In-Line Data Compression with Unified Storage Systems

In Line Data Compression: Ranges from 1.2 to 3.8x compression and skips dense blocks to minimize IO load To optimize the use of valuable data storage resources, data must be compressed when written to an array and then uncompressed when read. IceWEB’s Unified Data Storage Systems are designed to rapidly process these read-write compressions and decompressions by incorporating high-speed processors and an innovative cache design to ensure that using compression does not add significant overhead to I/O operations.

IceWEB IceSTORM OS

In addition, for blocks of data that are not very compressible — that is, the compression algorithm does not result in significant space savings — the system automatically skips compression for that data stream. This unique capability maximizes the efficiency of I/O throughput by simply avoiding attempts to compress and decompress dense blocks of data with minimal compression ability.

For blocks of data that do compress, we can eliminate I/Os, thus not only saving you capacity, but also reducing the overall I/O load on the system resulting in an overall performance increase for your array. Given all of this, there really is no reason not to be using compression for all your data; there is minimal downside, and a very significant upside with most applications.

Data Compression Ranges

What kind of compression ratio can you expect? First of all, a compression ratio is reported on files and file systems within the array. This ratio is simply the uncompressed size divided by compressed size. The following table shows the compression ratios you could expect with some common types of data:

File shares containing Office documents	1.2 to 1.6
File shares containing virtual machine images	1.5 to 3.0
Messaging databases	1.2 to 1.8
Structured databases	1.5 to 2.1
Uncompressed images	1.9 to 3.5
Uncompressed videos	2.2 to 3.8
HTML files	1.6 to 2.0
Executables	2.0 to 2.4