3.

Examples

lect03.ppt S-38.145 - Introduction to Teletraffic Theory – Spring 2005 1

3. Examples

Contents

• Model for telephone traffic

• Packet level model for data traffic
• Flow level model for elastic data traffic
• Flow level model for streaming data traffic

2
3. Examples

Classical model for telephone traffic (1)

• Loss models have traditionally been used to describe (circuit-

switched) telephone networks
– Pioneering work made by Danish mathematician A.K. Erlang (1878-1929)
• Consider a link between two telephone exchanges
– traffic consists of the ongoing telephone calls on the link

3
3. Examples

Classical model for telephone traffic (2)

• Erlang modelled this as a pure loss system (m = 0)

– customer = call
• λ = call arrival rate (calls per time unit)
– service time = (call) holding time
• h = 1/µ = average holding time (time units)
– server = channel on the link
• n = nr of channels on the link

µ
1
λ µ
µ
µ
n

4
3. Examples

Traffic process

channel-by-channel
occupation call holding time
6
channels

5
4
3
2
1

time
call arrival times
blocked call
nr of channels
occupied
nr of channels

6
5
4
3
2
1
0
traffic volume time

5
3. Examples

Traffic intensity

• The strength of the offered traffic is described by the traffic intensity a

• By definition, the traffic intensity a is the product of the arrival rate λ
and the mean holding time h:

a = λh
– The traffic intensity is a dimensionless quantity. Anyway, the unit of the
traffic intensity a is called erlang (erl)
– By Little’s formula: traffic of one erlang means that one channel is occupied
on average
• Example:
– On average, there are 1800 new calls in an hour, and the average holding
time is 3 minutes. Then the traffic intensity is

a = 1800 ∗ 3 / 60 = 90 erlang
6
3. Examples

Blocking

• In a loss system some calls are lost

– a call is lost if all n channels are occupied when the call arrives
– the term blocking refers to this event
• There are two different types of blocking quantities:
– Call blocking Bc = probability that an arriving call finds all n channels
occupied = the fraction of calls that are lost
– Time blocking Bt = probability that all n channels are occupied at an
arbitrary time = the fraction of time that all n channels are occupied
• The two blocking quantities are not necessarily equal
– Example: your own mobile
– But if calls arrive according to a Poisson process, then Bc = Bt
• Call blocking is a better measure for the quality of service experienced
by the subscribers but, typically, time blocking is easier to calculate

7
3. Examples

Call rates

• In a loss system each call is either lost or carried. Thus, there are
three types of call rates:
– λoffered = arrival rate of all call attempts
– λcarried = arrival rate of carried calls
– λlost = arrival rate of lost calls

λoffered λcarried

λlost

λoffered = λcarried + λlost = λ

λcarried = λ (1 − Bc )
λlost = λBc
8
3. Examples

Traffic streams

• The three call rates lead to the following three traffic concepts:
– Traffic offered aoffered = λofferedh
λoffered λcarried
– Traffic carried acarried = λcarriedh
– Traffic lost alost = λlosth λlost

aoffered = acarried + alost = a

acarried = a (1 − Bc )
alost = aBc
• Traffic offered and traffic lost are hypothetical quantities, but
traffic carried is measurable, since (by Little’s formula) it corresponds
to the average number of occupied channels on the link

9
3. Examples

Teletraffic analysis (1)

• System capacity
– n = number of channels on the link
• Traffic load
– a = (offered) traffic intensity
• Quality of service (from the subscribers’ point of view)
– Bc = call blocking = probability that an arriving call finds all n channels
occupied
• Assume an M/G/n/n loss system:
– calls arrive according to a Poisson process (with rate λ)
– call holding times are independently and identically distributed according to
any distribution with mean h

10
3. Examples

Teletraffic analysis (2)

