Performance comparison of a Web cache simulation framework
25 Mar 2005-Vol. 2, pp 281-284
TL;DR: This paper presents a powerful framework to simulate Web proxy cache systems and provides a comfortable environment to simulate and explore cache management techniques, and analyzed the details of a commercial proxy cache system.
Abstract: Performance comparison studies are primarily carried out through real systems or simulation environments. Simulation is the most commonly used method to explore new proposals due to both its flexibility and the relatively reduced time taken to obtain performance results. This paper presents a powerful framework to simulate Web proxy cache systems. Our tool provides a comfortable environment to simulate and explore cache management techniques. In order to validate our framework and show how accurate it executes, a performance comparison has been done. We analyzed the details of a commercial proxy cache system and compare its results with those obtained from our simulator using the most commonly replacement algorithm (LRU). For this purpose, the proposed environment was adapted to match the performance of the real proxy cache. Experimental results show that proxy cache hit ratio deviations fall very close to the real system, since then, never exceeds 3.42%.
Citations
More filters
TL;DR: This paper introduces and studies a caching algorithm that tracks the popularity of objects to make intelligent caching decisions and shows that when its parameters are set equal or close to their optimal values this algorithm outperforms traditional algorithms as LRU (least-recently used) and LFU (le least-frequently used).
Abstract: Due to its native return channel and its ability to easily address each user individually an IPTV system is very well suited to offer on-demand services. Those services are becoming more popular as there is an undeniable trend that users want to watch the offered content when and where it suits them best. Because multicast can no longer be relied upon for such services, as was the case when offering linear-programming TV, this trend risks to increase the traffic unwieldy over some parts of the IPTV network unless caches are deployed in strategic places within it. Since caches are limited in size and the popularity of on-demand content is volatile (i.e., changing over time), it is not straightforward to decide which objects to cache at which moment in time. This paper introduces and studies a caching algorithm that tracks the popularity of objects to make intelligent caching decisions. We will show that when its parameters are set equal or close to their optimal values this algorithm outperforms traditional algorithms as LRU (least-recently used) and LFU (least-frequently used). After a generic study of the algorithm fed by a user demand model that takes the volatility of the objects into account we will discuss two particular cases of an on-demand service, video-on-demand and catch-up TV, for each of which we give guidelines on how to dimension their associated caches.
67 citations
Cites background from "Performance comparison of a Web cac..."
...from observing requests on an existing web server [7]....
[...]
01 Dec 2009
TL;DR: This work aims at implementing SEMALRU- A Semantic and Least Recently Used Web cache replacement policy, which tunes the performance of the existing replacement algorithms through analyzing the semantic content of the document and the recency of the documents.
Abstract: The increasing demand for World Wide Web (WWW) services has made document caching a necessity to decrease download times and reduce Internet traffic. This work aims at implementing SEMALRU- A Semantic and Least Recently Used Web cache replacement policy. The basic LRU replacement policy is augmented with Semantic content of web pages to improve the efficiency of the replacement algorithms in terms of Hit Rate and Byte Hit Rate and to minimize the number of replacements made in cache. There are many well-known cache replacement policies based on size, recency, and frequency. This new improvised replacement policy attempts to implement cache replacement based on two parameters namely the semantics of the contents of web pages and the time of last access of the document. SEMALRU evicts documents that are less related to an incoming document or least recently used document which needs to be stored in the cache. This makes sure that only related documents are stored in the cache; hence the contents of the cache represent the documents of interest to the user and then ranked by recency. This policy tunes the performance of the existing replacement algorithms through analyzing the semantic content of the document and the recency of the document. A detailed algorithm to identify unrelated documents and documents that are least recently used has been devised. The policy was tested in a simulated environment with the related and unrelated set of user access pattern. . The parameters pertinent to cache replacement algorithms are computed and the results showing the improvement in the efficiency of the algorithm are furnished.
13 citations
03 Jun 2009
TL;DR: It is shown that when services like PLTV and CUTV gain in popularity, the transport capacity required in certain parts of the network risks to grow unwieldy, unless the content is replicated (i.e., cached) in the appropriate places in the network.
Abstract: One of the biggest advantages of interactive television (TV) is that it allows the viewer to watch the content at his or her most convenient time, either by pausing an ongoing broadcast or by selecting to view the content at a time later than the original airing time. The former service, often referred to as Pause Life TV (PLTV), and the latter, often referred to as Catch-Up TV (CUTV), require that an individual unicast flow is set up per user, whereas for the traditional Linear Programming TV (LPTV) the user just tunes in to a multicast flow that can serve many viewers. We first show that when services like PLTV and CUTV gain in popularity, the transport capacity required in certain parts of the network risks to grow unwieldy, unless the content is replicated (i.e., cached) in the appropriate places in the network. Subsequently, we show that a good caching algorithm that tracks the evolving popularity of the content and takes into account the initial popularity, allows keeping the required capacity under control. Finally, we discuss the trade-offs involved in determining the optimal cache location.
