What TCP/IP protocol headers can tell us about the web

Smith, F. Donelson; Campos, Félix Hernández; Jeffay, Kevin; Ott, David E.

doi:10.1145/378420.378789

Cited by 109 publications

(101 citation statements)

References 21 publications

Supporting

Mentioning

101

Contrasting

Order By: Relevance

“…The methodology we adopt to construct web model is closer to the work of Smith et al [12]. Additionally, we also model path characteristics and provide more comprehensive validation mechanism.…”

Section: Related Workmentioning

confidence: 97%

“…To reconstruct the data exchanges in the HTTP connections, we adopt a similar approach and heuristics from previous work [12]. One observation in their study is that when the server receives a HTTP request it will send TCP acknowledgments (ACKs) indicating the in-order byte sequence it has received, and all of the request messages will be ACKed before the corresponding HTTP response data is sent (note that here we assume there is no pipelining in use).…”

Section: Design Of Rampmentioning

confidence: 99%

“…For collection of traces we would like to thank the help of Vern Paxson of ICIR, and Alefiya Hussain and Jim Koda of ISI. We would like to thank F. Donelson Smith and Felix Hernandez Campos for providing part of their codes to us for reconstructing HTTP connection information [12]. We would also like to thank Purushotham Kamath, Alefiya Hussain and Debojyoti Dutta for providing helpful feedback from using RAMP in their work [20,21], and Tim Buchheim for helping integrate RAMP with other research activities.…”

Section: Acknowledgementsmentioning

confidence: 99%

See 2 more Smart Citations

A tool for RApid model parameterization and its applications

Lan

Heidemann

2003

Proceedings of the ACM SIGCOMM Workshop on Models, Methods and Tools for Reproducible Network Research - MoMeTools '03

View full text Add to dashboard Cite

The utility of simulations and analysis heavily relies on good models of network traffic. However, it is difficult to model and simulate the Internet traffic because of the network's great heterogeneity and rapid change. The statistical properties of Internet traffic not only constantly change over time but also vary in other dimensions such as locations and directions. Previously we have developed a tool RAMP that supports rapid parameterization of traffic models from live network measurements. In this paper, we first extend RAMP to support near-real-time trace-driven simulation. Next, we demonstrate the applications of RAMP via three case studies: generation of realistic traffic model for simulation, generation of high bandwidth synthetic network traces, and analysis and modeling of malicious traffic. Finally, we discuss some lessons we learned from using RAMP for traffic modeling.

show abstract

Section: Related Workmentioning

confidence: 97%

Section: Design Of Rampmentioning

confidence: 99%

Section: Acknowledgementsmentioning

confidence: 99%

See 1 more Smart Citation

A tool for RApid model parameterization and its applications

Lan

Heidemann

2003

Proceedings of the ACM SIGCOMM Workshop on Models, Methods and Tools for Reproducible Network Research - MoMeTools '03

View full text Add to dashboard Cite

show abstract

“…For brevity, we only discuss two sets of traces-NLANR traces collected over 24 hours from a link connecting New Zealand to US [18], and one hour long traces collected from the high-speed link connecting University of North Carolina at Chapel Hill (UNC) to its network service provider [20]. These traces contain 60-100 million packets.…”

Section: Packet Tracesmentioning

confidence: 99%

A case for run-time adaptation in packet processing systems

Kokku

Riché

Kunze

et al. 2004

SIGCOMM Comput. Commun. Rev.

View full text Add to dashboard Cite

Most packet processing applications receive and process multiple types of packets. Today, the processors available within packet processing systems are allocated to packet types at design time. In this paper, we explore the benefits and challenges of adapting allocations of processors to packet types in packet processing systems. We demonstrate that, for all the applications and traces considered, run-time adaptation can reduce energy consumption by 70-80% and processor provisioning level by 40-50%. The adaptation benefits are maximized if processor allocations can be adapted at fine timescales and if the total available processing power can be allocated to packet types in small granularities. We show that, of these two factors, allocating processing power to packet types in small granularity is more important-if the allocation granularity is large, then even a very fine adaptation time-scale yields meager benefits.

show abstract

“…The MTU (Maximum Transfer Unit) was set to 1448 bytes. For modelling HTTP traffic, we adopted the HTTP traffic model according to a latest study on HTTP traffic measurement [28]. The distribution of interpage and interobject time, number of objects within a Web page and size of an object are given in Table I.…”

Section: Test Methodologymentioning

confidence: 99%