Between current movies and television programs, zombies are everywhere. AMC’s “The Walking Dead” drew an astounding 15.8 million viewers for its Season 5 finale last March. The show has even spawned its own spinoff, “Fear the Walking Dead.”
The entertainment market isn’t the only industry seeing more zombies. IT teams are increasingly discovering these fearsome creatures within their own network infrastructures. These aren’t drooling, undead ghouls; these are underutilized servers siphoning off essential resources, taking a toll on the company’s computing power.
“If 10 percent of the servers within a 1000-server infrastructure are zombies, roughly $23,000 is spent a year keeping these ghost machines up and running.”
Zombie Servers are hiding in the Server Room
Zombie servers have become an increasingly prevalent issue, not only in enterprise infrastructures but also within some data centers as well. An estimate presented by The Wall Street Journal’s Robert McMillan pegged the number of zombie servers in the U.S. at more than 3 million. Together, these use more than 1 gigawatt of power, or roughly the same amount of power consumed by all the households in the city of Chicago.
“They’re servers … sucking up lots of power while doing nothing,” McMillan explained. “Most companies are far better at getting servers up and running than they are at figuring out when to pull the plug.”
This is more common than ever today. When enterprises deploy an application or integrate a new solution, they typically set up a server (be it physical or virtual) for the necessary support. However, when this system is migrated to the cloud, replaced or otherwise upgraded, some IT teams may overlook the original server. With the application the server was supporting now gone, these machines are needlessly consuming power.
Costs add up paying for that wasted power. Raritan pointed out that if 10 percent of the servers within a 1,000-server infrastructure are zombies, this means the organization supporting the network spends roughly $23,000 a year keeping these ghost machines up and running.
Addressing the zombie server problem
So what can an enterprise do to solve its zombie server problems? The first step is identifying the servers that are underutilized. This can be a laborious process, as it could mean tracing files and application activities back to specific database locations to ascertain which servers are functioning and which are simply sucking up resources.
“Things that should be turned off over time are not and unfortunately the longer they linger, the worse the problem becomes,” noted Paul Nally, Bruscar Technologies LLC principal. “They would have never been found by any other methodology other than walking around with a clipboard.”
Nally told McMillan that he once audited a data center with more than 1,000 servers, none of which had the necessary domain-name-system software configuration necessary to identify them. As a result, he and his team had to physically go to each server to determine which were zombies – a nightmarish task.
How managed hosting can help eliminate zombie servers
Once zombie servers are identified, the team can work to migrate what few application processes or activities these machines might still support. One of the most beneficial ways to address zombie servers is to leverage managed hosting services to ensure that available power and other resources are not wasted.
Within a managed hosting environment, the vendor’s on-site team maintains the company’s off-premises machines, taking responsibility for tasks including patching, monitoring and backups. In this way, the company can be sure that its critical applications and online activities are supported by an expertly maintained server environment. The service provider’s team ensures that all available computing resources are used appropriately, eliminating the chances of any zombie servers within the managed infrastructure.