What Is Data Reduction?

What Is Data Reduction?

Data reduction is a capacity optimization technique in which data is reduced to its simplest possible form to free up capacity on a storage device. There are many ways to reduce data, but the idea is very simple—squeeze as much data into physical storage as possible to maximize capacity. 

In this article, we’ll dive into the basics of data reduction to help you better evaluate storage vendors.

Benefits of Data Reduction

The main benefit of data reduction is pretty straightforward: The more data you can fit into a terabyte of disk space, the less capacity you’ll need to purchase. Data reduction can:

• Save energy
• Reduce your physical storage costs
• Decrease your data center footprint

Data reduction greatly increases the efficiency of a storage system and directly impacts your total spend on capacity.

All-Flash Arrays: Bringing the Benefits of Flash Memory to the Data Center

As you might guess, simply switching out your HDDs with SSDs is enough to increase the speed and performance of your NAS and SAN solutions. The benefits of an all-flash array are the same as the benefits of flash memory itself:

• Speed: Faster memory read-write and access times lead to improved speed and performance. The best all-flash arrays leverage NVMe over Fabrics (NVMe-oF) to maximize data transfer speeds and latencies throughout a SAN.
• Portability: SSDs are significantly smaller than HDDs. On a purely physical basis, flash memory has the advantage of space-per-capacity. On a cost-per-capacity basis, flash memory is quickly closing ground on HDD solutions.
• Durability: The lack of physical moving parts makes SSDs inherently less vulnerable to drops and shocks than their spinning-disk counterparts.

Data Compression vs. Data Deduplication

Data-reduction techniques can be broadly categorized into two main types:

• Data compression: This bit-rate reduction technique involves encoding information using fewer bits of data. Compression algorithms can be lossy (some information is lost, reducing the resolution of the data) and lossless (information is fully preserved by removing statistical redundancy).
• Data deduplication: Also known as dedupe, this process involves eliminating duplicate copies of data within a storage volume or across the entire storage system (cross-volume dedupe). It uses pattern recognition to identify redundant data and replace them with references to a single saved copy. 

In practice, you can employ a combination of techniques from both categories to reduce data in your system.

How Pure Storage Delivers on Data Reduction

Pure Storage® Purity Reduce uses five different data-reduction technologies to save space in its all-flash arrays:

• Pattern removal: Purity Reduce identifies and removes repetitive binary patterns to reduce the volume of data to be processed by the dedupe scanner and compression engine. 
• 512B aligned variable dedupe: A high-performance inline deduplication process with a variable block-size range of 4-32KB ensures only unique blocks of data are saved on flash.
• Inline compression: Purity Reduce uses an append-only write layout and variable addressing to remove the wasted space fixed-block architectures introduce.
• Deep reduction: Inline compression is followed by heavier-weight compression algorithms post-process to further increase space savings. 
• Copy reduction: Copies made on FlashArray™ only use metadata—Purity provides instant pre-deduplicated copies of data for xCopy commands, snapshots, replication, and clones.

Purity Reduce delivers the most granular and complete data reduction ratios in the flash storage industry:

• 4-6 to 1 for virtual server environments such as VMware or Hyper-V
• 2-4 to 1 for database environments including OLTP or OLAP
• 5-10 to 1 for virtual desktop environments (VDI) both persistent and non-persistent

Data reduction works on a wide variety of applications and data types, but the only way to know how it functions on your data is to try it.