Title: Modelling and Simulation of Systems
1Modelling and Simulationof Systems
2Useful References
- Introduction to Computer Simulation
- A System Dynamics Modelling Approach
- Roberts et al, Call No 003.3 INT
- Operations Research - An Introduction
- Hamdy Taha, 6E, Prentice-Hall, Call No CA003 TAH
- Network II.5 User Manual
3Model
- A standard or example for imitation or comparison
- Cheaper to construct a wrong model than a
wrong system - Enables a variety of ideas to be tried
- Simplifies complexity
- Reduces the scale of the system
- Provides a plan or map of the system
- Deconstructs a system into its components
4Modelling and Simulationof Systems
- System - a collection of interacting components
(elements) that function together for some
purpose - the circulatory system
- a computer network
- Simulation - to simulate is to imitate something
- A simulation involves a model that mimics
important components (elements) of the system
under study
5Modelling and Simulation
- A simulation may involve a
- physical model
- a mental conception
- a mathematical model
- a computer model
- a combination of the above
- Simulation has been practiced for centuries in
games, art and engineering, in these activities
participants - derive abstract models from concrete situations,
- they then interpret the models to provide
insights into the original situations
6Modelling and Simulation
- Given a model cannot include all the elements
and all the interactions between the elements
that make up a system the model is
simpler than the phenomenon it represents -
- model aircraft versus real aircraft --gt yet
engineers can use this physical model in a wind
tunnel to predict the flight characteristics of
the real aircraft - Newtons law of gravitational attraction is a
mathematical model of the force of gravity --gt
yet by interpreting its symbols we can learn
about a physical phenomenon
7Modelling and Simulation
- As a simplification of reality,
a model
will not produce reliable information
on every aspect of the system
being modelled
8Modelling and Simulation
- It is often useful to be able to
- predict the affect on a Systems Performance, of
a change to the system or change in the systems
workload - determine the minimum network resources required
to meet an applications requirements - Ask Questions like -
- How is Response Time (R/T) affected
? - How is Transaction Throughput
affected ? - Where is the bottleneck ?
- Will one 64kbps link be sufficient
to handle data traffic ?
9Modelling and Simulation
- Common Approaches
- 1) Do an After-the-Fact Study
- 2) Make a Simple Projection
- 3) Develop a Mathematical Model of the system
- 4) Develop a Computer Model of the system
101) Do an After-the-Fact Study
- Basically you make the changes then Wait-and-See
what happens - This tends to lead to
- unhappy clients, ie average R/T goes from 3secs
to 30secs - unwise purchases being made
- In effect, it is no option at all
- It is also a way of ensuring you have a short but
(in)glorious career
112) Make a Simple Projection
- Make a simple projection based on current
knowledge of the Systems behaviour - For example we could assume a Straight line
(y2x1) relationship between R/T (y) and System
load (x) based on the behaviour of the current
system up to a transaction load of 500 trans/sec - Refer fig Projected versus Actual Response time
- However what if there is an Exponential (y2x21)
relationship ? - R/T collapses between 0.8 and 0.9
system load
123) Develop a Mathematical Model
- A Mathematical Model can be expressed as a Set
of Equations that can be solved to to yield a set
of estimated results that allow us to predict
the affect of System Change on - Response Time
- Transaction Throughput
- Number of communications links required
- etc
- For Distributed Computing Systems which involve,
shared facilities - networks, databases,
printers, routers - Mathematical Models based on
Queuing Theory provide a reasonable fit to reality
134) Develop a Computer Model
- Given a simulation language - Simscript or
simulation package - Network II.5, the analyst
can model reality in great detail - Overheads - Time to build Model
- - Time to run simulation
- - Time to analyse results
- Due to the development of (cheap) powerful CPUs
simulation is increasingly used in training,
system design - Flight simulators
- Architectural design, walk through the proposed
building - Virtual Reality environments
14Modelling - Telephone Industry
- The Telephone industry developed a model of voice
traffic patterns in order to design and provision
a profitable circuit switching network - Engineers can
- Over Provision design a non-blocking switching
system that guarantees all users can get a dial
tone at anytime. Cost high -gt user
fees high -gt few users -gt low profit - Under Provision under design, too many users
will be denied service and again low profit - Modelling and Simulation aim to reduce the risk
of Over or Under provision, hence resulting in a
