“Legacy” is an interesting word. Aided by Gartner, Microsoft managed to convince the world in the mid-nineties that anything that wasn’t a Windows-based server was old and hence “legacy.” Who have thought just a few years ago that anything not flash would earn the very same moniker? Yet here we are.
All-flash arrays and hybrid storage arrays have established themselves as the standard. Storage areas networks (SANs) and network attached storage (NAS) using hard disk drives (HDDs), so long a staple of data storage, are now being sidelined as legacy.
So what should you be doing with all those antiques that are littering the data center?
It’s unlikely that they will earn a classic status like some American automobiles from the 50s and 60s. But that doesn’t mean they have no value. How should you harness existing assets as all-flash takes hold?
Here is what the experts say you should do with exist non-flash storage assets.
Hybrid Transition
One approach is to take advantage of existing storage hardware and retrofit in as many solid state drives (SSDs) as is feasible. Many can be changed out directly with HDDs. They slide right into the same slots. This creates a hybrid storage array composed partially of flash and partially of HDDs. That path would certainly be cheaper than buying brand-new, latest-generation all-flash arrays.
“Customers will need arrays from vendors that offer both flash and hybrid, depending on the workload and economic needs,” said Bob Fine, product marketing, midrange storage at Dell EMC. “With this need for different flavors of storage, customers need a single vendor that can deliver both, and use both arrays with the lowest operational costs including daily management, service and support.”
Greg Schulz, an analyst at StorageIO Group agrees that there is still a place for hybrid arrays.
“A hybrid array is the home run when it comes to leveraging your existing non-flash, non-SSD based assets today,” he said.
Storage Tiering
Storage tiering is another area where older arrays may come into their own. HDD-based storage can serve as the repository for lower tiers of storage. Or it can be partially converted to flash to form a hybrid array. Some hybrid arrays can be set up to automatically determine the age and frequency of actual data usage. Many hybrid arrays can use both flash drives and spinning disk drives together, in unison. This gives the user the option to dial in the amount of flash they need.
“They should choose a hybrid array than can automatically place data where it belongs to meet actual data use patterns,” said Fine. “Then you achieve the performance benefit of flash with the lower cost density of spinning media.”
Archival Storage
Brett Schechter, senior product marketing manager at Tintri, concurs. In fact, he’s seen many organizations do just that. Some are going a stage further and using legacy storage arrays as an archive.
“Many of our customers redeploy their legacy arrays and spinning drive assets as an archival tier,” said Schechter. “Much of the unstructured data that might be useful for data mining later, or even in legal proceedings, need not be on fast tier all-flash arrays. In this fashion, they maintain secure control of their data, and still have reasonably high performance if forced to access it.”
Wait for Technology Refreshes
Certainly, flash storage offers clear benefits. Thus, vendors are urging users to make the switch immediately. But unless there is a clear cut reason for doing so right now, it might be best to wait for the next technology refresh and make any changes then. Hardware will have to be changed out at that point in any case, so why rush into it any earlier without good reason? So keep the old stuff for another couple of years, if possible, to extract as much value as you can.
“As legacy storage comes up for technology refresh and enterprises move workloads off of them, we think they should always be considering all flash arrays for primary workloads and should buy that unless there is some specific and very good reason why they would choose not to do that,” said IDC analyst Eric Burgener. “With the release of big data flash arrays in 2016, we will see these systems competing more and more for secondary storage applications that have some performance sensitivity to them (not so much low latency but maybe high throughput and/or bandwidth requirements for content media serving, big data analytics, etc.).”
Giving Old Arrays a Boost
Aging arrays may be sluggish compared to their all-flash counterparts. But a little bit of flash can give them quite a boost if deployed smartly.
“A little flash or SSD storage in the right place as a cache or buffer can save you a lot of cash,” said Schulz. “Make your current assets more productive by adding a cache, buffering, tiering or micro-tiering solution in front of or on top of an old array to make them faster, sometimes even appearing like new.”
For example, you can use tiering and cache in operating systems such as Microsoft Windows Server 2016, VMware vSAN and vSphere, along with third-party micro-tiering tools like Enmotus among others. Likewise, you can leverage cache and tiering found in many storage systems and appliances, as well as software-defined storage tools, said Schulz.
Time to Die
Just like the famous line from the end of the movie “Blade Runner,” there comes a time for hardware when it is time to die. When that day comes, be sure to thoroughly erase all data using enterprise-class tools.
“When it comes time to dispose of your old HDDs as well as older early generation flash and SSDs, make sure to securely digitally shred using tools that do a deep cleaning (e.g. digital bleach) of those hard to reach persistent memory cells,” said Schulz. “Some tools such as those from Blancco, among others, in addition to doing the deep digital cleaning of flash SSD can also produce a report to use as a certificate of compliance.”
Photo courtesy of Shutterstock.