This paper introduces a framework for evaluating replica placement algorithms (RPA) for content delivery networks (CDN) as well as RPAs from other fields that might be applicable to current or future CDNs. First, the framework classifies and qualitatively compares RPAs using a generic set of primitives that capture problem definitions and heuristics. Second, it provides estimates for the decision times of RPAs using an analytic model. To achieve accuracy, the model takes into account disk accesses and message sizes, in addition to computational complexity and message numbers that have been considered traditionally. Third, it uses the "goodness" of produced placements to compare RPAs even when they have different problem definitions. Based on these evaluations, we identify open issues and potential areas for future research.

The proliferation of the Internet is leading to high expectation on the fast turnaround time. Clients abandoning their connections due to excessive downloading delays translates directly to profit losses. Hence, minimizing the latency perceived by end-users has become the primary performance objective compared to more traditional issues, such as server utilization. The two promising techniques to improve the Internet responsiveness are caching and replication. In this paper we present an overview of recent research in replication. We begin by arguing on the important role of replication in decreasing client perceived response time and proceed by illustrating the main topics that affect its successful deployment on the Internet. We analyze and characterize existing research, providing taxonomies and classifications whenever possible. Our discussion reveals several open problems and research directions. 1

The Replica Placement Problem (RPP) aims at creating a set of duplicated data objects across the nodes of a distributed system in order to optimize certain criteria. Typically, RPP formulations fall into two categories: static and dynamic. The first assumes that access statistics are estimated in advance and remain static, and, therefore, a one-time replica distribution is sufficient (1RPP). In contrast, dynamic methods change the replicas in the network potentially upon every request. This paper proposes an alternative technique, named Continuous Replica Placement Problem (CRPP), which falls between the two extreme approaches. CRPP can be defined as: Given an already implemented replication scheme and estimated access statistics for the next time period, define a new replication scheme, subject to optimization criteria and constraints. As we show in the problem formulation, CRPP is different in that the existing heuristics in the literature cannot be used either statically or dynamically to solve the problem. In fact, even with the most careful design, their performance will be inferior since CRPP embeds a scheduling problem to facilitate the proposed mechanism. We provide insight on the intricacies of CRPP and propose various heuristics.

The placement of data to maximize the performance and minimize the cost of a computing system is an optimization problem that has been studied extensively in several fields, including distributed databases, storage systems and, more recently, content delivery networks. However, little has been done to compare the various approaches and their applicability to different systems.

We consider the consistency control problem for replicated data in a distributed computing system (DCS) and propose a new algorithm to dynamically regenerate copies of data objects in response to node failures and network partitioning in the system. The DCS is assumed to have strict consistency constraints for data object copies. The new algorithm combines the advantages of voting based algorithms and regeneration mechanisms to maintain mutual consistency of replicated data objects in the case of node failures and network partitioning. Our algorithm extends the feasibility of regeneration to DCS on wide area networks, and is able to satisfy user queries as long as there is one current partition in the system. A stochastic availability analysis of our algorithm shows that it provides improved availability as compared to previously proposed dynamic voting algorithms. 1 Introduction In a distributed computing environment, two types of failures may occur: the processor at a given site may...

Caching and replication have emerged as the two primary techniques for reducing the delay experienced by end-users when downloading web pages. Even though these techniques may benefit from each other, previous research work tends to focus on either one of them separately. In particular, caching has been studied mostly in the context of proxy server systems, while replication is the technology behind Content Distribution Networks (CDNs). In this paper we investigate the potential performance gain by using a CDN server both as a replicator and as a proxy server. We develop an analytical model to quantify the benefit of each technique, under various system parameters, and propose a greedy algorithm to solve the combined caching and replica placement problem. Our simulation results indicate that a simple LRU caching scheme can improve significantly the response time of HTTP requests, when utilized over a replication-based infrastructure. Moreover, due to its simplicity, this hybrid approach does not affect the administrative overhead of the CDN architecture. q 2005 Elsevier B.V. All rights reserved.

Caching and replication have emerged as the two primary techniques for reducing the delay experienced by end users when downloading web pages. Even though these techniques may benefit from each other, previous research work tends to focus on either one of them separately. In this paper we investigate the potential performance gains by using a CDN server both as a replicator and as a proxy server. We assume a common storage space for both techniques, and develop an analytical model that characterizes caching performance under various system parameters. Based on the models predictions, we can reason whether it is beneficial to reduce the caching space in order to allocate extra replicas. The resulting problem of finding which object replicas should be created where, given that any free space will be used for caching, is NP-complete. Therefore, we propose a hybrid heuristic algorithm (based on the greedy paradigm), in order to solve the combined replica placement and storage allocation problem. Our simulation results indicate that a simple LRU caching scheme can considerably improve the response time of HTTP requests, when utilized over a replication-based infrastructure. 1.

Abstract—This paper presents data management and replication techniques used by the SNSP, a distributed operating system for sensor networks. Three replication schemes, a deterministic, a distributed probabilistic and an adaptive observation based scheme are compared. The first two were adapted from related work and the latter was developed for the SNSP. Results indicate that the distributed scheme has the best performance in terms of total cost per data access as well as increased data availability. When access cost without overhead is considered, the adaptive algorithm performs the best, but its overhead is higher because data items are replicated independently of accesses. Keywords-File allocation; Distributed operating system I.