10 citations
Additional excerpts
...In some studies this volatility of objects has been taken into account by feeding the caching algorithm with a trace obtained from observing requests on an existing web server [4]....
[...]
TL;DR: It is demonstrated that with the growing popularity of the CUTV service, the number of simultaneously running unicast flows on the aggregation parts of the network threaten to lead to an unwieldy increase in required bandwidth.
Abstract: The catch-up TV (CUTV) service allows users to watch video content that was previously broadcast live on TV channels and later placed on an on-line video store. Upon a request from a user to watch a recently missed episode of his/her favourite TV series, the content is streamed from the video server to the customer's receiver device. This requires that an individual flow is set up for the duration of the video, and since it is hard to impossible to employ multicast streaming for this purpose (as users seldomly issue a request for the same episode at the same time), these flows are unicast. In this paper, we demonstrate that with the growing popularity of the CUTV service, the number of simultaneously running unicast flows on the aggregation parts of the network threaten to lead to an unwieldy increase in required bandwidth. Anticipating this problem and trying to alleviate it, the network operators deploy caches in strategic places in the network. We investigate the performance of such a caching strategy and the impact of its size and the cache update logic. We first analyse and model the evolution of video popularity over time based on traces we collected during 10 months. Through simulations we compare the performance of the traditional least-recently used and least-frequently used caching algorithms to our own algorithm. We also compare their performance with a "perfect" caching algorithm, which knows and hence does not have to estimate the video request rates. In the experimental data, we see that the video parameters from the popularity evolution law can be clustered. Therefore, we investigate theoretical models that can capture these clusters and we study the impact of clustering on the caching performance. Finally, some considerations on the optimal cache placement are presented.
8 citations
Additional excerpts
...In some studies this volatility of objects has been taken into account by feeding the caching algorithm with a trace obtained from observing requests on an existing web server [6]....
[...]
23 Apr 2012
TL;DR: A novel content caching scheme referred as “Cache Management using Temporal Pattern based Solicitation” (CMTPS), to further minimize both service delays and load in the network for Video on Demand (VoD) applications is proposed.
Abstract: Caching is an effective technique to improve the quality of streaming multimedia services. In this paper, we propose a novel content caching scheme referred as “Cache Management using Temporal Pattern based Solicitation” (CMTPS), to further minimize both service delays and load in the network for Video on Demand (VoD) applications. CMPTS is based on the analysis of clients' requests over passed time intervals to predict the contents that will be solicited in the near future. By means of experimental tests, the CMTPS protocol is evaluated. The obtained results show that CMTPS outperforms LRU, in terms of peak traffic reduction and number of cache hits.
6 citations
Cites methods from "Performance comparison of a Web cac..."
...objects had been taken into a algorithm with traces obtained existing server [5][14]....
[...]
References
More filters
DOI•
01 Aug 2001
TL;DR: NCS (Network and Cache Simulator) is an HTTP trace-driven discrete event simulator of network and caching activity that resides between the high-level caching-only simulators prominant in much Web caching research, and the detailed simulators of networking protocols and traffic.
Abstract: NCS (Network and Cache Simulator) is an HTTP trace-driven discrete event simulator of network and caching activity. It is highly parameterized for maximal compatibility with previous caching and network simulations. In granularity, it resides between the high-level caching-only simulators prominant in much Web caching research, and the detailed simulators of networking protocols and traffic. In an effort to capture estimates of user-perceived latency, it simulates simplistic caching and prefetching functionality at various locations in a network comprising of client browsers, an optional intermediate proxy, and Web servers. Caching at the proxy and clients is optional. Additionally, it simulates many aspects of TCP traffic among these entities on a somewhat idealized network. In this report we motivate the development of NCS and describe its features and capabilities. We additionally provide a number of sample experiments showing the simulator’s utility in a variety of contexts.
15 citations
01 Jan 1999
TL;DR: This short note is written as a submission to the DARPA Workshop on Validation of Large Scale Network Simulation Models to identify issues critical to the validation of network models, and to develop potential strategies for the research community to address these issues.
Abstract: This short note is written as a submission to the DARPA Workshop on Validation of Large Scale Network Simulation Models. The purpose of the workshop is to identify issues critical to the validation of network models, and also to develop potential strategies for the research community to address these issues. 1 Current network simulation models
13 citations
15 Aug 2001
TL;DR: This paper argues for the validity of a new network and caching simulator by extensively comparing simulated results to both small- and large-scale realworld HTTP traffic and some of the preparation needed to use a large, well-known trace of Web usage.
Abstract: Simulation is a common technique used by Web researchers and provides many benefits. Verification and validation of a simulator are essential if the results of those simulations are to be believed. Unfortunately, even limited validation of simulators has been uncommon in the Web caching community. In contrast, this paper argues for the validity of a new network and caching simulator by extensively comparing simulated results to both small- and large-scale realworld HTTP traffic. In addition, we describe some of the preparation needed to use a large, well-known trace of Web usage.
8 citations