Once upon a time virtualization was just a concept on most organization’s roadmaps. Today it is a rarity that I talk to an organization that does not have a large percentage of its infrastructure virtualized. But this is creating new management challenges, one of which is keeping applications and their underlying virtualized infrastructure in proper alignment.
The advent of virtualization in organizations is putting their data centers and even their business processes in a huge state of flux. It used to be that as applications were purchased, the underlying resources that were needed to support them – servers, storage, backup, and to a certain degree even network – were procured at the same time and then largely dedicated to that application’s exclusive use regardless if it fully utilized them or not.
Virtualization breaks that entire model. Now backup, network, server and storage resources are purchased and deployed independently of applications with applications forced to share these common underlying resources. While that sounds fine in theory, organizations are finding the practical application of that a little more difficult than vendors might have them believe.
Organizations are in general running across this problem in two different ways. More mature organizations are encountering it as they converge IT operations and then look to automate them. It is as they evolve their infrastructure and look to automate the dynamic allocation of the right resources to the right applications that they discover it is still a manual process.
Alternatively there are those organizations that are taking a more revolutionary approach and going directly to this fully virtualized environment either because they are a relative new company or simply opting to outsource their entire data center to a third party. However these organizations are also not finding virtualization to be quite the panacea that they expected.
On the upside they have eliminated the headaches associated with managing the data center. The downside is that they still get caught unawares by changes made to this virtualization infrastructure as these changes unexpectedly and negatively impact their day-to-day operations. Further, they may find their provider cannot dynamically move their applications to the right resources during spikes or lulls in demand the way they were initially led to believe.
This is leading to the emergence of new technologies such as Dell’s Virtual Integrated System (VIS) architecture which is currently comprised of its Advanced Infrastructure Manager (AIM) and VIS Self-Service Creator. What makes AIM somewhat unique is that while it, as one might expect, manages Dell network, servers and storage, it also provides dips its toe in the heterogeneous waters by providing support for HP servers, Cisco switches and storage devices from other vendors.
Granted, Dell neither elaborates on what level of control it provides for HP servers and Cisco switches nor the breadth of support it provides for other storage vendors. However at least Dell recognizes with AIM that just because organizations are adopting virtualization and want a single pane of glass to manage it does not mean they are yet ready to have fully heterogeneous physical infrastructures.
But arguably more important at this early stage in the evolution is that we are at least starting to see the emergence of solutions like AIM that enable organizations to dynamically manage the placement of the applications that reside within their virtual infrastructure on the right physical device at the right time.
This goes somewhat to the heart of what Dell sees its Self-Service Creator providing: a rapid deployment of compute storage and server resources within a converged whole. This architecture enables organizations to put each of its virtualized applications on the right physical device while eliminating the time and effort that is normally associated with provisioning those resources.
While this approach has a tremendous amount of overlap with what cloud computing and cloud storage is in theory supposed to provide, in actuality all cloud computing and cloud storage architectures really do is pool compute and storage resources. To move the right application workload onto the right set of resources at the right time requires a higher level of intelligence that controls these resources as well as the applications that live on them.
For instance, a good example that Dell provided was that when a developer is ready to test a new application, using this model he can use AIM deploy this test/dev application it into the cloud onto resources that are assigned a lower priority or tier within the cloud infrastructure. But once it is ready to be moved into production, AIM can non-disruptively move it up to the appropriate pool of resources that have higher levels of performance with more aggressive recovery point and time objectives.
AIM will also in theory make it possible for organizations do a better job of capacity planning by having a better handle on the utilization of the physical resources within their virtualized environment. As most organizations know, doing effective capacity planning is more of an art than a science and, because of how dynamic environments have become, most organizations have found it more economical and easier to keep throwing resources at an application problem than really try to optimize it.
AIM looks to negate that approach by identifying where excess capacity (performance and/or storage) resides within this virtual pool as well as what resources are being stretched. In so doing, applications can be proactively and dynamically moved to where excess capacity is available or, if no resources are available, more capacity purchased and added to the pool.
Dell claims that in its early deployments with VIS that it is deploying workloads 90% faster, reducing the number of tasks required to deploy virtualized environments and overall providing for more responsive environments. Whether or not Dell is actually delivering on those claims, I am not certain. However what is certain is that if organizations are to fully leverage the benefits that a virtualized infrastructure provides then new infrastructure management tools such as Dell’s AIM are not just a nice-to-have, they are a necessity.