network (design) that will provide services most
of the time
15Modelling and Simulation
- Risk of Over Provisioning
- Money spent needlessly on hardware and services
that remain under utilized - Risk of Under Provisioning
- Users immediately suffer a fixed period of lost
productivity while additional time and money is
spent to upgrade the network - Refer fig 1.1 Degree of Network Provisioning
16Model Building - Phases
- Problem Definition
- Involves recognizing and defining a problem to
study that is amenable to analysis in systems
terms - contains variables that vary over time
- causes of the variability can be described
- causes can be contained in a closed system
- System Conceptualization
- Involves committing to paper the important
influences believed to be operating within a
system
17Model Building - Phases
- Model Representation
- Model represented in the form of computer code
- Model Behaviour
- Computer simulation can be used to determine how
all of the variables (elements) within the system
behave over time
18Model Building - Phases
- Model Evaluation
- Tests are performed to evaluate the models
quality and validity eg - checking for logical consistency
- matching model output against observed data
collected over time - formal statistical tests for variables used
within the simulation - Policy Analysis and Model Use
- The model is used to test alternative system
designs that might be implemented as the real
system
19Modelling and Simulationof Systems
- Lunar Lander Model
- Initial state
- Internal and External variables
- Deterministic
- Nondeterministic
- Two Types of Model
- Discrete Event Simulation
- Continuous Simulation
20Modelling and SimulationLunar Lander - A Simple
Game
- The aims are
- to land on the moon
- achieve a low rate of descent
- leave enough fuel for takeoff
- The game is a simulation that manipulates
variables representing - altitude
- rate of descent
- amount of fuel remaining
21Modelling and SimulationLunar Lander - A Simple
Game
- The simulation begins (the Initial State) by
displaying - initial rate of descent
- initial altitude
- amount of fuel available to burn
- Prompted for next amount of fuel to burn, then
calculate - fuel remaining after the burn
- the new altitude
- the new rate of descent
22Modelling and SimulationLunar Lander
- The variables can be classed as
- internal - rate of descent, altitude, amount of
fuel remaining - external - amount of fuel used in next burn
- These variables control the outcome of the
simulation - The variables are related via equations
once the amount of fuel
to be burned is assigned,
the value of the other variables can be
calculated
23Modelling and SimulationLunar Lander
- The model is Deterministic
- to the extent that the same sequence of fuel
quantities entered always results in the same
outcome - Once a winning sequence is discovered ,
the equations imply, it
will always produce a soft landing - The model is Nondeterministic
- to the extent that we cannot predict the sequence
of fuel amounts a player might enter
24Modelling and SimulationLunar Lander
- How can we make the game (model) more
nondeterministic ? - Start each new game with a random altitude and
rate of descent
25Modelling and SimulationTwo Types of Model
- Discrete Model
- deals with a system whose behaviour changes only
at given instants (in time) - eg when a customer enters or leaves the system
(bank) - Continuous Model
- deals with a system whose behaviour changes
continuously (in time) - where variables (attributes) of system undergo
smooth continuous changes - eg electric current - usually represented in terms of differential or
algebraic equations
26Modelling and Simulation Discrete Event
- Another characteristic of the Lunar Lander model
is its discrete outcome - When the fuel is entered for the next burn the
lander instantly assumes a new altitude and rate
of descent - The simulation computes the future state of the
modelled system at discrete time intervals,
often determined by events in the simulation
eg the amount of
fuel entered for the next burn
27Modelling and Simulation Continuous
- Newtons Mathematical Model of gravitational
attraction - is based on infinitesimal divisions of time and
space and produces a continuous range of results - generally described by deterministic
differential or algebraic equations that describe
the interactions among the different attributes
of the system
28Simulation for Network Planning and Management
- Network managers are likely to have at least two
major duties - planning new installations or network upgrades
- the day-to-day management of the network itself
29Planning New Installations orNetwork Upgrades
- When large systems are being designed,
the task may be quite complex
and it is
becoming more common to see planners resorting to
the use of modelling and simulation - Simulation will aid in
- Concept testing
- Comparison of alternatives
- Documentation of proposals
- Planned system must support the business function
