Paper Info
Title
Analysis of interleaved storage via a constant-service queuing system with Markov-chain-driven input
Abstract
A popular means of increasing the effective rate of main storageaccesses in a large computer is a multiplicity of memory modulesaccessible in parallel. Although such an organization usuallyachieves a net gain in access rate, it also creates new modes ofcongestion at the storage controller. This paper analyzes thevariables that describe such a congestion: queue lengths anddelays. A controller that maintains separate register sets toaccommodate the request queue of each module is considered. Thevarious processors attached to the storage are assumed to generate,in each memory cycle, a number of access requests with the samegiven distribution. The addresses specified by these requests(reduced to the module index) are further assumed to follow thestates of a first-order Markov chain. The analysis then becomes oneof a single-server queuing system with constant service time andindexed batch arrival process. Results are derived for severaldescriptors of the congestion and thus of the quality of serviceoffered by such an organization. The aim throughout is to embodythe results in a form readily suitable for numericalevaluation.
Year
DOI
Venue
1984
10.1145/828.2513
J. ACM
Keywords
DocType
Volume
congestion,batch arrival,constant service,queuing system,waiting times
Journal
31
Issue
ISSN
Citations 
3
0004-5411
1
PageRank 
References 
Authors
0.41
9
1
Name
Order
Citations
PageRank
Micha Hofri1342127.96