2016 Twenty Second National Conference on Communication (NCC) · 2016 · 13 citations · 3 references
Software MaintenanceNetwork TraceEngineeringTracereplay InfersSoftware EngineeringSimulationNetwork SimulatorSoftware AnalysisDelay-tolerant NetworkingApplication LayerSystems EngineeringModeling And SimulationNetwork PerformanceAdvanced NetworkingPhysical LayerComputer EngineeringApplication AnalysisComputer ScienceMobile ComputingSoftware DesignNetwork SimulationModel-based System EngineeringEdge ComputingSoftware TestingSimulation InfrastructureSystem Software
Ns-3 is a widely used as a the network simulator of choice by researchers. It contains many well tested and high quality models of network protocols. However, the application layer models of ns-3 are very simplistic, and do not capture all aspects of real life applications. As a result, there is often a huge gap between the results of real experiments and the corresponding simulations. This problem is particularly exacerbated for wireless simulations, where many networking phenomena like wireless channel contention crucially depend on the application traffic characteristics. One way to bridge the gap between experiments and simulations is to incorporate knowledge from network traces into simulations. To this end, our work builds a trace-based application layer simulator in ns-3. Given a network trace collected from a user, our TraceReplay application layer model automatically generates traffic that is faithful to the real application in the ns-3 simulator. TraceReplay infers and replays only application layer delays like user think times, lets the simulator control the lower layer phenomena. TraceReplay extracts application layer characteristics from a single trace, and replays this information across many users in simulation, by using suitable randomization. Our model is also generic enough to replay any application layer protocol. Validation of our simulation model shows that simulation results obtained using TraceReplay are significantly different from those using other models, and are closer to experimental observations.
3
A behavioral model of Web traffic
Hyoung-Kee Choi, J.O. Limb · 2005 · 201 citations
Web Traffic Model, Computational Social Science, Internet Traffic Analysis +14
Mahimahi: accurate record-and-replay for HTTP
Ravi Netravali, Anirudh Sivaraman, Somak Das et al. · 2015 · 185 citations
Network Conditions, Accurate Record-and-replay, Engineering +16
Michele C. Weigle, Prashanth Adurthi, Félix Hernández-Campos et al. · ACM SIGCOMM Computer Communication Review · 2006 · 130 citations