Virtualization. Cloud computing. Storage area networks. These technologies represent the cutting edge of IT, enabling...
system administrators to consolidate server racks and operate far more efficiently. Hardware, space and power costs plummet, leaving organizations with much-needed breathing room in their budgets.
But what happens to all of the leftover servers, switches and other hardware when a data center consolidates? Sadly, there's no server fairy that can swoop in and make all of this iron disappear in a puff of smoke. In the real world, it's important to plan for the hardware that is left behind when consolidation or simple age renders devices obsolete.
Repurposing hardware is not something that's given a lot of thought. Yet according to Dan Kusnetsky, VP of Research Operations at The 451 Group, planning is exactly what is needed for repurposing, as early in the process as feasible. Kusnetsky, an expert on the benefits of consolidation and virtualization, has kept an eye on this decidedly overlooked part of the process: what to do with all of those orphaned boxes after all is said and done. In a recent interview, he explained what many companies are doing with their orphan machines.
Pass used hardware down the chain of business priorities
First, the older hardware may not go anywhere. It's common for mission-critical VMs to be migrated to the newest hardware, where it benefits from better power consumption and superior resiliency features. Administrators can then migrate less important VMs to the older hardware. Similarly, the older hardware can continue on as backup/recovery servers, or "spare" systems that stand ready for use in an unexpected system outage. Since most good consolidation plans already make allowances for potential outages and legacy devices still take up floor space, power, and cooling, Kusnetsky downplayed this as a popular option.
Servers that no longer have a place in the main data center may still serve other departments within the organization. For example, send the devices to development and testing teams. Historically, these departments might be on the short end of the new hardware stick anyway, so if the hardware is more powerful than what the team had before, they might find themselves the recipients of the repurposed machines. This usually means the developers' original servers are distributed outside the company (as you'll see later on).
Another way that Kusnetsky sees companies re-tasking orphaned servers is by "giving them to business units or workgroups that had been managed remotely." After reimaging the devices, the once-remote locations will get a local server, which usually brings increased performance and security. However, this tactic may not be appropriate in every situation -- especially with organizations that are trying to centralize computing resources for their remote locations. It's easy to see that putting hardware back at a remote location will impose maintenance and management overhead that some businesses are trying to eliminate.
Removing older hardware from the company
Still, there are many instances where it just doesn't make sense to keep an older piece of hardware. Age, expensive maintenance contracts, a shortage of cost-effective replacement parts or lack of need may necessitate removing servers from the company entirely. In that case, Kusnetsky explained, the units could be sold out in the secondary market or even donated to universities and charitable organizations as a tax write-off.
Selling off used hardware to the secondary market appears to be an increasingly popular option, especially as the economy continues to struggle back from recession. John Lynch, Director of Sales and Marketing at Asset Recovery Centers LLC, definitely agrees and points to a recent project involving the inventory and disbursement of about $1 billion worth of used client equipment. This is sometimes termed a "takeout" in secondary market lingo. Lynch explained that the $1 billion figure for this particular takeout represents the original cost of the equipment, not the current value, which has yet to be determined. Lynch said that the secondary market is seeing a definite increase in business -- and not just from current consolidation projects.
Lynch explained that during the bumpy ride the economy has been experiencing since 2006, many companies in the throes of reorganizing, consolidating and dying would often neglect to move their equipment to the secondary market. Instead, "they put it in the corner, or in a warehouse somewhere." Often this decision was made, he added, because companies in distress wouldn't have time to deal with equipment issues as personnel and business decisions commanded more attention.
Today, however, organizations aren't trying to save anymore. "It makes no sense to keep warehousing this hardware," Lynch said. This activity is filling the secondary market with current equipment as well as machines that could be more than five years old -- ancient by some IT standards. But, in the secondary market, "everything has value," Lynch emphasized. Even dated equipment can be sold as scrap and its core materials extracted and reused. Slightly newer equipment can be used as parts, and still-functioning devices can be distributed anywhere in the world.
Given the different choices to handle orphaned servers, which is the best solution for a business to use?
Use a variety of reallocation and disposition tactics for older hardware
Kusnetsky emphasized that he has seen no real preference for any of these solutions across the companies he has observed. "Each organization has its own unique situation," Kusnetsky said, which is why it's important that the business and technical aspects of the company are equally considered when deciding what to do with older and orphaned hardware. "What is the business situation of the company? Where will the hardware fit? Can it fit?"
The variety of scenarios a company might face is almost endless. For instance, Kusnetsky outlined an environmental situation that's currently very prevalent in data centers located in high-density areas within California or New York. There, the very limits of power production will force companies to take a hard look at their power consumption before and after hardware is repurposed. If re-using older hardware might mean using more energy, then such firms would have little choice but to remove the hardware in favor of newer and more power-efficient models.
This complexity means that plans must be made that encompass the needs of IT and the general welfare of the business. Older hardware can find a new home in your company, but that home must be well-prepared.
Survey your company to be sure there is no business unit that is lacking for more powerful hardware. That includes remote units as well as development and testing teams. Also, consider moving low-use apps to the older hardware, but make sure you don't give your company a net increase in power or space consumption that negates the gains of the consolidation.
If you can't find a new home for the hardware, the secondary market is full of vendors that can assist you in getting value from your hardware. If you're looking to reduce your tax bill, donations to worthy organizations are an excellent way to go. Smaller data centers might even consider a barter arrangement with partners and vendors to obtain needed products or services that might otherwise be beyond the reach of their current budget.
What did you think of this feature? Write to SearchDataCenter.com's Matt Stansberry about your data center concerns at firstname.lastname@example.org.