0% found this document useful (0 votes)
51 views

3-Queuing Delay & Analysis

This document discusses queuing delay and queuing analysis in packet switched networks like IP networks. It reviews the components of end-to-end delay including propagation, transmission, queuing, and processing delays. It introduces queuing theory concepts like traffic intensity and how average queuing delay increases as traffic intensity approaches and exceeds 1. The document discusses single server and multi-server queuing systems and key metrics like arrival rate, service rate, utilization, average queue length, waiting time and residence time. Formulas for analyzing these queuing systems are presented along with examples of applying queuing theory to modeling systems like routers.

Uploaded by

Krezia
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
51 views

3-Queuing Delay & Analysis

This document discusses queuing delay and queuing analysis in packet switched networks like IP networks. It reviews the components of end-to-end delay including propagation, transmission, queuing, and processing delays. It introduces queuing theory concepts like traffic intensity and how average queuing delay increases as traffic intensity approaches and exceeds 1. The document discusses single server and multi-server queuing systems and key metrics like arrival rate, service rate, utilization, average queue length, waiting time and residence time. Formulas for analyzing these queuing systems are presented along with examples of applying queuing theory to modeling systems like routers.

Uploaded by

Krezia
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 35

Queuing Delay and

Queuing Analysis

1
RECALL: Delays in Packet
Switched (e.g. IP) Networks
 End-to-end delay (simplified) =
– (dprop + dtrans + dqueue + dproc) … on each link

A B
 Where:
 Propagation delay (dprop) = d/s (dependent on path)
 Transmission delay (dtrans) = L/R (dependent on path)
 Queuing delay (dqueue) = (dependent on load)
 Processing delay (dproc) = (minimal-insignificant/node)
 Number of links (Q) = (dependent on path)
2
Introduction
Projected vs. Actual Response
Time

Why??

3
Queuing Analysis
Queueing delay (revisited)
 R: link bandwidth
(bps)

queueing delay
average
 L: packet length
(bits)
 a: average packet
arrival rate traffic intensity
= La/R

La/R ~ 0
 La/R ~ 0: avg. queueing delay small
 La/R -> 1: avg. queueing delay large
 La/R > 1: more “work” arriving
than can be serviced, average delay infinite!
La/R -> 1
4
Queuing Analysis
Introduction- Motivation
 Address how to analyze changes in
network workloads (i.e., a helpful tool
to use)
 Analysis of system (network) load and
performance characteristics
– response time
– throughput
 Performance tradeoffs are often not
intuitive
 Queuing theory, although
mathematically complex, often makes
analysis very straightforward 5
Queuing Analysis
Important Note
 Queuing theory is heavily dependent
on basic probability theory (a pre-
requisite for our graduate program)
 If you need to refresh your
knowledge in this area, please review
the Stallings textbook, Chapter 7:
Overview of Probability and
Stochastice Processes.
 I will not test you specifically on
probability theory, but will reference
it in coverage of the queuing topics
addressed in this module.
6
Queuing Analysis
Single-Server Queuing System

Queuing
Items Arriving
Items Departing
(rate: )
(rate: R)
System
(Delay Box)
(message, packet, cell)

Items Lost

7
Queuing Analysis
Router output port functions

datagram
link
switch buffer(s) layer line
fabric protocol termination
queueing (send)

Queue Queue server

 buffering/queuing required when datagrams arrive


from fabric faster than the transmission rate
 scheduling discipline chooses among queued
datagrams for transmission
 sending discipline (servicing the queue) on the
output link as determined by link protocol 8
Introduction
The Fundamental Task of
Queuing Analysis

Given: Determine:
• Arrival rate,  • Items waiting, w
• Service time, Ts • Waiting time, Tw
• Number of servers, N • Items queued, r
Queuing Analysis
• Residence time, Tr 9
Parameters for Single-Server
Queuing System

Comments, assuming queue has infinite capacity:


