Heuristic Algorithm for Calculating the Size of Tiers in a Hierarchical Storage System
Abstract
Introduction: A hierarchical storage system which serves many applications is discussed. Levels in the hierarchy differ in their memory type and response time. For each application, a certain quality of the service is assigned in terms of latency. The allocation of the address space among applications and memory levels affects the response time of the system. The purpose of the work is automatic memory allocation at every level taking into account the request flow characteristics in order to meet the service quality demands. Results: A memory partitioning optimization problem is formulated, and a heuristic algorithm is proposed for its solution. The cache hit rate and distribution function of popularity of the requested memory blocks are used as flow characteristics. An assessment of the storage system efficiency is calculated by simulating a storage system fed by a request flow saved from real storage systems. The calculated assessments show up to 10% gain for the proposed algorithm. Practical relevance: The proposed algorithm can improve the service quality in a hierarchical storage system.Published
2015-10-20
How to Cite
Malichenko, D. (2015). Heuristic Algorithm for Calculating the Size of Tiers in a Hierarchical Storage System. Information and Control Systems, (5), 100-105. https://doi.org/10.15217/issn1684-8853.2015.5.100
Issue
Section
Information coding and transmission