30Concept Testing
- Models of networks can be developed
and using a simulation
design tool tested (Proof of Concept)
to see which model is the most likely to be
suitable as the basis of
building a workable network
31Comparison of alternatives
- A good simulation design tool should permit rapid
creation of a simulated environment
and also
simple modification of the network type for
comparison of differing systems prior to final
selection - This should be able to be done easily and
efficiently - This gives the designer the opportunity to
test different systems
and network alternatives and come up
with a workable design
32Documentation of proposals
- The simulation design tool should also support
high-level reporting as an aid
in producing documented proposals
for evaluation by
management
33Supporting the Business Function
- The results of the simulation should allow the
building of a workable system - that satisfies the functional requirements
defined in the application specifications - that carries the required load with an adequate
response time - that provides an efficient interface with the
people who will be using it - that is easily expandable
- that is not unreasonable in terms of its
implementation cost and ongoing support costs
34Network II.5
- Is a good example of a simulation design tool
- It permits the rapid development and
analysis of prototype networks
35Network II.5 - From User Manual
- Is a design tool which takes a computer system
description you specify and provides measures of - hardware utilization
- software execution
- message delivery
- response times
- contention
- Is used both to evaluate the ability of a
proposed system (configuration) to meet the
required workload and to evaluate competing
designs
36Network II.5 - From User Manual
- Is designed to model a wide variety of computer
architectures, from a single processor to a
complex system of processors and storage devices
connected in a network - It allows the portions of a computer system of
special interest to be modelled at a detailed
level while the rest of the system is modelled at
a coarser level - It is a tool not a computer language with an
interactive graphical interface than can be used
after a short period of training
37Day-to-Day Management
- In modern networks where switching is common,
the level of network monitoring
(of traffic
generated by each node in the network)
that is easy in shared-media (non-switched)
networks is now not possible - The volume of traffic generated by Monitoring
agents can be detrimental to network performance
(refer lecture 11) - The monitors cannot cover the whole network
(refer lecture 11)
38Day-to-Day Management
- Simulating the functioning of the network is one
way of being able to monitor the level of service
that the network provides - This can be done by testing the network
in a simulated environment
using
real traffic
that is generated by
real nodes on the network - The traffic details can be sampled for a period
of time for each node (eg using a Monitoring
agent like RMON)
and then imported into the simulation
environment
39Day-to-Day Management
- RMON a Monitoring Agent
- collects statistics (packet counts, error rates
etc) with management station retrieving data
after some time duration - Refer lecture 11
40Day-to-Day Management
- The traffic that is recorded can be used by the
simulation design tool to generate more traffic - The simulated traffic created is based upon the
samples of real traffic recorded - Network II.5 uses a tool called TRAFLINK III that
can import recorded data from a number of popular
monitoring tools
41TrafLink III - From User Manual
- Provides the link between Network II.5 and other
software packages - It allows you to import network topology
information and traffic traces into Network II.5
and use the developed model as a foundation for
what-if scenarios - It enables you to represent your existing network
topology as a Network II.5 model
42Introduction to Network Performance
43Network Performance
- Networks are built to support business
applications - word processing
- computer-aided drawing
- electronic mail
- imaging etc
- Critical Success Factor - each business
application must be performed in real time - Real Time - definition depends on the application
and can vary from msecs to hours
44- What Real Time Responses are Expected to satisfy
the following Business Applications ? - Word Processing
- Computer aided drawing
- Electronic mail
45Real Time Response Expected
- Word Processing
- user loads document, makes changes, writes back
to file server - read/write operations should be a few seconds at
most - Computer aided drawing
- drawing activities are iterative
- users expect subsecond response time when loading
a library symbol or deleting a component from a
drawing - Electronic mail
- may take minutes or hours without users becoming
upset with network performance
46- How is User Productivity affected if Response
Time is poor ?