1. At  = 1, server is working 100% of the time (saturated), so items are queued
(delayed) until they can be served. Departures remain constant (for same L).
2. Traffic intensity, u = L/R. Note that Ts = L/R, so:
max = 1 / Ts = 1 / (L/R) is the theoretical maximum arrival rate, and that
Lmax/R = u = 1 at the theoretical maximum arrival rate
10
Queuing Analysis
Queuing Process Example: Single-
Server FIFO Queue
Depth of General Expression:
the Queue
(r)
TRn+1 = TSn+1 + MAX[0, Dn – An+1]

11
Queuing Analysis
General Characteristics of
Network Queuing Models
 Item population
– generally assumed to be infinite therefore,
arrival rate is persistent through time
 Queue size
– infinite, therefore no loss
– finite, more practical, but often immaterial
 Dispatching discipline
– FIFO, typical
– LIFO (when is this practical?)
– Relative/Preferential, based on QoS
12
Queuing Analysis
Multiserver Queuing System

Comments:
1. Assuming N identical servers, and  is the utilization of each server.
2. Then, N is the utilization of the entire system, and the maximum utilization is
N x 100%.
3. Therefore, the maximum supportable arrival rate that the system can handle is:
max = N / Ts = NR/L
13
Queuing Analysis
Multiple Single-Server Queuing
Systems

14
Chapter 8 Overview of Queuing Analysis
Basic Queuing Relationships

Single
General Multiserver
Server
r = Tr Little’s Formula  = Ts  = Ts
N
w = Tw Little’s Formula r=w+ u = Ts = N

Tr = Tw + Ts r = w + N

15
Queuing Analysis
Kendall’s notation
 Notation is X/Y/N, where:
X is distribution of interarrival times
Y is distribution of service times
N is the number of servers
 Common distributions
G = general distribution if interarrival times or
service times
 GI = general distribution of interarrival time with
the restriction that they are independent
 M = exponential distribution of inter-arrival times
(Poisson arrivals – p. 167) and service times
 D = deterministic arrivals or fixed length service

Queuing Analysis
M/M/1? M/D/1? 16
Important Formulas for Single-
Server Queuing Systems

Note Coefficient of variation:


if Ts = Ts => exponential
if Ts = 0 => constant
17
Queuing Analysis
Important Formulas for Single-
Server Queuing Systems

18
Queuing Analysis
Mean Number of Items in System
(r)- Single-Server Queuing
Ts/Ts = Coefficient of variation

M/M/1

M/D/1

19
Queuing Analysis
Mean Residence Time – (Tr)
Single-Server Queuing

M/M/1

M/D/1

20
Queuing Analysis
Network Queue Performance:
Key Fact

The higher the variability in arrival rate


at the router, relative to the service
time on the output link(s), i.e., Ts/Ts
(coefficient of variation) the poorer the
performance of the router, especially at
high rates of utilization.

21
Queuing Analysis
Multiple Server Queuing Systems

Multiserver
Queuing
System

Multiple Single-
Server Queuing
System

22
Queuing Analysis
Important Formulas for Multiserver
Queuing

Note:
Useful only in
M/M/N case,
with equal
service times
at all N
servers.

23
Queuing Analysis
Multiple Server Queuing Example
(p. 203)

Single server
M/M/1 (2nd Floor)

Multiserver
M/M/? (2nd Floor)

Multiple
Single server

M/M/1 (1st Floor)

M/M/1 (2nd Floor)

24
Queuing Analysis
MultiServer vs. Multiple Single-
Server Queuing System Comparison
(from example problem, pp. 203-204)

Single server case (M/M/1):


Single server utilization:  = 10 engineers x 0.5 hours each / 8 hour work day
= 5/8 = .625
Average time waiting: Tw = Ts / 1 -  = 0.625 x 30 / .375 = 50 minutes
Arrival rate:  = 10 engineers per 8 hours = 10/480 = 0.021 engineers/minute
90th percentile waiting time: mTw(90) = Tw/ x ln(10) = 146.6 minutes
Average number of engineers waiting: w = Tw = 0.021 x 50 = 1.0416 engineers

25
Queuing Analysis
Example: Router Queuing
Internet  = 5 packets/sec

9600 L = 144 octets
bps

From data provided:


