Four Ways To Save Storage Space

A large red file, and a smaller green file beside it

Posted on Tuesday, November 8, 2016

Historical attempts to manage data growth have failed, simply because they are at odds with modern storage provisioning techniques. But the CTO is still expected to control costs and prevent wasted spend wherever possible. Which is why simply storing everything is not a legitimate long-term approach to your corporate needs.

Instead you must manage what is retained and how with a proper data storage strategy. The following four tips will help you better manage the challenge you face.

1. Stop wasting space

Your existing data store is bound to contain hundreds of thousands of duplicate files and records that consume valuable capacity. These duplicates may also be present in your backups and archives, taking up valuable space there too.

De-duplication can be costly and time-consuming, but the dividends are significant. By using hardware and/or software de-duplication solutions you can automate much of the process and prevent creation of copied data in future too.

Managing what is stored will free up space in both active and archive storage space, providing some additional “breathing room” to plan and deploy the next capacity increase.

2. Stop thinking in terms of thin provisioning

Traditionally, the CTO was encouraged to buy capacity upgrades at the last possible moment – a strategy known as thin provisioning. And although this approach does help to alleviate the cost of additional disk arrays, it is not the most efficient for your storage growth needs.

The problem is that thin provisioning relies on being able to accurately predict data growth – something no one has ever managed successfully yet.

Thin provisioning is also at odds with software defined storage, which creates and uses a large, shared pool of storage that can be allocated and deployed automatically using policy-based provisioning.

This may mean purchasing purposeful storage to meet business requirements such as high performance flash for analytics where storage can be deployed– or redeployment of existing post-warranty assets for long-term retention. Reduced management overheads and increased infrastructure flexibility will help to cancel out those additional costs however.

3. Stop using short term fixes

The de-duplication proposal above is actually just a short-term fix. Removing duplicates will free up capacity, but those savings will be quickly consumed by legitimate data growth.

You could of course invest in compression technologies to further reduce the footprint of your data, but performance issues aside, this too is a short term fix. Yes, you resolve capacity issues in the short term – but all you have really achieved is a deferment of the wider problem.

A better alternative is to create a tiered approach and to align the value of information with cost of retention. The information requiring the least access performance should be moved to the least expensive storage based on demand policies. This elasticity creates the greatest flexibility because typically 80% of data is considered stale and not required for daily operation.

Capacity demands are never going to reduce, so you must become more strategic in your thinking.

4. Stop ignoring data lifecycle management

Not all of the data your business stores needs to be online all the time. In fact, much of the information held for Big Data purposes is stale, and maybe even useless.

Revisit your data lifecycle management (DLM) strategy to create new rules about validity, and what happens to data that no longer serves a purpose. You can build a tiered approach that aligns the value of information with cost of retention.

You can then automate these rules, shifting data to low-cost archive systems whenever necessary. You can then be sure that all spare capacity is available at all times.

The information requiring the least access performance should be moved to the least expensive storage based on demand policies. This elasticity creates the greatest flexibility because typically 80% of data is considered stale and not required for daily operation.

You can then automate these rules, shifting data to low-cost archive systems whenever necessary. A proper data lifecycle management plan should ensure that all spare capacity is available at all times.

Big Data does not mean all data

Data growth is a fact of life for the modern enterprise. But with a little strategic planning and investment, you can help to increase availability and performance, whilst containing costs.

For more help and advice, including how to repurpose your post warranty assets to deliver increased capacity at no additional cost, please get in touch.