Replication is a well-known technique to improve the accessibility of Web sites. It generally offers reduced client latencies and increases a site’s availability. However, applying replication techniques is not trivial, and various Content Delivery Networks (CDNs) have been created to facilitate replication for digital content providers. The

The rapid growth of time-critical information services and business-oriented applications is making quality of service (QoS) support increasingly important in content distribution. This paper investigates the problem of placing object replicas (e.g., web pages and images) to meet the QoS requirements of clients with the objective of minimizing the replication cost. We consider two classes of service models: replica-aware service and replica-blind service. In the replica-aware model, the servers are aware of the locations of replicas and can therefore direct requests to the nearest replica. We show that the QoS-aware placement problem for replica-aware services is NP-complete. Several heuristic algorithms for efficient computation of suboptimal solutions are proposed and experimentally evaluated. In the replica-blind model, the servers are not aware of the locations of replicas or even their existence. As a result, each replica only serves the requests flowing through it under some given routing strategy. We show that there exist polynomial optimal solutions to the QoS-aware placement problem for replicablind services. Efficient algorithms are proposed to compute the optimal locations of replicas under different cost models.

The addition of storage capacity in network nodes for the caching or replication of popular data objects results in reduced end-user delay, reduced network tra#c, and improved scalability.

Clusters of Web caches are extensively used by different types of organizations, including companies, universities, ISPs, and CDNs. To model Web caches, we must account for two types of stochastic events: objects being pulled into/out of the cache cluster at random times, and caches going up and down at random times. Detailed stochastic models of such complex systems quickly become intractable. In this paper we propose a stochastic fluid model which captures the salient characteristics of a cache cluster. The stochastic fluid model replaces the object arrivals to the cluster and departures (object modification/expiration) with a fluid flow, but maintains the up/down dynamics of the original system. The model can be applied to a variety of cluster routing policies, and provides a simple means to estimate the hit rate. We compare the results of the stochastic fluid model with that of a simulation of the real system. We find the fluid model to not only be a close approximation, but also to exhibit the key qualitative properties of the original system. We conclude that stochastic fluid models show great potential in modeling a variety of content distribution systems.

Abstract — Although cooperation generally increases the amount of resources available to a community of nodes, thus improving individual and collective performance, it also allows for the appearance of potential mistreatment problems through the exposition of one node’s resources to others. We study such concerns by considering a group of independent, rational, self-aware nodes that cooperate using on-line caching algorithms, where the exposed resource is the storage of each node. Motivated by content networking applications – including web caching, CDNs, and P2P – this paper extends our previous work on the off-line version of the problem, which was limited to object replication and was conducted under a game-theoretic framework. We identify and investigate two causes of mistreatment: (1) cache state interactions (due to the cooperative servicing of requests) and (2) the adoption of a common scheme for cache replacement/redirection/admission policies. Using analytic models, numerical solutions of these models, as well as simulation experiments, we show that online cooperation schemes using caching are fairly robust to mistreatment caused by state interactions. When this becomes possible, the interaction through the exchange of miss-streams has to be very intense, making it feasible for the mistreated nodes to detect and react to the exploitation. This robustness ceases to exist when nodes fetch and store objects in response to remote requests, i.e., when they operate as Level-2 caches (or proxies) for other nodes. Regarding mistreatment due to a common scheme, we show that this can easily take place when the “outlier ” characteristics of some of the nodes get overlooked. This finding underscores the importance of allowing cooperative caching nodes the flexibility of choosing from a diverse set of schemes to fit the peculiarities of individual nodes. To that end, we outline an emulation-based framework for the development of mistreatment-resilient distributed selfish caching schemes.

Large scale hierarchical caches for web content have been deployed widely in an attempt to reduce delivery delays and bandwidth consumption and also to improve the scalability of content dissemination through the world wide web. Irrespectively of the specific replacement algorithm employed in each cache, a de facto characteristic of contemporary hierarchical caches is that a hit for a document at an $l$-level cache leads to the caching of the document in all intermediate caches (levels $l-1,\ldots,1$) on the path towards the leaf cache that received the initial request. This paper presents various algorithms that revise this standard behavior and attempt to be more selective in choosing the caches that get to store a local copy of the requested document. As these algorithms operate independently of the actual replacement algorithm running in each individual cache, they are referred to as \emph{meta algorithms}. Three new meta algorithms are proposed and compared against the de facto one and a recently proposed one by means of synthetic and trace-driven simulations. The best of the new meta algorithms appears to be leading to improved performance under most simulated scenarios, especially under a low availability of storage. The latter observation makes the presented meta algorithms particularly favorable for the handling of large data objects such as stored music files or short video clips. Additionally, a simple load balancing algorithm that is based on the concept of meta algorithms is proposed and evaluated. The algorithm is shown to be able to provide for an effective balancing of load thus possibly addressing the recently discovered ``filtering-effect'' in hierarchical web caches. -level cache leads to the caching of the document in all intermediate caches (levels########## ) on the path towards the leaf cache that received the initial request. This paper presents various algorithms that revise this standard behavior and attempt to be more selective in choosing the caches that get to store a local copy of the requested document. As these algorithms operate independently of the actual replacement algorithm running in each individual cache, they are referred to as meta algorithms. Three new meta algorithms are proposed and compared against the de facto one and a recently proposed one by H. Che, Y. Tung, and Z. Wang [1] by means of synthetic and trace-driven simulations. The best of the new meta algorithms appears to be able to lead to improved performance under most simulated scenarios, especially under a low availability of storage. The latter observation makes the presented meta algorithms particularly favorable for the handling of large data objects such as stored music files or short video clips. Additionally, a simple load balancing algorithm that is based on the concept of meta algorithms is proposed and evaluated. The algorithm is shown to be able to provide for an effective balancing of load thus possibly addressing the recently discovered "filtering-effect" in hierarchical web caches (C. Williamson [2]).

