Citation Context

... pages are a more important dynamic in the Web. We discuss implications of these dierences in Sections 3 and 4. 2.3.3 Kazaa Is Not Zipf Much has been written about the Zipf-like qualities of the WWW [=-=7]. In -=-fact, researchers commonly quote the Zipf parameter of the popularity distributions seen in their traces [14, 27], in part to demonstrate that their results are \correct." This Zipf property of W...

Citation Context

...ffic, we explore two questions in this section. First, Zipf's law [16] is widely used to model skewed distributions because of its simple form and has been recently applied to research on Web caching =-=[17]-=- and Internet topology models [18]. Earlier work [19] suggests that the Zipf's law also applies to the signaling traffic for Gnutella. An interesting question is whether Zipf's law is suitable for mod...

Citation Context

...ues when using the algorithms. We then conclude in Section VIII. II. PREVIOUS WORK There has been a considerable amount of research on Web performance, ranging from Web workload characterization [3], =-=[4]-=-, [23] to developing techniques to enhance Web performance. Two primary techniques for enhancing Web performance are caching and replication. Previous work has studied many aspects of caching and repl...

Citation Context

...e approximate it with a beta distribution [10], B(x), (a = 0.12, fl = 4.2), shown by the dotted line in Figure 3. We use a beta distribution rather than the more commonly considered Zipf distribution =-=[6]-=- for two reasons: First, the Zipf distribution is discrete and applies to a specific count of objects, whereas we • want a continuous distribution that applies to an indeterminate count of objects. Se...

Citation Context

...ol, web clients running on users' machines request objects from web servers. Previous studies have examined many aspects of the web, including web workloads [2, 8, 15, 29], characterizing web objects =-=[3, 11]-=-, and even modeling the hyperlink structure of the web [6, 21]. These studies suggest that most web objects are small (5-10KB), but the distribution of object sizes is heavy-tailed and very large obje...

Citation Context

...cludes. 2 Related work Web tracing and caching are highly active research areas. Recent studies of Web traffic include analyses of Web access traces from the perspective of browsers [11, 26], proxies =-=[2, 5, 6, 7, 10, 13, 14, 18, 20, 23, 30]-=-, and servers [1, 4, 29]. Earlier tracing studies were limited in request rate, number of requests, and diversity of population. The most recent tracing studies have been larger and more diverse. In a...

Citation Context

... adversely affect the performance of the applications running over it. Plaxton et al. [1] provides an elegant algorithm that achieves nearly optimal latency on graphs that exhibit power-law expansion =-=[2]-=-, at the same time, preserving the scalable routing properties of the DHT-based system. However, this algorithm requires pair-wise probing between peers to determine latencies and it is unlikely to sc...

Citation Context

... at the prefix and AS level follows Zipf's law [19]. Studies of the World Wide Web have shown that a small fraction of the requests, resources, and servers are responsible for the bulk of the traffic =-=[20, 21]-=-. The small number of heavy hitters has important implications for traffic engineering. On the positive side, since the leading heavy hitters account for so much traffic, care in routing just these de...

Citation Context

... the use of appropriate cache management policies that appropriate for Web workloads. A number of measurements have been done to exploit the access properties at clients, proxies, and servers [1] [6] =-=[8]-=- [25] [26]. 5 Conclusion As Web service becomes more and more popular, users are suffering network congestion and server overloading. Great efforts have been made to improve Web performance. Web cachi...