• Then the quantitive relation between the three factors (system, traffic,
and quality of service) is given by Erlang’s formula:

an
Bc = Erl(n, a ) := n!
n i
∑ ai!
i =0

n!= n ⋅ (n − 1) ⋅ K ⋅ 2 ⋅1, 0!= 1

• Also called:
– Erlang’s B-formula
– Erlang’s blocking formula
– Erlang’s loss formula
– Erlang’s first formula
11
3. Examples

Example

• Assume that there are n = 4 channels on a link and the offered traffic is
a = 2.0 erlang. Then the call blocking probability Bc is
24 16
4! 24 2
B c = Erl( 4 , 2 ) = = = ≈ 9 .5 %
22 23 24 1+ 2 + 4+8 16
+ 24 21
1+ 2 + 2!
+ 3!
+ 4! 2 6

• If the link capacity is raised to n = 6 channels, then Bc reduces to

26
Bc = Erl( 6 , 2 ) = 6! ≈ 1 .2 %
2 3 4 5 6
1 + 2 + 22! + 23! + 24! + 25! + 26!

12
3. Examples

Capacity vs. traffic

• Given the quality of service requirement that Bc < 1%, the required
capacity n depends on the traffic intensity a as follows:

n(a ) = min{i = 1,2,K | Erl(i, a ) < 0.01}

100

80

60
capacity n
40

20

20 40 60 80 100
13
traffic a
3. Examples

Quality of service vs. traffic

• Given the capacity n = 20 channels, the required quality of service

1 − Bc depends on the traffic intensity a as follows:

1 − Bc (a ) = 1 − Erl(20, a )
1

0.8

0.6
quality of service
1 − Bc 0.4

0.2

20 40 60 80 100
14
traffic a
3. Examples

Quality of service vs. capacity

• Given the traffic intensity a = 15.0 erlang, the required quality of service
1 − Bc depends on the capacity n as follows:

1 − Bc (n) = 1 − Erl(n,15.0)
1

0.8

0.6
quality of service
1 − Bc 0.4

0.2

10 20 30 40 50
15
capacity n
3. Examples

Contents

• Model for telephone traffic

• Packet level model for data traffic
• Flow level model for elastic data traffic
• Flow level model for streaming data traffic

16
3. Examples

Packet level model for data traffic (1)

• Queueing models are suitable for describing (packet-switched) data

traffic at packet level
– Pioneering work made by many people in 60’s and 70’s related to
ARPANET, in particular L. Kleinrock (http://www.lk.cs.ucla.edu/)
• Consider a link between two packet routers
– traffic consists of data packets transmitted along the link

R
R R
R

17
3. Examples

Packet level model for data traffic (2)

• This can be modelled as a pure queueing system with a single server

(n = 1) and an infinite buffer (m = ∞)
– customer = packet
• λ = packet arrival rate (packets per time unit)
• L = average packet length (data units)
– server = link, waiting places = buffer
• C = link speed (data units per time unit)
– service time = packet transmission time
• 1/µ = L/C = average packet transmission time (time units)

µ
λ

18
3. Examples

Traffic process

packet status (waiting/in transmission)

waiting time
transmission time

time
packet arrival times
number of packets in the system
4
3
2
1
0
time
link occupation
1

0
time

19
3. Examples

Traffic load

• The strength of the offered traffic is described by the traffic load ρ

• By definition, the traffic load ρ is the ratio between the arrival rate λ
and the service rate µ = C/L:
λ λL
ρ= =
µ C
– The traffic load is a dimensionless quantity
– By Little’s formula, it tells the utilization factor of the server, which is the
probability that the server is busy

20
3. Examples

Example

• Consider a link between two packet routers. Assume that,

– on average, 50,000 new packets arrive in a second,
– the mean packet length is 1500 bytes, and
– the link speed is 1 Gbps.
• Then the traffic load (as well as, the utilization) is

