Introduction to the improvement of data compression in new version of Vnx (II.)

The new version of data compression mainly has the following improvements, this article will detail the display capacity and compression of the improvement of rice.

• Conversion of thick LUN to thin LUN when block compression is enabled
• Shows the capacity saved by compression
• Deeper-level file compression algorithms

Solution:

1. New version of the introduction of the show because of compression and save capacity

This feature shows the capacity saved by compression. Unlike the capacity savings that thin resource allocation (Thin provisioning) can bring, this can inform users of the savings resulting from compression.

The new features show that the saving capacity does not include the capacity saved by the 0 detection, the following example will help us to understand further.

A 2TB raid Group lun or thick lun, when compression is enabled, the 0 detection mechanism removes all free space, and the LUN first converts to a 150GB thin LUN. When the compression is complete, the final thin LUN is compressed into 100GB. At this time because of compression and save space is only 50GB, most of the space saved by 0 detection mechanism at work.

Benefits of new features

Users can determine the increasing benefits of using compression. This is necessary when using compression can have an impact on performance, and users need enough information to make a cost-benefit analysis.

View new features with Unisphere

Viewing the compression space of a single LUN in the LUN properties interface

View the storage Pool property interface for compression savings for the entire storage pool

To view new features using the command line

2. New version of the introduction of depth compression algorithm

• The new algorithm can save up to 30% of capacity compared to the fast (fast) compression algorithm
• For archived files, saving capacity is more important than access time
• Requires more CPU resources and a longer initial compression time
  

Depth algorithm and fast algorithm comparison

Algorithm name	Compress time	Decompression time	Size scale of files and source files after compression	Configuration	Start supported version of File
Deep Deep	9 seconds.	2.11 seconds.	35%	Manual configuration	V7.1
Quick Fast	4 seconds.	1.28 seconds.	50%	Default	V5.6

As the above table data is obtained by testing the data set of 414MB, rapid fast algorithm can compress it to the original 50% in 4 seconds, and the depth deep algorithm can compress to 35%, just need one more times.

View New algorithm with Unisphere

Enable deep compression on the file system. Property interface, click to Reset, and then select the compression method