Given the ongoing infatuation with zombies it's a fair bet you'll see plenty of them this Halloween. Better also watch out for zombies in your enterprise.
I can only assume that online medical journals are out of date since I could not find a scientific definition of "zombie" anywhere, so let me offer up the following: a zombie is something that is dead but doesn't know it yet. An attribute of zombies is that while they were once value-creating entities, they are now consuming resources without significant added value. Brain-eating is optional.
Where are the zombies in your enterprise? How about virtual machines that were quickly and easily spun up but are no longer needed or used? SharePoint sites? File shares, intranet pages, user desktop software? User accounts? How about entire business or IT services?
Virtualization sprawl is perhaps the most visible symptom that, like easy credit, easy provisioning comes at a price. Automation has made us increasingly efficient at spinning up resources like VMs, but we typically lack a process to spin them down, or for deciding which resources should be spun up in the first place. This open-loop approach to resource provisioning results in an accumulating operational cost burden including poor use of software licenses, disk space, and staff time to maintaining these unproductive resources. And in the case of unterminated user accounts, an accumulating security risk.
With apologies to George Romero and Woody Harrelson, we can kill or prevent zombie resources with a closed-loop approach to provisioning using a Service Catalog. By managing virtualization, Sharepoint, file shares, user accounts etc. as services in a Service Catalog we create opportunities for zombie control through policy enforcement and cost transparency. The number of people who are authorized to requests resources can be narrowed through role-based permissions. Some services can be offered as monthly, quarterly or annual subscriptions; at the end of the subscription period the requestor could be given the option to renew the subscription. A lack of affirmative response could result in cancelling the subscription. The Service Catalog can then initiate workflow to automatically terminate the resource. Datacenter automation and discovery, meanwhile, can "sweep up" by detecting unauthorized resources like rogue VMs.
From the outset, the Service Catalog should expose a cost for the subscription. Whether or not formal chargebacks are used, the requestor (and his or her management) would see a report for the ongoing cost for subscriptions. With up-front transparency in the catalog itself, some judicious users may refrain from requesting resources in the first place if the value doesn't justify the cost.
In the cases above we haven't eliminated any services, we're just eliminating zombie resources by controlling who gets access to them and for how long. But how do you know if there isn't dead weight in the service portfolio itself? By following a Service Portfolio Management strategy you can gather information on service consumption, cost, quality and business value throughout the service lifecycle, then periodically sit down with your business customers to analyze the portfolio as a whole to see whether some services have outlived their value and should be investigated for retirement or outsourcing. A regular pruning of zombie services from your service portfolio can free up resources for more valuable, non-brain-eating services that have more potential to drive business value.
Some might say that offering such infrastructure resources as an internal service is part of the internal cloud phenomenon. I guess that depends on one's definition of "cloud", but since this article has already reached its buzzword quota with portfolio, catalog, virtualization and zombie, I'll leave that conclusion to you, gentle reader.