On the surface, it would seem to make sense that management is a task best performed in an organization by managers. When you apply that ethic to the emerging structure of data centers, which now use virtualization and private cloud foundations, you realize there are changes that can be made. Casting business resources as cloud services moves the budgeting process from capital expenditures to operating expenditures. And for more organizations, it means relocating management responsibilities from IT administration to a newly combined resource administration.
For these managers newly tasked with administering clouds along with people, admin tools don't make much sense. In a sweeping restructuring of its key virtualization management tools suite this morning, VMware is introducing a completely renovated dashboard for monitoring virtual data center operations, with graphs and 100-point-scale ratings designed to make better sense to people who might not, at first glance - or even second - know what any of this means.
Retooling VM administration as analytics"This transition from an architecture standpoint requires a new approach to IT management," states Martin Klaus, VMware's director of product marketing for vCenter Operations Management Suite, in an interview with RWW. With managers having a more marketing-centered view of the world, the new vCenter Ops will take more cues from the marketing mindset, beginning with using analytics more prominently in boiling down streams of data into need-to-know bullet points.
"As there are more moving parts that come and go more quickly, especially when you think about self-service portals," Klaus continues, "where the demand for more resources cannot be predicted up front, you need something that allows you to use analytics that give the IT administrator much more predictive controls over what's happening in environments, to intervene and intercept issues that are building before those issues impact the end users."
The revised dashboard is designed to communicate more ideas in shorter spaces. Perhaps you've noticed dashboards these days taking their cues from mobile apps, which have learned (out of necessity) to communicate greater information in shorter spaces. VMware group product manager Jai Malkani says the inspiration for nugget-izing vCenter's information into nuggets came from a recent reassessment of how its customers divide responsibilities among themselves.
"The customer teams working today in a cloud environment, what are they really focused on, and what's the top thing on their minds?" remarks Malkani. "As I worked with some 140-odd customers in the beta program over the last year, [I found] the main two areas that an ops team focuses on are: insuring and restoring service levels, making sure the problems are resolved and the environment is up and running at all times; and being pro-active toward optimizing the environment for efficiency and costs."
Administration is a two-cycle engine
Some of VMware's competitors perceive such tasks as workflows that can be diagrammed using flowchart tools. VMware is gambling that these two task areas that Malkani has identified are instead perpetual and ongoing cycles, where problems are identified, mitigated, and resolved; and ideas are generated, perfected, and implemented on a continual basis.
With respect to the cycle on the left, Malkani explains, the cloud administrator is focused on three areas: 1) The applications profile of the VM that appears to be the source of the problem. In other words, an app on that VM may be the actual source of the problem, but any remedial measure involving that VM will affect the entire profile of the apps or services in its purview; 2) Determining whether the problem is on account of how a problem app may be behaving in its environment, or instead the characteristics of the VM which presents that environment to the app - such as available capacity, or the current state of security monitoring; 3) Whether a corrective action is available in an immediately accessible manner - preferably, something which the admin can simply do and be done with it. A manager acting as an admin here would like to be given a multiple choice question, and choose the answer that appears to best resolve the problem - or to use Malkani's phrase for it, "closes the loop."
The "ideas" part of VMware's cycle are what Malkani calls optimizations - little improvements that come incrementally, instead of in great batches or overhauls. We see this concept emerging in so-called "resilience architectures," which replace typical crisis remediation methods with regular workflows that mitigate problems by their very nature, so that the responses to problems are essentially the same as everyday maintenance. "Do this [on the right-hand side] so that the left-hand side doesn't happen in the first place," he illustrates.
Fewer silos sounds like a good idea
The drive toward easier-to-understand metrics, and reducing and compartmentalizing dashboards with graphs and icons, is not a trend that was started by VMware. It's an increasingly competitive field, with an emerging ecosystem around management tools that fill in the gaps that management suites have left open.
Speaking with RWW, VMware's Martin Klaus admitted that one of his company's explicit goals for vCenter Ops was the reduction of the need for certain third-party tools, or for anyone else to come in thinking they need to patch the holes in vCenter. Its strategy here takes a cue from its competition, even borrowing some of its language: The new suite will incorporate tools that were offered for VMware's previous vCenter Operations tools as add-ons, including the Chargeback Manager tool that produces forecasts of future expenditures when current conditions are left as they are, compared to the savings recouped from making adjustments.
"Before, there were too many element levels, management tools in place that made for overlaps in some of the data," says Klaus. "But there's only one person sitting in front of these tools. And it was not possible for him to correlate the data that each of these tools were collecting. So with vCenter Operations, you can now see the data from third-party monitoring tools - from the storage layer, the networking layer, the database layer, the Web server layer - coming together. And now you can take a step back: What is really needed in terms of net-new data points? In this case, you can reduce the overall number of monitoring tools that are needed in the environment."
General availability of vCenter Operations Management Suite begins now, with various licensing rates beginning at $50 per VM.