A Longitudinal View of HTTP Traffic

Callahan, Tom; Allman, Mark; Paxson, Vern

doi:10.1007/978-3-642-12334-4_23

Cited by 28 publications

(22 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While the server needs to support HTTP/1.1 for persistent and pipelined request, the browser ultimately has to issue the request. Callahan et al [2] observed a drastic change in the number of requests per connection and attributes the change to a different default browser version in the environment they monitor.…”

Section: Impact Of Browsermentioning

confidence: 99%

Pitfalls in HTTP Traffic Measurements and Analysis

Schneider

Ager

Maier

et al. 2012

Passive and Active Measurement

View full text Add to dashboard Cite

Abstract.Being responsible for more than half of the total traffic volume in the Internet, HTTP is a popular subject for traffic analysis. From our experiences with HTTP traffic analysis we identified a number of pitfalls which can render a carefully executed study flawed. Often these pitfalls can be avoided easily. Based on passive traffic measurements of 20.000 European residential broadband customers, we quantify the potential error of three issues: Non-consideration of persistent or pipelined HTTP requests, mismatches between the Content-Type header field and the actual content, and mismatches between the Content-Length header and the actual transmitted volume. We find that 60 % (30 %) of all HTTP requests (bytes) are persistent (i. e., not the first in a TCP connection) and 4 % are pipelined. Moreover, we observe a Content-Type mismatch for 35 % of the total HTTP volume. In terms of Content-Length accuracy our data shows a factor of at least 3.2 more bytes reported in the HTTP header than actually transferred.

show abstract

Section: Impact Of Browsermentioning

confidence: 99%

Pitfalls in HTTP Traffic Measurements and Analysis

Schneider

Ager

Maier

et al. 2012

Passive and Active Measurement

View full text Add to dashboard Cite

show abstract

“…Studies are based on aggregated traffic captured at internet hubs or large organizations (e.g., universities) access links. These studies expose general trends in HTTP usage and workload [5], [2], or focus on some particular Web 2.0 applications such as interactive maps. [6], [7], [8], [9], [10] Our work comes as a complement, by providing a fine-grained study of the various Web applications.…”

Section: Introductionmentioning

confidence: 99%

On Traffic Patterns of HTTP Applications

Augustin

Mellouk

2011

2011 IEEE Global Telecommunications Conference - GLOBECOM 2011

View full text Add to dashboard Cite

Abstract-HTTP has been the most popular internet protocol for 30 years. Until recently, its role has been limited to a traditional transfer of hypertext documents. However, its flexibility and interoperability cause it to be progressively involved in a much wider range of applications, from video and audio streaming to email, chat and documents editing. Understanding the behavior of modern Web applications is a crucial step to apply QoS or security policies on this traffic. This paper studies 20 popular, Web applications that are representative of 12 application types. We describe a method to isolate and capture browser-generated traffic and plot time series with an RRDTool database. We show that modern Web applications present very diverse traffic patterns, and propose a description and classification of these patterns.

show abstract

“…Instead, we considered measuring median and average HTTP object download times, but find that latency is a reliable predictor for such times. (During preliminary testing we discovered that round-trip-time correlates to download time for a median-sized-400-byte [17]-object with a Pearson correlation coefficient of approximately 0.83, for both a default path as well as a poisoned path that affected 30% of the replica clients.) Hence, we measure three basic network performance primitives: latency, throughput, and jitter.…”

Section: Measuring Performance Improvementsmentioning

confidence: 98%

“…Google research showed that a 500-millisecond increase in latency caused a 20% traffic drop [8], while it was reported that a latency increase of 100 milliseconds can produce a 1% drop in revenue for Amazon [28]. Accordingly, recent years have seen many optimizations to accelerate the delivery of such services, ranging from better transport protocols [40] to browser enhancements [4,7] to new compression and site optimization algorithms [1,17].…”

Section: Introductionmentioning

confidence: 99%

Quantifying the benefits of joint content and network routing

Valancius

Ravi

Feamster

et al. 2013

SIGMETRICS Perform. Eval. Rev.

View full text Add to dashboard Cite

Online service providers aim to provide good performance for an increasingly diverse set of applications and services. One of the most effective ways to improve service performance is to replicate the service closer to the end users. Replication alone, however, has its limits: while operators can replicate static content, wide-scale replication of dynamic content is not always feasible or cost effective. To improve the latency of such services many operators turn to Internet traffic engineering. In this paper, we study the benefits of performing replica-to-end-user mappings in conjunction with active Internet traffic engineering. We present the design of PECAN, a system that controls both the selection of replicas ("content routing") and the routes between the clients and their associated replicas ("network routing"). We emulate a replicated service that can perform both content and network routing by deploying PECAN on a distributed testbed. In our testbed, we see that jointly performing content and network routing can reduce round-trip latency by 4.3% on average over performing content routing alone (potentially reducing service response times by tens of milliseconds or more) and that most of these gains can be realized with no more than five alternate routes at each replica.

show abstract

A Longitudinal View of HTTP Traffic

Cited by 28 publications

References 16 publications

Pitfalls in HTTP Traffic Measurements and Analysis

Pitfalls in HTTP Traffic Measurements and Analysis

On Traffic Patterns of HTTP Applications

Quantifying the benefits of joint content and network routing

Contact Info

Product

Resources

About