A change management process is a critical component for data storage systems, especially in the areas of disaster recovery planning (to ensure that local changes replicate to a remote site) and storage networking (to prevent problems between interdependent software).
Tools can help, but there are only a few capable of complete end-to-end mapping from the application to the storage environment, according to Russ Fellows, a senior partner at Evaluator Group in Broomfield, Colo.
In this podcast, Fellows offers advice on effective change management for data storage systems.
You can either read the podcast transcript below or download the MP3 file to listen to at your leisure.
Download for later:
Change management process
• Internet Explorer: Right Click > Save Target As
• Firefox: Right Click > Save Link As
SearchStorage.com: In what areas of storage is it most important to have a formal change management process in place?
Fellows: The most important area is around business continuity and, of course, disaster recovery is related to that. Another is storage network change management. Basically, any area where a lot of interactions can occur between vendor products is where a lot of problems arise, and the storage network is certainly one of those. So [it's important to] understand all the dependencies of applications, and the dependencies between each application, down to the operating system, device drivers, I/O interfaces, and down to the network and the storage.
SearchStorage.com: What are the most critical best practices that would apply generally to change management with storage systems?
Fellows: Change management practices apply equally across an IT shop regardless of the underlying technology. It's something that's important to have if you have specialties -- the operating system admins, system administrators, network administrators and storage administrators in good communication and coordinating with each other.
Many people became enamored with the ideas of ITIL [Information Technology Infrastructure Library] for change management, but more recently, companies have started to back off that because of the rigorous processes they require. They found that in some cases those processes can be expensive and time-consuming, and don't really add value to the overall availability of the systems. So, just like we saw with ISO in the 1990s, ITIL isn't necessarily a silver bullet and the answer to all change management problems.
SearchStorage.com: What types of tools can help storage administrators with change management?
Fellows: In our discussions with end users, we find a lot of different tools in use from different vendors. A lot of them look at solving a particular area like storage networks or application to networks -- one area of the problem of the entire stack. There are very few products out there that can do a complete mapping, end to end, from an application environment all the way down to the storage. Unfortunately, problems between versions aren't usually well understood before changes can be implemented. So changes are done locally, but OS administrators [and] network administrators need to have the awareness of the potential impact. It's really an application-centric viewpoint that's the most important and, as I mentioned, there aren't too many tools that do that end to end, so the administrators will have to fill in the gaps; in many cases, they still rely on [an] Excel [spreadsheet] listing of all the assets they have.
SearchStorage.com: In what ways has server virtualization had an impact on storage from a change management perspective?
Fellows: It sort of has a two-fold effect. In some ways, it's made change management a little bit more difficult. It's introduced one more layer in the overall stack of things that you have to manage and maintain. That's a very critical part of the application environment, and that in turn has [an impact] on I/O interfaces all the way down to the storage.
But, in other ways, it's made change management better and easier because now it's much easier to have multiple versions of application environments, also known as virtual machines [VMs]. That includes a specific OS and device driver, Java virtual machines and everything needed for an application all maintained together. Now people can rapidly clone those virtual machines, and applications can be moved from a VM that might have a problem to another VM. So, in that regard, server virtualization in general has made changes easier and made it easier to recover from problems.
SearchStorage.com: As you mentioned, change management processes are especially critical in disaster recovery plans. What tips do you have for IT shops with respect to disaster recovery?
Fellows: Ensuring that you have processes in place to manage and maintain parity between different sites is definitely one of the most important for business continuity. Corporations … have disaster recovery sites, and some smaller companies are now using hosted DR sites. They might not have the need for an entire site, but they have hosted remote sites. And, in all those cases, you have to have very rigorous processes to make sure that any local changes or upgrades that you do are replicated to the remote site.
[It's] common to make a big change locally and replicate that to the other site. But then you do minor tweaks locally, and for whatever reason, those don't ever get replicated to the other site. Then, when the time comes and you need to have a working environment, the configurations aren't appropriate and you'll have problems. So, disaster recovery and business continuity is a critical area for change management.