Multiple Access Protocols

1
Multiple Access Protocols

• Dr. R. K. Rao

2
Multiple Access Protocols

• Many algorithms exist for allocation of Multiple
  Access Channel. To begin with let us investigate
  representative algorithms
• Pure ALOHA
• Slotted ALOHA
• Reservation ALOHA

3
Pure ALOHA

• In 1970s, Norman Abramson and his team at
  University of Hawaii devised the algorithm
• The basic idea used in the algorithm is
  applicable to any system in which uncoordinated
  users are competing for the use of single shared
  channel
• The algorithm is referred to as Random Multiple
  Access Protocol or Pure ALOHA

4
Random Multiple Access Protocol

• Users transmit whenever they have data to be sent
• There will be collisions and the colliding frames
  are destroyed
• However, due to the broadcasting nature of the
  channel, a sender can always find out whether or
  not its frame was destroyed by listening to the
  channel, the same way other users do

5
Random Multiple Access Protocol

• With a LAN, the feedback is immediate
• However, with a satellite, there is a delay of
  270 ms, before the sender knows if the
  transmission was successful
• If the frame was destroyed, the sender just waits
  a random amount of time and sends it again
• This kind of system where users share a common
  channel resource is referred to as Contention
  System

6
Pure ALOHA System
7
Pure ALOHA System

• Frames have the same length
• Whenever two frames try to occupy the channel at
  the same time, there will be collision and both
  fames are garbled
• Question Can this system work? If yes, What is
  the efficiency or throughput of the system?

8
Formal Description of the Algorithm

• Transmission Mode Users transmit at any time
  they desire, encoding their transmission with an
  error detection code
• Listening Mode After a message transmission, the
  user listens for the acknowledgement (ACK) from
  the receiver. Transmissions from different users
  will sometimes overlap in time, causing errors in
  the data in each of the colliding partners. The
  user then receives a negative acknowledgement
  (NAK)

9
Formal Description of the Algorithm

• Retransmission Mode When NAK is received, the
  messages are simply retransmitted. Colliding
  users retransmit after a random amount of delay
• Timeout Mode If, after a transmission, the user
  does not receive either an ACK or NAK within a
  specified time, the user retransmits the message

10
Message Arrival Statistics

• Let us assume that there are infinite users
• is the total traffic arrival rate
  (packets/sec)
• is the successful traffic rate
  (packets/sec)
• is the traffic rejection rate (packets/sec)

11
Message Arrival Statistics

• Let be the number of bits per packet
• Then, the successful traffic or throughput
  (bits/sec) can be written as
• The total traffic can be written as

12
Message Arrival Statistics

• Let the Channel Capacity (maximum bit rate) be
• Then, the normalized throughput can be written as
• And, the normalized total traffic becomes

13
Message Arrival Statistics

• Maximum and Minimum Values of normalized
  throughput and total traffic
• Throughput Range
• Total Traffic Range

14
Message Arrival Statistics

• Packet transmission time
• The throughput and total traffic can be written
  in terms of packet transmission time. Thus,

15
Time window for successful transmission

• A user can successfully transmit a packet when
• as long as no other user began one within the
  previous seconds
• Or starts one within the next seconds
• Thus, a window of 2 seconds is needed for
  each message to be successful

16
Poisson Arrival Process

• The message arrival statistics for unrelated
  users of a communication system is often modeled
  as a Poisson Process
• For a process with , the packet arrival
  rate, the probability of arrivals in
  sec is given by

17
Probability of Successful Packet Transmission

• The packet will be successfully transmitted if no
  packets arrive in a duration of 2 sec, for
  the process with packet arrival rate of
  packets/sec. Thus the probability of successful
  transmission is

18
Normalized throughput of pure ALOHA

• Thus the normalized throughput of pure ALOHA
  system can be written as
• As increases, increases until a point
  is reached where further traffic increases create
  a large number of collision rate to cause a
  reduction in throughput

19
Throughput of pure ALOHA
20
Throughput of pure ALOHA

• The maximum value of throughput 1/2e0.18
• This maximum occurs at 0.5
• In pure ALOHA only 18 of the Channel Resource
  can be utilized
• Simplicity of control is achieved at the expense
  of Channel Capacity

21
Example-1

• A group of stations share a 56 kbps pure ALOHA
  channel. Each station outputs a packet on the
  average of once every 10s, even if the previous
  one has not yet been sent (i.e. the stations
  buffer the packets). Each packet is comprised of
  3000 bits. What is the maximum number of
  stations that can share this channel, assuming
  that the arrival process is Poisson?

22
Example-1 (Contd.)

• With pure ALOHA, the maximum usable capacity is
• Since the delay increases without bound for
  throughput greater than 0.184
• Each station sends 3000bits/10sec or 300bps.
  Thus the maximum number of stations that can
  share the channel is

23
Example-2

• A group of three stations share 56 kbps pure
  ALOHA channel. The average bit rate transmission
  from each of the three stations is 7.5 kbps, 10
  kbps, and 20 kbps. The size of each packet is
  100 bits. Find the normalized total traffic on
  the channel, the normalized throughput, the
  probability of successful transmission, and the
  arrival rate of successful packets. Assume the
  arrival process is Poisson.

