Four ways to reduce cost of storage ownership

Storage is expensive to acquire and operate, but several techniques offer you ways to keep those costs under control.

Storage is expensive stuff. You can expect to pay big dollars for arrays and then surprising amounts for the disks they house. Vendor maintenance and licensing fees are inescapable, as are rises in the price of the electricity you need to operate and cool your storage equipment.

Throw in the fact that Australia has a shortage of skilled and experienced storage experts, with the few in the market able to command large salaries, and storage can quickly accrue substantial ownership costs.

Happily, there are a few ways you can control those costs and reduce the total cost of ownership (TCO) for storage. Here are four suggestions you may find useful

1. Achieve high utilisation rates

Many organisations acquire storage arrays and then don’t use all of the disk capacity they contain. This means they overinvest when they purchase storage, a wasteful use of capital. It can also mean that they invest in so much capacity that they are committed to using an old array even after more advanced systems reach the market, leaving them unable to afford products that could enable innovation.

There are a few ways to improve utilisation, namely:

•         Thin provisioning – When storage arrays are first configured, they are “provisioned”, a term that describes reserving some of their capacity so it can be allocated to a particular application or system. Provisioning often sees IT staff apply a mixture of optimistic and defensive thinking that sees them reserve a lot of storage capacity in order to ensure they don’t run out of capacity in the future. That capacity is locked away from other users.
In practice, this often means some storage capacity is orphaned, hidden and eventually becomes unused. Thin provisioning allows administrators allocate the capacity users ask for, but until they actually use it the disk space is available for other users, which stops the problem of under-utilised storage hidden away by past provisioning requests.

•         Virtualisation –Storage has nearly always been considered as a physical resource, so the disks inside an array have been bound to that array. Virtualisation allows you to consider all the capacity in all your arrays as a single pool of storage. By doing so, virtualisation means you can identify unused storage capacity on arrays around your organisation and put them to work. This will often mean using underutilised capacity, which will boost utilisation rates.

•         Careful procurement – Storage buyers sometimes acquire more storage than they need, either through optimism or because they are offered a good deal on lots of capacity. More careful procurement will see you acquire less storage capacity, but use more of it.

2. Reduce the amount of data you store

Filing cabinets are simple and wonderful tools - a few folders and some sliding drawers let you arrange and preserve documents that, if left in an unruly pile, would occupy lots of space and often spill over to make a mess. Used well, you get more files in less space.

You cannot just file away your digital data in neat packages. But you can arrange it more neatly so it occupies less disk space. And if you need less disk space, you can operate fewer disks, defer purchases and reduce cost.

Two of the techniques you can use to store more data in less space are:

•         Data compression – Many file formats aren’t very efficiently designed, making it possible to compress them so they occupy less disk space. Most storage management software can perform compressions. Expect compression ratios of at least 2:1

•         Data deduplication – In a modern business, lots of data is replicated. Some of the copies you’ll have come about because one department wants its own version of data to work with, so makes a copy. Lots of data is also replicated because you do the same things a lot: every letter your organisation has ever sent has the same details on the letterhead. Data deduplication looks for both kinds of duplicate data, stores one version instead of lots, and then automagically rebuilds documents so that when you open them the duplicated data re-appears, ready for use.
Data deduplication is nearly always more powerful than compression and can help you to store a lot more data on the arrays you already have.

Compression and deduplication reduce TCO by making it possible to buy and operate less disk.

3. HSM and tiering

Most organisations have some very important data, and plenty of less important material.

E-mail from the last week, for example, is probably being accessed quite frequently. But email from three years ago is hardly ever read or accessed.

Most organisations know about these access patterns, but many simply leave all their email on one storage appliance. That kind of behaviour is a missed opportunity because different types of storage array and different types of disk have different performance characteristics and therefore cost different amounts. Storing last year’s email on this year’s fast-and-expensive solid state disk is therefore a very poor idea!

Two approaches address this problem.

One is hierarchical storage management (HSM), a technique that sees old and infrequently-accessed data migrated to cheap-to-operate storage devices. HSM could mean a move to tape or to an old storage array. In either case, it leaves your fast, new, highly-capable storage kit to operate with data that needs its speed, thereby making better use of your investment and controlling costs.

Tiering is a very similar idea, but sees data that applications need to access a lot placed on fast disks and in fast arrays. Many high-volume websites place frequently-accessed data in a tier of very fast disk, less popular data on fast disk and their least popular data on old, slow cheap disk.

Tiering differs from HSM in its dynamism – it often analyses data on the fly and shuffles data between tiers often and automatically – and in its use to assist applications perform better. HSM, by contrast, is more often applied to older data that is not used by live applications.

4. Simplify storage management

Many organisations acquire different storage devices over time. As most of these devices include management software, organisations find they need multiple tools to manage storage.

The complexity and cost involved is a significant contributor to storage TCO.

Some vendors now offer software that can manage arrays from their rivals. Third party storage management tools do likewise.

In both cases, organisations can expect to reduce TCO by removing the need for multiple pieces of management software and people with the skills to operate them all.

 

This was first published in August 2011

Dig deeper on Data management processes

-ADS BY GOOGLE

SearchCIO.com.au

SearchSecurity.com.au

SearchDataBackup

SearchDisasterRecovery

SearchSMBStorage

SearchStorageChannel

Close