ρ = 50,000 ∗1500 ∗ 8 / 1,000,000,000 = 0.60 = 60%

21
3. Examples

Delay

• In a queueing system, some packets have to wait before getting served

– An arriving packet is buffered, if the link is busy upon the arrival
• Delay of a packet consists of
– the waiting time, which depends on the state of the system upon the
arrival, and
– the transmission time, which depends on the length of the packet and the
capacity of the link
• Example:
– packet length = 1500 bytes
– link speed = 1 Gbps
– transmission time = 1500*8/1,000,000,000 = 0.000012 s = 12 µs

22
3. Examples

Teletraffic analysis (1)

• System capacity
– C = link speed in kbps
• Traffic load
– λ = packet arrival rate in pps (considered here as a variable)
– L = average packet length in kbits (assumed here to be constant 1 kbit)
• Quality of service (from the users’ point of view)
– Pz = probability that a packet has to wait “too long”, i.e. longer than a given
reference value z (assumed here to be constant z = 0.00001 s = 10 µs)
• Assume an M/M/1 queueing system:
– packets arrive according to a Poisson process (with rate λ)
– packet lengths are independent and identically distributed according to the
exponential distribution with mean L

23
3. Examples

Teletraffic analysis (2)

• Then the quantitive relation between the three factors (system, traffic,
and quality of service) is given by the following formula:

Pz = Wait(C , λ ; L, z ) :=
 λL exp(−( C − λ ) z ) = ρ exp(− µ (1 − ρ ) z ), if λL < C ( ρ < 1)
C L
 1, if λL ≥ C ( ρ ≥ 1)
• Note:
– The system is stable only in the former case (ρ < 1). Otherwise the number
of packets in the buffer grows without limits.

24
3. Examples

Example

• Assume that packets arrive at rate λ = 600,000 pps = 0.6 packets/µs

and the link speed is C = 1.0 Gbps = 1.0 kbit/µs.
• The system is stable since

ρ = λCL = 0.6 < 1

• The probability Pz that an arriving packet has to wait too long (i.e.
longer than z = 10 µs) is

Pz = Wait(1.0,0.6;1,10) = 0.6 exp( −4.0) ≈ 1%

25
 3. Examples

Capacity vs. arrival rate

• Given the quality of service requirement that Pz < 1%, the required link
speed C depends on the arrival rate λ as follows:

C (λ ) = min{c > λL | Wait(c, λ ;1,10) < 0.01}

2
1.75
1.5
1.25
link speed C 1
(Gbps) 0.75
0.5
0.25

0.2 0.4 0.6 0.8 1

26
arrival rate λ (packets/µs)
 3. Examples

Quality of service vs. arrival rate

• Given the link speed C = 1.0 Gbps = 1.0 kbit/µs, the quality of service
1 − Pz depends on the arrival rate λ as follows:

1 − Pz (λ ) = 1 − Wait(1.0, λ ;1,10)
1

0.8

0.6
quality of service
1 − Pz 0.4

0.2

0.2 0.4 0.6 0.8 1

27
arrival rate λ (packets/µs)
 3. Examples

Quality of service vs. capacity

• Given the arrival rate λ = 600,000 pps = 0.6 packets/µs, the quality of
service 1 − Pz depends on the link speed C as follows:

1 − Pz ( R ) = 1 − Wait(C ,0.6;1,10)
1

0.8

0.6
quality of service
1 − Pz 0.4

0.2

0.6 0.7 0.8 0.9 1

28
link speed C (Gbps)
3. Examples

Contents

• Model for telephone traffic

• Packet level model for data traffic
• Flow level model for elastic data traffic
• Flow level model for streaming data traffic

29
3. Examples

Flow level model for elastic data traffic (1)

• Sharing models are suitable for describing elastic data traffic at flow
level
– Elasticity refers to the adaptive sending rate of TCP flows
– This kind of models have been proposed, e.g., by J. Roberts and his
researchers (http://perso.rd.francetelecom.fr/roberts/)
• Consider a link between two packet routers
– traffic consists of TCP flows loading the link

R
R R
R
30
3. Examples

Flow level model for elastic data traffic (2)

