Storage Capacity Allocation Algorithms for Hierarchical Content Distribution

Abstract

Abstract: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 traffic, and improved scalability. The problem of allocating an available storage budget to the nodes of a hierarchical content distribution system is formulated; optimal algorithms, as well as fast/efficient heuristics, are developed for its solution. An innovative aspect of the presented approach is that it combines all relevant subproblems, concerning node locations, node sizes, and object placement, and solves them jointly in a single optimization step. The developed algorithms may be utilized in content distribution networks that employ either replication or caching/replacement.

Meta algorithms for hierarchical Web caches

“…Thus, networks that may tolerate the 10% increase in network traffic may achieve a reduced overall delay by employing the load balancing mechanism. These results appear to be consistent with our previous results pertaining to the tradeoff between load balancing and hit distance, for the case of object replication [20].…”

“…The average number of hops to achieve a request hit is used to compare the performance of the various meta algorithms. For simplicity only hierarchical request forwarding has been used; see [20] for the effect of request peering between sibling caches. As the focus of this work is on the meta algorithms and not on the replacement algorithms themselves, the standard LRU replacement is assumed to run in all caches.…”

“…The equal allocation of storage has been employed as a baseline storage allocation strategy, often applied in practice, under which the different meta algorithms are compared; see [20] for algorithms that optimize the allocation of ¡ to the nodes of the hierarchy. Before comparing the different meta algorithms we present a graph that pertains to the effect of the parameter of the Prob meta algorithm (Prob is the only parametric new algorithm).…”

Joint object placement and node dimensioning for Internet content distribution

“…Producing the required object placement not only identifies the set of information object for each node but also returns the fraction of the total storage capacity that must be allocated to each node. In [6] we have formulated a specific instance of this problem and proposed heuristic algorithms for it.…”