Essary, David and Amer, Ahmed
(2009)
Space-Efficient Predictive Block Management.
Technical Report.
Department of Computer Science, University of Pittsburgh, Pittsburgh.
(Unpublished)
Abstract
With growing disk and storage capacities, the amount of required metadata for tracking all blocks in a system becomes a daunting task by itself. In previous work, we have demonstrated a system software effort in the area of predictive data grouping for reducing power and latency on hard disks. The structures used, very similar to prior efforts in prefetching and prefetch caching, track access successor information at the block level, keeping a fixed number of immediate successors per block. While providing powerful predictive expansion capabilities and being more space efficient in the amount of required metadata than many previous strategies, there remains a growing concern of how much data is actually required. In this paper, we present a novel method of storing equivalent information, SESH, a Space Efficient Storage of Heredity. This method utilizes the high amount of block-level predictability observed in a number of workload trace sets to reduce the overall metadata storage by up to 99% without any loss of information. As a result, we are able to provide a predictive tool that is adaptive, accurate, and robust in the face of workload noise, for a tiny fraction of the metadata cost previously anticipated; in some cases, reducing the required size from 12 gigabytes to less than 150 megabytes.
Share
| Citation/Export: |
|
| Social Networking: |
|
Details
Metrics
Monthly Views for the past 3 years
Plum Analytics
Actions (login required)
 |
View Item |
![[img]](http://d-scholarship.pitt.edu/style/images/fileicons/text_plain.png)
![[feed]](/images/feed-icon-32x32.png){kind=icon}