COMPUTER NETWORK: MODELING AND SIMULATION -Abhaykumar Kumbhar Computer Science Department.

COMPUTER NETWORK:MODELING AND SIMULATION

-Abhaykumar Kumbhar Computer Science Department

Motivation:

• Bridge between Real life application and Theoretical world.

Contents:• Introduction.• Modeling.• Simulation.• How to develop a Model.• Classification .• NS simulator.• OPNET.• An Ethernet Case Study.• Requirements of a good Network Simulating Tool• Evolution.• Conclusion and Future Work

Introduction:• Performance analysis of computer networks :

increase in size and geographical extent.• The size of the networks and the inherent

complexity of network protocols complicate this analysis.

• Analysis techniques such as queuing models have difficulty modeling dynamic behavior retransmission timeouts and congestion.

• Simulation offers a better method of studying computer networks, since one can simulate the details of actual protocols.

• Simulation is an analysis and evaluation tool.• Every simulation is tight to a model that

represents the real world.• Inconsistent modeling

• Modeling itself is a crucial step towards meaningful results.

• Aim :systematic modeling techniques -increase the quality and validity of results

Modeling:

• Constructing a conceptual framework that describes a system.

Modeling Taxonomy:

Simulation:

system under study(has deterministic rules governing its behavior)

observer

“real” life

system boundary

exogenous inputsto system

(the environment)

computer programsimulates deterministic rules governing behavior

pseudo random inputsto system

(models environment)

program boundary

observer

“simulated” life

Why Simulation?

• study system performance, operation• real-system not available, is complex/costly or

dangerous (eg: space simulations, flight simulations)

• quickly evaluate design alternatives (eg: different system configurations)

• evaluate complex functions for which closed form formulas or numerical techniques not available

When to Use Simulation?• Whenever Mathematical Analysis Is Difficult

or Impossible.

• For Validating Analytic Models.

• For Experimentation Without Disturbing an Operational System

How to develop a model:

• Determine the goals and objectives• Build a conceptual model• Convert into a specification model• Convert into a computational model• Verify• Validate

Simulation Model Development Methodology

Simulation Model Verification and Validation

Classification of Simulation Tools:• GPPL: General Purpose Programming Language• PSL: “Plain” Simulation Language• SP: Simulation Package

NS Simulator:

• Began as REAL in 1989• Developed by UC Berkeley• Public domain SP• Object-oriented• Written in C++ and object-oriented tcl (Otcl)• Network components are represented by

classes

Ns class hierarchy:

OPNET:

• Developed by OPNET Technologies Inc.• Commercial SP• Object-oriented• Totally menu-driven package• Built-in model libraries contain most popular

protocols and applications• Simulation task made easy

Modeling hierarchy in OPNET

An Ethernet Case Study:• Bridging the Gap Between Reality and

Simulations• We set up our test-bed using two nodes

connected by a 3 meter cross-over Ethernet cable.

• a simple test-bed experiment and attempt to replicate the obtained results by simulation in ns-2,QualNet and OPNET.

• Network details• set packet size of the application to 1472

bytes. This was to ensure that the maximum Ethernet frame of 1500 bytes would be transmitted

Delay experienced by packets in saturated case with non-blocking sockets

particularly interested in the saturation performance of the system and we chose multiple rates near the maximum link bandwidth

• The maximum delay of the system is experienced when the source bufferis full.

• This delay is given as max delay = tx delay ×buffer size in packets.

• The transmission delay for one packet is the minimum delay.

• The buffer size is then calculated as max delay/min delay. Hence a buffer size of 78 packets was arrived at 94/1.2 ≈ 78packets ≤ 128KB.

• Setting up simulations in QualNet:--the link propagation delay being negligible, set it to 25 μs to model test-bed network characteristics.--queue size from its default value of 50 KB to 117 KB, which corresponds to 78 packets of 1500 bytes each.

• Setting up simulations in OPNET– Ethernet Station Advanced (ESA).– The ESA has a traffic generator built over the

Ethernet MAC.– the queue length to infinity.– OPNET models queues as a set of sub-queues,

allowing the possibility of different application traffic to map on to different queues. By default only one sub-queue is created in the Ethernet MAC process model and this is a hidden parameter. This sub queue is accessed as a process interface of the Ethernet MAC process model. The sub-queue packet capacity was set to 78 packets.

• Setting up simulations in NS-2– configurable parameters: node objects, link

bandwidth, maximum length and type of the interface queue and delay.

– queue length to 78. – maximum propagation delay is 25 μs for 10 Mbps.

• Results:

Figure :Throughput

Figure: Packet Delivery Fraction

Figure : Average End-to-End Delay

Requirements for Network Simulation tools :

• Model development simplicity• Modeling flexibility• Fast modeling• Animation• Different kinds of implemented components• Component adaptability• Creating new components• Static capabilities of a simulator• Graphs

Evolution of Network Simulation Tools:

• “Zeroth ” Generation — General Purpose Languages- Fortran, C/C++, Pascal, Basic

• “First” Generation — General Purpose Queueing System Simulations

-GPSS, SLAM, SIMSCRIPT

• “Second” Generation — Application Specific: Computer Systems and Wide-Area Communication Networks

• “Third” Generation — Integration of Second Generation Languages-With a Graphics-Oriented Analysis and Modeling Environment

-SES/Workbench-OPNET

Conclusion and Future work:

• Best way to learn Protocols.

References:

1:Advanced Modelling and Simulation Methodsfor Communication Networks Jože Mohorko, Fras Matjaž, Klampfer Saša.

2:www.sciencedirect.com3:www.ieeexplore.org

• THANK YOU

–QUESTIONS??