poorly with many active stations, or when all active
stations are in the same subtree
Carrier Sense Multiple Access (CSMA)
■ A fundamental advance: check whether the medium is active
before sending a packet (i.e carrier sensing)
■ Unlike polling/probing a node with something to send doesn’t
have to wait for a master, or for its turn in a schedule
■ If medium idle, then can send
■ If collision happens, detect and resolve
■ Works when ‘a’ is small
Simplest CSMA scheme
■ Send a packet as soon as medium becomes idle
■ If, on sensing busy, wait for idle -> persistent
■ If, on sensing busy, set a timer and try later -> non-persistent
■ Problem with persistent: two stations waiting to speak will collide
How to solve the collision problem
■ Two solutions
■ p-persistent: on idle, transmit with probability p:
◆ hard to choose p
◆ if p small, then wasted time
◆ if p large, more collisions
■ exponential backoff
◆ on collision, choose timeout randomly from doubled range
◆ backoff range adapts to number of contending stations
◆ no need to choose p
◆ need to detect collisions: collision detect circuit => CSMA/CD
Summary of CSMA schemes
Ethernet
■ The most widely used LAN
■ Standard is called IEEE 802.3
■ Uses CSMA/CD with exponential backoff
■ Also, on collision, place a jam signal on wire, so that all stations
are aware of collision and can increment timeout range
■ ‘a’ small =>time wasted in collision is around 50 microseconds
■ Ethernet requires packet to be long enough that a collision is
detected before packet transmission completes (a <= 1)
◆ packet should be at least 64 bytes long for longest allowed
segment
■ Max packet size is 1500 bytes
◆ prevents hogging by a single station
More on Ethernet
■ First version ran at 3 Mbps and used ‘thick’ coax
■ These days, runs at 10 Mbps, and uses ‘thin’ coax, or twisted
pair (Category 3 and Category 5)
■ Ethernet types are coded as <Speed><Baseband or
broadband><physical medium>
◆ Speed = 3, 10, 100 Mbps
◆ Baseband = within building, broadband = on cable TV
◆ Physical medium:
✦ “2” is cheap 50 Ohm cable, upto 185 meters

✦ “T” is unshielded twisted pair (also used for telephone wiring)

✦ “36” is 75 Ohm cable TV cable, upto 3600 meters

Recent developments
■ Switched Ethernet
◆ each station is connected to switch by a separate UTP wire
◆ line card of switch has a buffer to hold incoming packets
◆ fast backplane switches packet from one line card to others
◆ simultaneously arriving packets do not collide (until buffers
overflow)
◆ higher intrinsic capacity than 10BaseT (and more expensive)
Fast Ethernet variants
■ Fast Ethernet (IEEE 802.3u)
◆ same as 10BaseT, except that line speed is 100 Mbps
◆ spans only 205 m
◆ big winner
◆ most current cards support both 10 and 100 Mbps cards (10/100
cards) for about $80
■ 100VG Anylan (IEEE 802.12)
◆ station makes explicit service requests to master
◆ master schedules requests, eliminating collisions
◆ not a success in the market
■ Gigabit Ethernet
◆ aims to continue the trend
◆ still undefined, but first implementation will be based on fiber links
Evaluating Ethernet
■ Pros
◆ easy to setup
◆ requires no configuration
◆ robust to noise
■ Problems
◆ at heavy loads, users see large delays because of backoff
◆ nondeterministic service
◆ doesn’t support priorities
◆ big overhead on small packets
■ But, very successful because
◆ problems only at high load
◆ can segment LANs to reduce load
CSMA/CA
■ Used in wireless LANs
■ Can’t detect collision because transmitter overwhelms colocated
receiver
■ So, need explicit acks
■ But this makes collisions more expensive
◆ => try to reduce number of collisions
CSMA/CA algorithm
■ First check if medium is busy
■ If so, wait for medium to become idle
■ Wait for interframe spacing
■ Set a contention timer to an interval randomly chosen in the
range [1, CW]
■ On timeout, send packet and wait for ack
■ If no ack, assume packet is lost
◆ try again, after doubling CW
■ If another station transmits while counting down, freeze CW and
unfreeze when packet completes transmission
■ (Why does this scheme reduce collisions compared to
CSMA/CD?)
Dealing with hidden terminals
■ CSMA/CA works when every station can receive transmissions
from every other station
■ Not always true
■ Hidden terminal
◆ some stations in an area cannot hear transmissions from others,
though base can hear both
■ Exposed terminal
◆ some (but not all) stations can hear transmissions from stations not
in the local area
Dealing with hidden and exposed terminals
■ In both cases, CSMA/CA doesn’t work
◆ with hidden terminal, collision because carrier not detected
◆ with exposed terminal, idle station because carrier incorrectly
detected
■ Two solutions
■ Busy Tone Multiple Access (BTMA)
◆ uses a separate “busy-tone” channel
◆ when station is receiving a message, it places a tone on this
channel
◆ everyone who might want to talk to a station knows that it is busy
✦ even if they cannot hear transmission that that station hears

◆ this avoids both problems (why?)

