AMASS Overview

DesignPreventsThrashingCacheOptimizesRequests

In a storage environment, there are many volumes but only a few drives. If several requests come in for many different volumes, the potential exists for AMASS to spend most of its time moving media in and out of drives and little of its time actually processing requests.

The following items help AMASS handle many simultaneous requests, thus preventing thrashing as well as optimizing requests:

Request queue sorting

Read-ahead

Write-behind

Prioritizing algorithm

The AMASS cache resides on a hard disk attached to the UNIX server where AMASS is installed. The cache implementation follows all UNIX file system conventions for synchronous IO, sync, and fsync functions.

Data caching provides the following benefits:

Faster system performance

Protection against thrashing

In addition, a large cache allows large files to be queued faster before being moved to a library thus increasing throughput.

After files are in the cache, multiple writes to the same volumes are grouped into single operation that minimizes volume movement and maximizes throughput. Therefore, a high aggregate throughput is achieved through the following items:

Grouping write operations in the cache

6-00026-01 Rev A

Online Archiving with AMASS

1-9