• The simplest model is a single server (n = 1) pure sharing system

with a fixed total service rate of µ
– customer = TCP flow = file to be transferred
• λ = flow arrival rate (flows per time unit)
• S = average flow size = average file size (data units)
– server = link
• C = link speed (data units per time unit)
– service time = file transfer time with full link speed
• 1/µ = S/C = average file transfer time with full link speed (time units)

∞ µ
λ

31
3. Examples

Traffic process

transfer time
flow duration with full link rate

extra delay

time
flow arrival times
number of flows in the system
4
3
2
1
0
time
relative transmission rate for single flows
1

1/2
1/41/3
0
time 32
3. Examples

Traffic load

• The strength of the offered traffic is described by the traffic load ρ

• By definition, the traffic load ρ is the ratio between the arrival rate λ
and the service rate µ = C/S:
λ λS
ρ= =
µ C
– The traffic load is (again) a dimensionless quantity
– It tells the utilization factor of the server

33
3. Examples

Example

• Consider a link between two packet routers. Assume that,

– on average, 50 new flows arrive in a second,
– average flow size is 1,500,000 bytes, and
– link speed is 1 Gbps.
• Then the traffic load (as well as, the utilization) is

ρ = 50 ∗1,500,000 ∗ 8 / 1,000,000,000 = 0.60 = 60%

34
3. Examples

Throughput

• In a sharing system the service capacity is shared among all active

flows. It follows that all flows get delayed (unless there is only a single
active flow)
• By definition, the ratio between the average flow size S and the average
total delay D of a flow is called throughput θ,

θ =S/D
• Example:
– S = 1 Mbit
– D=5s
– θ = S/D = 0.2 Mbps

35
3. Examples

Teletraffic analysis (1)

• System capacity
– C = link speed in Mbps
• Traffic load
– λ = flow arrival rate in flows per second (considered here as a variable)
– S = average flow size in kbits (assumed here to be constant 1 Mbit)
• Quality of service (from the users’ point of view)
– θ = throughput
• Assume an M/G/1-PS sharing system:
– flows arrive according to a Poisson process (with rate λ)
– flow sizes are independent and identically distributed according to any
distribution with mean S

36
3. Examples

Teletraffic analysis (2)

• Then the quantitive relation between the three factors (system, traffic,
and quality of service) is given by the following formula:

C − λS = C (1 − ρ ), if λS < C ( ρ < 1)
θ = Xput(C , λ ; S ) := 
0, if λS ≥ C ( ρ ≥ 1)
• Note:
– The system is stable only in the former case (ρ < 1). Otherwise the number
of flows as well as the average delay grows without limits. In other words,
the throughput of a flow goes to zero.

37
3. Examples

Example

• Assume that flows arrive at rate λ = 600 flows per second and the link
speed is C = 1000 Mbps = 1.0 Gbps.
• The system is stable since

ρ = λCS = 1000
600 = 0.6 < 1

• Throughput is

θ = Xput(1000,600;1) = 1000 − 600 = 400 Mbps = 0.4 Gbps

38
 3. Examples

Capacity vs. arrival rate

• Given the quality of service requirement that θ ≥ 400 Mbps, the

required link speed C depends on the arrival rate λ as follows:

C (λ ) = min{c > λS | Xput(c, λ ;1) ≥ 400} = λS + 400

1400
1200
1000
800
link speed C
(Mbps) 600
400
200

200 400 600 800 1000

39
arrival rate λ (flows per second)
3. Examples

Quality of service vs. arrival rate

• Given the link speed C = 1000 Mbps, the quality of service θ depends
on the arrival rate λ as follows:

θ (λ ) = Xput(1000, λ ;1) = 1000 − λS , λ < 1000/S

1000

800

600
throughput θ
(Mbps) 400

200

200 400 600 800 1000

40
arrival rate λ (flows per second)
3. Examples

Quality of service vs. capacity

• Given the arrival rate λ = 600 flows per second, the quality of service θ
depends on the link speed C as follows:

