Simulationbased STRESS Testing Case Study: A Multicast Routing Protocol Systematic Testing of Robust - PowerPoint PPT Presentation

1 / 21
About This Presentation
Title:

Simulationbased STRESS Testing Case Study: A Multicast Routing Protocol Systematic Testing of Robust

Description:

Introduction of new services (e.g. IP multicast) ... tracer. Network. interface. Node. 16. USC INFORMATION SCIENCES INSTITUTE. Root. S1. C. A. B ... – PowerPoint PPT presentation

Number of Views:71
Avg rating:3.0/5.0
Slides: 22
Provided by: ahmed3
Category:

less

Transcript and Presenter's Notes

Title: Simulationbased STRESS Testing Case Study: A Multicast Routing Protocol Systematic Testing of Robust


1
Simulation-based STRESS Testing Case Study A
Multicast Routing ProtocolSystematic Testing of
Robustness by Examination of Selected Scenarios
  • Ahmed Helmy, Deborah Estrin
  • Computer Science Department/ISI
  • University of Southern California
  • ahelmy,estrin_at_usc.edu,
  • http//catarina.usc.edu/ahelmy,estrin

2
Outline
  • Motivation / Problem Statement
  • Related Work
  • The STRESS framework
  • Problem dimensions solution for PIM-SM
  • host events, topology, faults
  • Simulation framework
  • Case study results
  • Summary Future Work

3
Motivation and Problem
  • Approach
  • Develop systematic methodologies for testing
    Internet protocol robustness
  • Programmatically simulate failure modes to
    analyze behavior

4
Related Work
Tend to abstract out network failures Complexity
increases with multicast
5
STRESS Framework
6
STRESS block diagram
Loss Failures
Link Tracing
Routed Topology
Simulation set-up
Simulation Engine
Host Scenarios
Simulation Tracing
Scenario Generation
Output Analysis
7
Problem Dimensions
  • A Test Scenario may include
  • Routed Topology
  • Host Events
  • Faults (selective message loss)

Faults
Topology
Host Events
8
Representative Host Scenarios
  • Host Events simple scenario reduction filter
  • Send (S), Join (J1,J2), Leave (L1,L2)

9
Representative Topologies
  • Remove topologies that experience same errors and
    transitions
  • Transitions are dependent on protocol-specific
    mechanisms
  • Need protocol-specific analysis to get the
    representative topologies

10
Unicast vs. Multicast
S
S
R1
R1
R2
R2
R3
R3
R4
R4
Multiple unicasts
Multicast
11
PIM-SM Overview
S
R1
Root
R2
R3
R2
R1
R4
12
Topological Equivalence
  • N-router LAN topology, running PIM-SM

13
Equivalent Topologies
S1
R1
R2
R2
S2,R1
Topology 2
Topology 1
14
Simulation
  • Used Network Simulator (NSv2) and the VINT system
    (http//catarina.usc.edu/vint)
  • Added LANs, selective loss, protocol tracing and
    profiling support
  • Implemented PIM-SM shared tree mechanisms
    (http//catarina.usc.edu/pim)
  • Join/Prune, Assert, Register

15
Simulation Framework
16
Results Faulty Join Scenario
C
A
B
R1
R2
17
Results Fault Register Scenario
C
A
S2, R1
18
Results Coverage
Number of Scenarios with Selective Loss
Total 352 (1st topo) 296 (2nd topo) 648
Scenarios
Protocol Procedure Coverage
of 91 traced procedures
19
Summary
  • Presented STRESS framework for systematic testing
    of protocol robustness
  • Used representative scenarios and equivalent
    topologies to reduce space searched
  • Automated fault investigation through simulation
  • Captured several design errors in PIM-SM and
    suggested fixes

20
Future Work
  • Automate Scenario and Topology Generation using
    search techniques
  • Forward Search
  • Fault-oriented backward search
  • Achieve better completeness and coverage
  • Extend the framework for performance analysis of
    end-to-end multicast protocols

21
Thank You!
Merci
Ahmed Helmy ahelmy_at_usc.edu http//catarina.usc.edu
/ahelmy
Write a Comment
User Comments (0)
About PowerShow.com