Optimal proactive caching in peer-to-peer network

Abstract

As a promising new technology with the unique properties like high efficiency, scalability and fault tolerance, Peer-to-Peer (P2P) technology is used as the underlying network to build new Internetscale applications. However, one of the well known issues in such an application (for example WWW) is that the distribution of data popularities is heavily tailed with a Zipf-like distribution. With consideration of the skewed popularity we adopt a proactive caching approach to handle the challenge, and focus on two key problems: where(i.e. the placement strategy: where to place the replicas) and how(i.e. the degree problem: how many replicas are assigned to one specific content)? For the where problem, we propose a novel approach which can be generally applied to structured P2P networks. Next, we solve two optimization objectives related to the how problem: MAX_PERF and MIN_COST. Our solution is called PoPCache, and we discover two interesting properties: (1) the number of replicas assigned to each content is proportional to its popularity; (2) the derived optimal solutions are related to the entropy of popularity. To our knowledge, none of the previous works has mentioned such results. Finally, we apply the results of PoPCache to propose a P2P base web caching, called as WebPoPCache. By means of web cache trace driven simulation, our extensive evaluation results demonstrate the advantages of PoPCache and Web-PoPCache.

Community-Based Caching for Enhanced Lookup Performance in P2P Systems

“…Consequently, n will observe a lower demand for kc, giving it the opportunity to cache ka. Therefore, in contrast to previous solutions [5], [6], [7], [9], local statistics are adequate to provide a customized service to each of the communities. Next, we discuss a distributed cache-capacityallocation mechanism based on the local statistics and structure of the overlay topology.…”

“…If key kL (indexed at node L) is requested more frequently than kK, nJ should cache kL and its corresponding value. Therefore, in contrast to previous solutions [5], [6], [7], [9], local statistics are adequate to determine what keys to cache at a node. Query arrivals in P2P systems show flash-crowds, and diurnal and seasonal effects [1], [5].…”

“…Query arrivals in P2P systems show flash-crowds, and diurnal and seasonal effects [1], [5]. Therefore, statistics such as periodic, network-wide query counts or arrival rate estimates, used in [5], [6], [7], [9], [19], are inadequate to decide effectively when and what to cache. Moreover, such sampling messages introduce a significant overhead.…”

“…Content popularity profiles in P2P systems follow a Zipf's-like distribution [4], [5], [6]. However, resources popularly shared within an individual community typically do not rank high in popularity in the context of the overall P2P system [2] and often are inconspicuous in the system-wide behavior.…”

“…Therefore, such communities are unable to benefit from performance enhancements such as caching and replication that focus only on the most popular resources. For example, Beehive [5] and PoPCache [6], [7] force a large fraction of peers (in a structured P2P system [8]) to cache the most popular resources regardless of their interests. In spite of requiring large caches and many probing messages to estimate the global popularity, these solutions are inconsiderate of moderately popular resources.…”