Disk Array Implementation Tips - EnterpriseStorageForum.com

Disk Array Implementation Tips

We have recently been posting buying guides concerned with the latest disk arrays. Let’s follow those up with a story on disk array implementation. In some cases, old arrays are just fine and will last you long into the future. But in others, they can become a liability.

So which should you retain and which should you dump? If you listen to the vendors, all you tend to hear is “Buy New.” But that may not always be the smartest course.

Here, then, are some tips to help you decide.

Skip All Flash

While there are plenty of good all-flash arrays around and some applications and organizations need them, they are overkill in many cases. Dell is one of the all-flash array vendors, but it is realistic enough to note that 97% of its customers using flash also use lower cost HDDs in a hybrid array configuration.

“The majority of organizations have needs for supporting both high performance applications as well as less accessed data that doesn’t require expensive storage,” said Travis Vigil, executive director, product management, Dell Storage.

His advice to those adding some flash into the enterprise storage arena is to avoid having separate arrays for each – one for flash and several for disk, for example. Vigil said that storage deployments benefit the most from a single SAN that handles both ends of the spectrum at the same time.

“Hybrid flash storage with auto tiering continues to be the best way to optimize cost for performance as it combines the faster flash with the slower yet lower cost spinning disks,” said Vigil. 

Retain Slower and Bigger Disks

Vigil also has advice concerning how to decide which disk to buy in the future. He reckons 15k and even 10k disk drives are gradually becoming obsolete. Why? Their higher price point for only marginally better performance isn’t worth it any more as flash is gobbling up the higher performance market. Where flash can’t compete, though, is in quantity.

“As the cost of flash comes down, look for 15K and 10K drives to be replaced with SSDs while most customers will gravitate towards 7.2K HDDs for their colder data,” said Vigil.


Bob Madaio, Senior Director, Product marketing, Hitachi Data Systems, commented that older systems can sometimes bog companies down with disparate units that are hard to manage. His suggestions is to add a virtualization layer to sit on top of your storage hardware. That way, the old can coexist with the new and you can get the maximum lifespan out of them. 

“If the systems you currently have deployed are limiting your flexibility, consider adding a level of storage virtualization/abstraction,” said Madaio. “Doing this, you can keep and repurpose your existing assets, add new assets to support new applications, or move information between systems in a much easier fashion.”

Read the Manual…

An unfortunate fact of our high-tech world is that whether it is a cell phone, a laptop or a high-end storage array, few users ever find out about 90% of the features. So for all these devices, it pays to take the time to read the manual. Perhaps most of the fancy bells and whistles are irrelevant. But a handful will likely save you time and money if you take the time to find out about them.

“Many customers don¹t realize that they¹re not fully leveraging all of the features available in their arrays, including some that can help extend the life of an array,” said Jon Siegal ­ Vice President of Product Marketing, Core Technologies Division, EMC.

He gave the example of an EMC VNX. He indicated that there are many ways to leverage this technology to continue to gain value from older storage assets, such as using VMX for file server consolidation. In addition, you can increase the flash amount on VMX for increased performance and scalability. Or you can introduce EMC ViPR to the account to leverage older arrays to create a pool of storage capacity, provide object support, lower operational management costs, add additional data services and extend the life of the VNX arrays, added Siegal.

Match the Array to the Application

Some say have more disk, some say have more flash. But who really knows? The user who has a specific application to run. You can avoid problems and surprises by knowing your application performance, availability, capacity and economic needs.

Armed with that data, the user can better evaluate the capabilities of different storage systems. That enables you to look beyond the basic speeds and feeds, get clarification from the vendor of what they actually support along with up-front and on-going costs.

“Look for solutions that work for you and your environment while supporting your application needs,” said Greg Schulz, an analyst with StorageIO Group.

A Little Goes a Long Way

In Schulz’ view, flash is something that works well – but in moderation. He advises users to use it strategically where it can have the biggest impact. 

“A little bit of flash SSD in the right location can go a long way being effective, but on the other hand, a lot of flash will cost you a lot of cash,” he said. “Disk drives are best for storing large amounts of data cost effectively and for backing up or protecting data on SSDs.”

Photo courtesy of Shutterstock.

Comment and Contribute


(Maximum characters: 1200). You have characters left.



Storage Daily
Don't miss an article. Subscribe to our newsletter below.

Thanks for your registration, follow us on our social networks to keep up-to-date