θ (C ) = Xput(C ,600;1) = C − 600 S , C > 600 S

400
350
300
250
throughput θ 200
(Mbps) 150
100
50

650 700 750 800 850 900 950 1000

41
link speed C (Mbps)
3. Examples

Contents

• Model for telephone traffic

• Packet level model for data traffic
• Flow level model for elastic data traffic
• Flow level model for streaming data traffic

42
3. Examples

Flow level model for streaming CBR traffic (1)

• Infinite system is suitable for describing streaming CBR traffic at flow

level
– The transmission rate and flow duration of a streaming flow are insensitive
to the network state
– This kind of models applied in 90’s to the teletraffic analysis of CBR traffic in
ATM networks
• Consider a link between two packet routers
– traffic consists of UDP flows carrying CBR traffic (like VoIP) and loading the
link

R
R R
R
43
3. Examples

Flow level model for streaming CBR traffic (2)

• Model: an infinite system (n = ∞)

– customer = UDP flow = CBR bit stream
• λ = flow arrival rate (flows per time unit)
– service time = flow duration
h = 1/µ = average flow duration (time units)
•
• Bufferless flow level model:
– when the total transmission rate of the flows exceeds the link capacity, bits
are lost (uniformly from all flows)

µ
1
λ µ
•
•
• ∞
44
3. Examples

Traffic process

flow durations

time
flow arrival times

total bit rate lost traffic

(number of flows)

carried traffic time

45
3. Examples

Offered traffic

• Let r denote the bit rate of any flow

• The strength of offered traffic is described by average total bit rate R
– By Little’s formula, the average number of flows is

a = λh
– This may be called traffic intensity (cf. telephone traffic)
– It follows that

R = ar = λhr

46
3. Examples

Loss ratio

• Let N denote the number of flows in the system

• When the total transmission rate Nr exceeds the link capacity C, bits
are lost with rate

Nr − C
• The average loss rate is thus

E[( Nr − C ) + ] = E[max{Nr − C ,0}]

• By definition, the loss ratio ploss gives the ratio between the traffic lost
and the traffic offered:

E[( Nr −C ) + ] 1 +
ploss = E[ Nr ]
= ar
E [( Nr − C ) ]
47
3. Examples

Teletraffic analysis (1)

• System capacity
– C = nr = link speed in kbps
• Traffic load
– R = ar = offered traffic in kbps
– r = bit rate of a flow in kbps.
• Quality of service (from the users’ point of view)
– ploss = loss ratio
• Assume an M/G/∞ infinite system:
– flows arrive according to a Poisson process (with rate λ)
– flow durations are independent and identically distributed according to any
distribution with mean h

48
3. Examples

Teletraffic analysis (2)

• Then the quantitive relation between the three factors (system, traffic,
and the quality of service) is given by the following formula

∞ i −a
1 a
ploss = LR(n, a) := a ∑ (i − n) i! e
i = n +1

• Example:
– n = 20
– a = 14.36
– ploss = 0.01

49
3. Examples

Capacity vs. traffic

• Given the quality of service requirement that ploss < 1%, the required
capacity n depends on the traffic intensity a as follows:

n(a ) = min{i = 1,2,K | LR(i, a ) < 0.01}

100

80

60
capacity n
40

20

20 40 60 80 100
50
traffic a
3. Examples

Quality of service vs. traffic

• Given the capacity n = 20, the required quality of service 1 − ploss

depends on the traffic intensity a as follows:

1 − ploss (a ) = 1 − LR(20, a )
1

0.8

0.6
quality of service
1 − ploss 0.4

0.2

20 40 60 80 100
51
traffic a
3. Examples

Quality of service vs. capacity

• Given the traffic intensity a = 15.0 erlang, the required quality of service
1 − ploss depends on the capacity n as follows:

1 − ploss (n) = 1 − LR(n,15.0)

1

0.8

0.6
quality of service
1 − ploss 0.4

0.2

10 20 30 40 50
52
capacity n
3. Examples

THE END

53