24
Example-2 (Contd.)

• Arrival Rate in packets/sec.
• Total Arrival Rate in packets/sec
• Packet transmission time is

25
Example-2 (Contd.)

• Normalized Total Traffic is
• Normalized throughput is
• Probability of successful transmission is
• Arrival Rate of successful packets

26
Slotted ALOHA

• By introducing a small amount of coordination
  among users, the performance of the pure ALOHA
  can be improved
• Such a scheme is referred to as S-ALOHA or
  Slotted-ALOHA system
• As with pure ALOHA system, in S-ALOHA the packet
  size is constant
• Packets are required to be sent in the slot time
  between synchronization pulses and can be started
  only at the beginning of the time slot.

27
Slotted ALOHA

• This simple change reduces the rate of collisions
  by half, since only packets transmitted in the
  same slot can interfere with one another.
• Normalized throughput of S-ALOHA system is thus
  given by

28
Slotted ALOHA Operation
29
Slotted ALOHA Throughput

• The maximum value of throughput 1/e0.37
• This maximum occurs at G 1.0
• In S-ALOHA 37 of the Channel Resource can be
  utilized
• There exists tradeoff between Channel utilization
  and Coordination

30
Example-3

• A group of S-ALOHA stations generate a total of
  120 requests per second, including both original
  and retransmissions. Each request is for 12.5 ms
  duration slot.
• (a) What is the normalized total traffic?
• (b) What is the probability of a successful
  transmission on the first attempt?
• (c) What is the probability of exactly two
  collisions before a successful transmission?

31
Example-3 (Contd.)

• (a) The normalized total traffic is
• (b) Probability of successful transmission is
• (c) Probability of 2 collisions before a
  successful transmission is

32
Reservation ALOHA (R-ALOHA)

• Significant improvement in performance can be
  achieved over ALOHA system by using Reservations
• The R-ALOHA system has two basic modes
• Unreserved Mode (Quiescent State)
• 1. A time frame is established and divided into
  small reservation subslots
• 2. Users use these subslots to reserve message
  slots
• 3. After requesting a reservation, the user
  listens for an acknowledgement and a slot
  assignment

33
Reservation ALOHA (R-ALOHA)

• Reserved Mode
• 1. The time frame is divided into M1 slots
  whenever a reservation is made
• 2. The first M slots are used for message
  transmission
• 3. The last is subdivided into subslots to be
  used for reservations
• 4. Users send message packets only in their
  assigned portions of the M slots

34
5 Slots, 6 Subslots R-ALOHA System
35
R-ALOHA System

• In the quiescent state, with no reservations,
  time is partitioned into short subslots for
  making reservations
• Once reservation is made, the system is
  configured so that 5 message slots followed by 6
  reservation subslots becomes the timing format
• In the Figure the station seeks to reserve three
  message slots
• The reservation acknowledgement advises the
  station where to locate its data packets.

36
R-ALOHA System

• Since the control is distributed, all stations
  receive the downlink transmission and are aware
  of the reservation format
• The acknowledgement need not disclose any more
  than the location of the first slot to use.
• When there are no reservations taking place, the
  system reverts back to its quiescent mode

37
Performance of S-ALOHA and R-ALOHA Systems

• Delay vs. Throughput is used as the Performance
  measure for Multiple Access Schemes
• For normalized throughput (between 0 and 1) the
  delay equal to 0 until it is equal to 1 then the
  delay increases without bound

38
Ideal Delay vs. Throughput
39
Delay vs. ThroughputS and R-ALOHA Systems
40
Delay vs. ThroughputS and R-ALOHA Systems

• For throughput of less than approximately 0.20,
  the S-ALOHA manifests less delay than does
  R-ALOHA
• For throughput between 0.20 and 0.67, it is
  apparent that R-ALOHA is superior since the delay
  is less
• Why does S-ALOHA perform better at low traffic
  intensity?

41
Delay vs. ThroughputS and R-ALOHA Systems

• The S-ALOHA does not require the overhead of the
  reservation subslots as does R-ALOHA
• At low values of throughput (less than 0.2),
  R-ALOHA pays the price of greater delay due to
  greater overhead.
• For throughput (greater than 0.2 and less than
  0.67), the collisions and retransmissions
  inherent in S-ALOHA system cause it to incur
  greater delay more quickly than R-ALOHA

42
Channel Utilization

• Normalized throughput is a measure of channel
  utilization.
• It can be found by forming the ratio of the
  successfully transmitted message traffic, in bits
  per second, to the total message traffic
  including the rejected messages, in bits per
  second.

43
Channel Utilization

• Calculate the normalized throughput of a channel
  that has a maximum data rate of 50 kbps and
  operates with 10 ground stations, each station
  transmitting at the average rate of 2
  packets/sec. The system format provides for 1350
  bits/packet. Which of the three ALOHA schemes
  could be successfully used with the channel?

44
Channel Utilization

• The normalized throughput of the system is
• Only R-ALOHA scheme could be used, since with
  each of the other (pure and slotted) schemes, 54
  of the resource cannot be utilized