• Ts = L/R = (144x8)/9600 = .12sec
•  = Ts = 5 packets/sec x .12sec = .6
Determine:
1. Tr= Ts / (1-) = .12sec/.4 = .3 sec
2. r =  / (1-) = .6/.4 = 1.5 packets
ln(1-.90)
3. mr(90) = - 1 = 3.5 packets For 3 & 4, use:
ln (.6)
ln(1-.95) ln(1 – y/100)
4. mr(95) = - 1 = 4.8 packets mr(y) = -1
ln (.6) ln 
26
Queuing Analysis
Priorities in Queues – Two
priority classes

27
Queuing Analysis
Priorities in Queues – Example
Tr
        
 64Kbps

Router queue services two packet sizes: Find the average Queuing Delay (Tr)
• Long = 800 octets through the router:
• Short = 80 octets 1 Ts 1 + 2 Ts 2
• Lengths exponentially distributed Tr1 = Ts1 +
1 - 1
• Arrival rates are equal, 8packets/sec
• Link transmission rate is 64Kbps .08 x .01 + .8 x .1
= .01 + = 0.098 sec
• Short packets are priority 1, 1-.08
• Longer packets are priority 2
Tr 1 - Ts 1
Tr2 = Ts2 +
From data above, calculate: 1-
Ts 1 = Lshort/R = (80 x 8) / 64000 = .01 sec .098 - .01
= .1 + = 0. 833 sec
Ts 2 = Llong/R = (800 x 8) / 64000 = .1 sec 1 - .88
1 =  Ts 1 = 8 x 0.01 = 0.08
1 2
2 =  Ts 2 = 8 x 0.1 = 0.8 Tr = T +
 r1  Tr2
 = 1 + 2 = 0.88
= .5 x .098 + .5 x .833 = 0.4655 sec
28
Chapter 8 Overview of Queuing Analysis
Network of Queues

29
Queuing Analysis
Elements of Queuing Networks

30
Queuing Analysis
Queuing Networks

31
Queuing Analysis
Jackson’s Theorem and Queuing
Networks
 Assumptions:
– the queuing network has m nodes, each providing
exponential service
– items arriving from outside the system at any node
arrive with a Poisson rate
– once served at a node, an item moves immediately to
another with a fixed probability, or leaves the network
 Jackson’s Theorem states:
– each node is an independent queuing system with Poisson
inputs determined by partitioning, merging and tandem
queuing principles
– each node can be analyzed separately using the M/M/1
or M/M/N models
– mean delays at each node can be added to determine
mean system (network) delays
32
Queuing Analysis
Jackson’s Theorem - Application
in Packet Switched Networks Internal load:
L
 =  i
i=1
Packet Switched where:
Network
 = total on all links in network
i = load on link i
L = total number of links

Note:
• Internal > offered load
External load, offered to network:
• Average length for all paths:
N N
E[number of links in path] = /
 =   jk
j=1 k=2 • Average number of item waiting
where: and being served in link i: ri = i Tri
 = total workload in packets/sec • Average delay of packets sent
jk = workload between source j through the network is:
1 L Mi
and destination k T=   (See p. 210)
i=1 Ri - Mi
N = total number of (external)
sources and destinations where: M is average packet length and 33
Queuing Analysis Ri is the data rate on link i
Estimating Model Parameters
 To enable queuing analysis using
these models, we must estimate
certain parameters for the network:
– Mean and standard deviation of arrival
rate
– Mean and standard deviation of service
time (or, packet size)
 Typically, these estimates use
sample measurements taken from an
existing system

34
Queuing Analysis
Sample Means for Underlying
Exponential Distribution
Sampling:
• The mean is generally
the most important
quantity to estimate:


N
( ) = 1 Xi
N i=1
• Sample mean is itself a
random variable
• Central Limit Theorem:
the probability
distribution tends to
normal as sample size,
N, increases for
virtually all underlying
distributions
• The mean and variance
of X can be calculated
as:
E[ ]= E[X] = 
Var[ ]= 2x/N

35
Queuing Analysis

You might also like