47Real Time Response Expected
- User Productivity
- Declines if applications do not respond quickly
enough, as User - frustration sets in
- attitudes deteriorate
- work output slows down
- confidence in the network becomes low
- Network Performance must be sufficient to support
the users/applications response time
expectations
48Common Performance Metrics
- User Response Time
- determined by adding up all the delays that
application packets endure before a response is
returned to user - delays - client side think time,
- network delays in forwarding packets,
- server side processing time etc
- Two Definitions
- Network Utilization
- current use of available bandwidth, 10, 20 etc
- ability to provide additional bandwidth on demand
- ability to provide dedicated bandwidth to an
application
49Network Performance
- good performance must be available day and night
- response time must be provided when needed most
- end of fiscal quarter
- end of month
- 9am - 11am each morning to handle Peak
Interactive Traffic Load etc - good performance must be maintained when
- additional users (remote or local) added to
network - new department moves onto network
- new application is deployed enterprise wide
- additional network printers installed etc
50Additional Users
- If additional users (remote or local) are to be
added to network - need to anticipate performance impact in advance
- if analysis indicates
- extra traffic can be carried -gt proceed
- response time will degrade -gt delay until
selection and implementation of appropriate
network upgrades
51Network Performance
- The ability to predict network performance allows
us to - Keep user response time low
- Increase user productivity
- Supply adequate network bandwidth
- Provide for future growth
- Ensure successful deployment of new applications
- Validate response time goals of new network
designs - Troubleshoot for bottlenecks
- Choose among several competing network
applications - Choose the best alternative network topology
52- What are the Costs of a Poorly Performing Network
?
53Costs of a Poorly Performing Network
- Putting a dollar figure on network down time is
an exercise in estimation - A Network is considered to be down when users can
no longer work - If the network is down
- A salesperson writing a quotation for a customer
risks losing business if the quotation cannot be
delivered on time - A point-of-sale clerk who cannot authorize a
customers credit card will lose a sale and
possibly a customer - Electronic mail between a contractor and a
supplier is delayed and a critical order date is
missed
54Costs of a Poorly Performing Network
- For many companies the enterprise network (their
computer systems and networks) has become Mission
Critical that is, failure or poor performance
means the company is out of business
55Costs of a Poorly Performing Network
- Large Australian company with offices across all
states and New Zealand - Has central corporate office and numerous branch
offices and manufacturing sites interconnected by
a WAN - Branch offices
- provide sales and support services to customers
- transmit sales orders to the corporate office
- communicate with manufacturing sites for product
support - Corporate office
- communicates with appropriate manufacturing sites
to schedule product shipments to customers
56Costs of a Poorly Performing Network
- The enterprise network is Mission Critical to
such a company. Performance problems can cost
dollars in lost sales. - Other mission critical systems ?
57Costs of a Poorly Performing Network
- Productivity - Loss Estimate
- Full-time administrative assistant 60 per hr
includes on-costs - 10 assistants
- Access to File server is slow in the afternoons
-gt 10 of an assistants time is unproductive - Estimate
- 48weeks x 5days/week x 8hrs/day 1920 hrs
- 0.10 x 1920 x 60 x 10 115,200 lost annually
58Costs of a Poorly Performing Network
- Productivity - Loss Estimate
- Software Company - pending product release
- 40 developers and testers
- 7 development systems acting as a repository for
source code and each equipped with software
development tools - Periodical build is done, all source code and
data files are compiled and linked to produce
final product - Turnaround time to be competitive these builds
should take less than 12 hrs so that testing can
resume in the morning - If builds start to take longer, the next release
will slip because developers cannot test until
late next morning
59Costs of a Poorly Performing Network
- Productivity - Loss Estimate
- Assume a new release will generate a 1 million
in revenue over 6 mths - A slip of 1 mth represents a direct loss of
167,000 - A slip in release date may also mean a competitor
takes market share - The loss of productivity due to a late delivery,
missed deadline, or an inoperative application is
just as real, though not easily estimated in
dollar terms
60Costs of a Poorly Performing Network
- Lost sales opportunities
- Low customer and user satisfaction
- Slipped schedules
- Low user morale
- Out of business
- Loss of market share
61Performance Studies
- Allow us to answer some Real World network
performance questions as exemplified by following
examples - How does User Response Time change ?
