In this paper we present an online method for managing a goal-oriented buffer partitioning in the distributed memory of a network of workstations. Our algorithm implements a feedback mechanism which dynamically changes the sizes of dedicated buffer areas and thereby the buffer hit rate for the different classes in such a way that user-specified response time goals are satisfied. The aggregated size of the buffer memory across all network nodes remains constant and only the partitioning is changed. The algorithm is based on efficiently approximating the trajectory of the per-class response time curves as a function of the available buffer. Changes in the workload that would lead to violation of response time goals are counteracted by accordingly adjusting the buffer allocation. For local replacement decisions, we integrate a cost-based buffer replacement algorithm to fit into our goal-oriented approach. We have implemented our algorithm in a detailed simulation prototype and we present first results obtained from this prototype.

Presents a method for distributed caching to exploit the aggregate memory of networks of workstations in data-intensive applications. In contrast to prior work, the approach is based on a detailed cost model as the basis for optimizing the placement of variable-size data objects in a distributed, possibly heterogeneous two-level storage hierarchy. To address the online problem with a priori unknown and evolving workload parameters, the method employs dynamic load tracking procedures and an approximative, low-overhead version of the cost model for continuous reoptimization steps that are embedded in the decisions of the underlying local cache managers. The method is able to automatically find a good tradeoff between an "egoistic" and an "altruistic" behavior of the network nodes, and proves its practical viability in a detailed simulation study under a variety of workload and system configurations.