Multiple Access Collision Avoidance
■ BTMA requires us to split frequency band
◆ more complex receivers (need two tuners)
■ Separate bands may have different propagation characteristics
◆ scheme fails!
■ Instead, use a single frequency band, but use explicit messages
to tell others that receiver is busy
■ In MACA, before sending data, send a Request to Sent (RTS) to
intended receiver
■ Station, if idle, sends Clear to Send (CTS)
■ Sender then sends data
■ If station overhears RTS, it waits for other transmission to end
■ (why does this work?)
Token passing
■ In distributed polling, every station has to wait for its turn
■ Time wasted because idle stations are still given a slot
■ What if we can quickly skip past idle stations?
■ This is the key idea of token ring
■ Special packet called ‘token’ gives station the right to transmit
data
■ When done, it passes token to ‘next’ station
◆ => stations form a logical ring
■ No station will starve
Logical rings
■ Can be on a non-ring physical topology
Ring operation
■ During normal operation, copy packets from input buffer to
output
■ If packet is a token, check if packets ready to send
■ If not, forward token
■ If so, delete token, and send packets
■ Receiver copies packet and sets ‘ack’ flag
■ Sender removes packet and deletes it
■ When done, reinserts token
■ If ring idle and no token for a long time, regenerate token
Single and double rings
■ With a single ring, a single failure of a link or station breaks the
network => fragile
■ With a double ring, on a failure, go into wrap mode
■ Used in FDDI
Hub or star-ring
■ Simplifies wiring
■ Active hub is predecessor and successor to every station
◆ can monitor ring for station and link failures
■ Passive hub only serves as wiring concentrator
◆ but provides a single test point
■ Because of these benefits, hubs are practically the only form of
wiring used in real networks
◆ even for Ethernet
Evaluating token ring
■ Pros
◆ medium access protocol is simple and explicit
◆ no need for carrier sensing, time synchronization or complex
protocols to resolve contention
◆ guarantees zero collisions
◆ can give some stations priority over others
■ Cons
◆ token is a single point of failure
✦ lost or corrupted token trashes network

✦ need to carefully protect and, if necessary, regenerate token

◆ all stations must cooperate

✦ network must detect and cut off unresponsive stations

◆ stations must actively monitor network

✦ usually elect one station as monitor
Fiber Distributed Data Interface
■ FDDI is the most popular token-ring base LAN
■ Dual counterrotating rings, each at 100 Mbps
■ Uses both copper and fiber links
■ Supports both non-realtime and realtime traffic
◆ token is guaranteed to rotate once every Target Token Rotation
Time (TTRT)
◆ station is guaranteed a synchronous allocation within every TTRT
■ Supports both single attached and dual attached stations
◆ single attached (cheaper) stations are connected to only one of the
rings
ALOHA and its variants
■ ALOHA is one of the earliest multiple access schemes
■ Just send it!
■ Wait for an ack
■ If no ack, try again after a random waiting time
◆ no backoff
Evaluating ALOHA
■ Pros
◆ useful when ‘a’ is large, so carrier sensing doesn’t help
✦ satellite links

◆ simple
✦ no carrier sensing, no token, no timebase synchronization

◆ independent of ‘a’
■ Cons
◆ under some mathematical assumptions, goodput is at most .18
◆ at high loads, collisions are very frequent
◆ sudden burst of traffic can lead to instability
✦ unless backoff is exponential
Slotted ALOHA
■ A simple way to double ALOHA’s capacity
■ Make sure transmissions start on a slot boundary
■ Halves window of vulnerability
■ Used in cellular phone uplink
ALOHA schemes summarized
Reservation ALOHA
■ Combines slot reservation with slotted ALOHA
■ Contend for reservation minislots using slotted ALOHA
■ Stations independently examine reservation requests and come
to consistent conclusions
■ Simplest version
◆ divide time into frames = fixed length set of slots
◆ station that wins access to a reservation minislot using S-ALOHA
can keep slot as long as it wants
◆ station that loses keeps track of idle slots and contends for them in
next frame
Evaluating R-ALOHA
■ Pros
◆ supports both circuit and packet mode transfer
◆ works with large ‘a’
◆ simple
■ Cons
◆ arriving packet has to wait for entire frame before it has a chance to
send
◆ cannot preempt hogs
◆ variants of R-ALOHA avoid these problems
■ Used for cable-modem uplinks