62How does User Response Time Change ?
- Situation Resource Utilization is low and there
are few active users. Examples - when there are only a few digital images accessed
per hour -gt Users do not have to
wait - when there is little new mail to process -gt the
Email system responds quickly - when there are only a few small reports printed
per hour -gt No waiting for
print jobs - In this situation Response time
- will remain constant
- is independent of the number of users, can add
more users without and increase
63How does User Response Time Change ?
- Situation Resource Utilization High and Many
active users - In this situation Response time
- Increases as new users are added
- will not remain constant
- Long queues and lengthy wait times, users
discover another way to measure response time -
their position in the print queue - Refer fig 1-2
- Queuing theory and simulation methods can provide
estimates of how response time behaves versus the
number of users
64How does User Response Time Change ?
- Situation Resource Utilization High and Many
active users - Given this situation Response Time increases with
increases in - the number of active users
- user workload intensity
- the volume of data moved
- queue depth
- level of network utilisation
65Should a Server be Upgraded or a Second Server
installed ?
- Response time includes delay introduced by the
server - As workload on the server increases, its
utilisation approaches 100 and user response
time degrades - Should we upgrade server by
- increasing number of disks
- installing faster disks
- replacing LAN adaptor with 32 bit version
- adding more RAM
- upgrading CPU
- Where is the bottleneck ?
66Should a Server be Upgraded or a Second Server
installed ?
- Another option is to purchase a second server and
redistribute the user workload between them - Queuing Theory can be used to answer this
question.
67Should a Server be Upgraded or a Second Server
installed ?
- Queuing Theory - it is always better to have one
fast server than two half speed servers - When the single fast server is run at
- low utilisation, its response time will be almost
half of the half-speed servers - high utilisation, its response time will be about
equal to a half-speed servers - What other issue may come into consideration ?
68How many PCs can be supported on the a LAN before
Response Time deteriorates ?
- Analysis
- PC generates about 1 utilisation
- Keep LAN utilisation below 40 as response time
becomes erratic as LAN utilisation exceeds about
40 - Users say response time is more important than
utilisation - If we know up to 50 of users are active at any
time, how many PCs can we attach to the LAN ? - A LAN analyser or an SNMP management station can
be used to get a more accurate measure of the
traffic generated by the PCs - What Media Access Control technique is being used
?
69How many PCs can be supported on the a LAN before
Response Time deteriorates ?
- Max Utilisation Users Active Nos of PCs
Traffic per PC - Max Utilisation 40/100
- Users Active 50/100
- Traffic per PC 1/100
70Can we add a NEW application to the Enterprise
Network ?
- The ability of the network to support additional
traffic comes under scrutiny when a new
application is to be deployed, - new application and database servers
- additional user workstations
- LAN and WAN traffic will increase.
- Analysis
- benchmark new application on an isolated LAN and
measure LAN traffic generated by each transaction - determine location of application and database
servers - determine number and type of transactions that
will occur at each location
71Can we add a NEW application to the Enterprise
Network ?
- Analysis
- determine network topology
- measure existing work traffic flows
- Input topology and traffic details into a
simulation tool - perform simulation to provide response time
figures for each user site and utilisation
figures for WANs and LANs - Result - a reproducible documented analysis that
answers how the new application can be deployed
72How do we decide WAN Link Speeds ?
- Available Bandwidth
- varies with offered traffic load
- goes to near-zero at saturation, response time
becomes more unpredictable at this point, ie the
variation of response time around its mean
increases dramatically - Network designers can provide and control
bandwidth to achieve a certain level of response
time. - eg they can fine-tune a routers priority
queues to favour - interactive over bulk traffic
73How do we decide WAN Link Speeds ?
- Need to consider
- network topology
- location of clients and servers
- work habits of the users
- response time requirements
- throughput requirements
- transaction traffic loads offered by clients and
servers - Using techniques such as mean-value analysis,
worst-case analysis, queuing theory, or
simulation, link speeds can be adjusted to meet
the requirements