I recently participated in a Ziff Davis Enterprise eSeminar on cloud computing strategies, which you can listen to online and download the slides at eseminarslive.com. Based on the audience questions and the fact that nearly 100 percent of attendees stayed to the end, I'd say the topic was of interest, so I'm continuing the discussion here.
Much of what I talked about was culled from listening with a keen ear to both the case study presentations and audience questions at technical conferences. An emerging area of interest that IT managers should pay attention to is chargeback. Virtualization platforms in the enterprise are catching up with public cloud offerings in being able to track compute, network and storage usage. To paraphrase my favorite Spider-Man aphorism, "With great billing comes great accountability." In other words, as business units start getting a bill for unused or underutilized virtual systems, life cycle management-especially the termination part of that life cycle-will assert itself.
It's also become clear to me that virtualization projects-in particular those intended to create a cloudlike environment in a private data center-benefit from being rolled out in phases and are in fact hurt by an all-or-nothing approach. One example from VMworld: A major grocery retailer rolled out a successful server virtualization project by starting with just a few stores in a region. As the deployment lessons were learned (some applications ran only on guest systems for which the vendor had to do development work to support) and operational kinks were worked out (some management system timing delays over great distance interfered with reports), the grocer's IT staff gained the experience necessary to make the investment pay off.
I worked with Aaron Goldberg, a Ziff Davis market expert, on the presentation. During the course of our online discussion, I realized what was so important about all the resource pooling upon which virtualization and cloud computing is based. In essence, well-crafted resource pools should lower the technical ability of the person requesting an IT resource to accomplish a business task.
Further, the hard, expensive work of creating compute, network and storage pools and ensuring that applications can work with these pooled resources requires that IT functional groups work together at an almost unprecedented level. IT managers should explore every option to work with peer experts in adjacent technical fields. Without internal cooperation, internal cloud platforms are almost certain to be stunted instead of stunning.
As we progressed through the eSeminar, I returned to a theme I've sounded for some time: Virtualization in and of itself is losing its edge as a way for IT to stand out as a "good guy" in the organization. Increased server utilization, workload mobility and effective HA/DR (high availability/disaster recovery) are becoming standard operating procedure. Effective and efficient management of virtual resources is what will distinguish IT managers in the current phase of data center operations. And effective resource management will only gain in importance as cloud services continue to advance.