Cooperating proxy caches are groups of HTTP proxy servers that organize to share cached objects. This paper develops analytical models for proxy cooperation based on speedup in user response time. Speedup expressions are derived for the cooperation upper bound, a proxy mesh, and a three-level proxy hierarchy. The equations allow comparisons of fundamental design choices by separating the delivery organization from the discovery mechanism. For the mesh and hierarchy models, the discovery mechanisms include ideal discovery, ICP query, and distributed metadata directories. Equations are evaluated using parameter estimates from experiments and cache trace analysis. Results indicate that proxy cooperation is marginally viable from the standpoint of average user response time, and that the miss penalty for the hierarchy renders it less viable than the mesh. Proxy cooperation can, however, reduce the variability in user response time and the number of long delays. A trace-driven simulation shows that caching constraints have little effect on cooperation performance due to request filtering by lower level caches.

A Bloom filter is a simple randomized data structure that answers membership query with no false negative and a small false positive probability. It is an elegant data compression technique for membership information and has broad applications. In this paper, we generalize the traditional Bloom filter to Weighted Bloom Filter, which incorporates the information on the query frequencies and the membership likelihood of the elements into its optimal design. It has been widely observed that in many applications, some popular elements are queried much more often than the others. The traditional Bloom filter for data sets with irregular query patterns and non-uniform membership likelihood can be further optimized. We derive the optimal configuration of the Bloom filter with query-frequency and membershiplikelihood information, and show that the adapted Bloom filter always outperforms the traditional Bloom filter. Under reasonable frequency models such as the step distribution or the Zipf’s distribution, the improvement of the false positive probability of the weighted Bloom filter over that of the traditional Bloom filter has been evaluated by simulations.

This paper studies the optimal placement of web files for en-route web caching. It is shown that existing placement policies are all solving restricted partial problems of the file placement problem, and therefore give only sub-optimal solutions. A dynamic programming algorithm of low complexity which computes the optimal solution is presented. It is shown both analytically and experimentally that the file-placement solution output by our algorithm outperforms existing enroute caching policies. The optimal placement of web files can be implemented with a reasonable level of cache coordination and management overhead for en-route caching; and importantly, it can be achieved with or without using data prefetching.

Although coope"+E=@ ge+E=@DT incre@DT the amount ofre@zO"@D available to a community of node" thus improving individual andcollezED e pezzO@DT@=fi it also allows for the appe arance of pote tial mistreDT@z t proble" throughthe ee osition of one nodefi refi=fi+O toothefifl We study suchconce@D by consideT@fl a group ofinde e nde t, rational, se"fiO ware node that coopeEz" using on-line caching algorithms,wheo the eh ose reorit isthe storage at eD h no de Motivate by content networking applications -- including we b caching, C Ns, and P2P -- this pap e eDfi""" our prez=z" work onthe o#-line ve"=OO ofthe probleD which was conducte unde agameEDTEzO++D frame ork, andlimite to objeE re@Oz@DTEflz We ide tify and invefi""zzD twocause of mistre+DTE t: (1) interactions(due tothe coopeD@zz e se@zz=fiD ofrefiOzEfiD and (2)the adoption of a common scheme for cache managefiD t policiez Using analytic mode"" numeEfiED solutions ofthe= modeE+ as we@ as simulationea eatio ts, we show that on-line coope"zzOO scheEE using caching are fairly robust to mistreDT"z t cause bystate intefl@""DT"fl To app eD in a substantial manneE the intefi"==ED throughthe eh hange of miss-stre@= has to be veE inte@"fl making itfe@+E=O forthe mistreODT node to de"fi+ andreD" toeE@fifi=DT""=z This robustneT cestn toefiEO whe node fefl h andstore obje"+ inre" onse torez"fl rez"flfiDT i.e.,whe the opefi@= as Lee l-2 cache (or proxieO forothe nodeE ReEz+E"D mistreT+"fi tdue to a common schefi= we show that this caneDE"fl take place whe the "outlie" characteTO"zzD of some # E-mail:{nlaout, gsmaragd, beE" matta}@cs.bu.e=O istavrak@di.uoa.gr + Computer Scuter Dept, Boston University, Boston,Massac husetts, USA.