Virtualization Management – VMware and Dell, Big 2 or Big 6?

Virtualization Management

With Dell buying Quest, and VMware buying DynamicOps, the Virtualization Management landscape has been forever changed. Now Dell is a full fledged systems management vendor, and VMware has crossed the line into managing both their own and other hypervisors, and being able to construct clouds that even include non-virtualized resources. This gives rise to a very interesting question. Are Dell and VMware turning into traditional systems management vendors like CA, IBM, HP, and BMC or are they preparing to disrupt the existing systems management business just like VKernel (part of Quest, and now Dell) and DynamicOps did when they were startups?

How Does Virtualization Management Differ from Systems Management?

Traditional systems management is really all about managing very large, very complex, very heterogeneous systems that do not change very rapidly. Processes like ITIL are used to document the steps used to hand off tasks between groups, document the hand off criteria, and document the required response times. The state of the system at any point in time is supposed to be reflected in a CMDB – but the CMDB is rarely fully up to date with even a slowly changing physical infrastructure. True performance (response time and latency) is rarely measured – most of the time it is inferred from normal or abnormal resource utilization statistics.

The complexity of the undertaking, wide variety of supported systems, and wide variety of management functionality in traditional systems management leads to software that is expensive to purchase, time consuming and expensive to install and configure, and expensive and time consuming to maintain and operate. The dollars involved in systems management decisions cause these decisions to be made at the highest levels in the IT organization. This creates the first of many problems for the traditional systems management business as the decision makers tend not to be terribly technical and tend to be disconnected from the actual use of the products. This means that in many cases, systems management products are purchased that do not meet the technical needs of the people who actually need to use these products to do their jobs.

Virtualization Management on the other hand benefits from virtualization itself to create a problem that is in some respect dramatically simplified. Virtualization allows multiple different systems (for example Windows and Linux servers) to be managed consistently. Virtualization also creates a new measurement point – they hypervisor itself which “sees everything” and which can therefore be a valuable source of management data.

New Problems Created by Virtualization

Despite the fact that as stated above, virtualization has some simplifying effects upon the environment, it also creates an entirely new set of problems:

  1. Static and slowly changing systems become dynamic and rapidly changing systems. This one factor alone turns the offerings from traditional systems management vendors into legacy boat anchors. ITIL is simply not agile and responsive enough, and legacy CMDB’s cannot be updated quickly enough to keep up with the rate of change.
  2. Resource utilization becomes a lousy proxy for performance. In the physical world, most systems and applications are managed for performance by software that looks at resource utilization and performance is inferred from whether or not resource utilization is within normal bands or not. This approach does not work at all in virtualized systems as resources are too shared and abstracted from the underlying hardware to be useful.
  3. Timekeeping problems in virtual machines (documented in this VMware White Paper) make the running management agents of the kind that legacy tools typically rely upon into a worse than useless proposition (consumes time and money and adds no value).
  4. Owners of applications hold virtualized systems to higher standards than the physical systems that the applications used to run on. This basically boils down to a lack of trust. Application owners trusted the systems when they were dedicated to their applications and they could over-provision to their hearts (and budgets) content – but they do not trust dynamic and shared systems. Therefore owners of virtualized systems have to produce better and different management data about the operation of those systems than was necessary in the physical world.
  5. Virtualization enables and invites even more rapid changes. The ease with which a new server can be provisioned invites the provisioning of more servers and leads to sprawl. Private clouds enable business constituents to provision their own servers when they need them. Image management facilitates the frequent updating of servers driven by Agile Development.
  6. Virtualization in some cases leads to workloads running in public clouds or hosted virtual environments where the resources are not only abstracted, they are now owned by a different business entity making key resource metrics not only irrelevant but largely unavailable to application owners.

In summary, we are moving from infrequently changing applications running on static infrastructure to rapidly changing applications running on dynamic and distributed infrastructure. Legacy systems management tools from the legacy Big 4 systems management vendors were not built to meet these requirements. Furthermore, these requirements are so different than the case that the Big 4 tools were designed to meet that virtualization effectively breaks these tools and makes them simply not work in the new environment.

The New Virtualization Management Market

Owners of virtualized environments recognized these changes long ago, and starting looking for tools that worked the met these new needs and use cases. Two very significant things have happened as a result. The first is that a new buyer has emerged for management tools. That new buyer is the person that owns the virtualized environment. The second is that a new model for delivering and selling software has been popularized by the success of SolarWinds and Splunk. That model is all about downloading the product and trying it for free, in production, before buying it. This turns the old enterprise sales model of taking the CIO and his VP’s out for a couple of rounds of golf on its head. Now the people who have to live with the software are making the buying decision and the sales rep’s from IBM, CA, BMC and HP are not even at the table nor aware that an evaluation is occurring.

Decisions for VMware and Dell

By their actions last week, both VMware and Dell have clearly stated that they intend to be major players in the virtualization management market. The question is, are they going to behave like typical big vendors and try to cram ELA’s for suites of software down the throats of their customers (as IBM, CA, HP, and BMC do) or are they going to behave like the startups that got this whole virtualization management business rolling in the first place.

The first piece of evidence in this regard came when Quest bought VKernel. Instead of folding VKernel into the vFoglight team, Quest did the opposite. They saw the value of VKernel’s “easy to try and easy to buy” model and VKernel became the virtualization performance and capacity management division of Quest. If Dell sticks with this decision and builds upon it then we will know the answer. And the answer will be that Dell will seek to disrupt the traditional systems management business instead of simply trying to participate in it. More evidence in this regard emerged during the analyst call on the Dell acquisition of Quest when John Swainson, the President of Dell Software said that he was glad that he “did not have a legacy revenue stream to try to protect”. Obviously if Mr. Swainson is not worried about his own legacy revenue stream he is going to be even less worried about the legacy systems management revenue streams at IBM, CA, HP, and IBM.

On to VMware. VMware has been making it very clear for several years that it views virtualization as a force that disrupts the existing systems management business and that it intends to create a new virtualization management business around the virtualization platform with VMware as a leading participant in that new business.  However, VMware’s execution on this front has looked and felt a lot more like that of a traditional systems management vendor than that of an agile startup in the virtualization management business. A lot of companies have been acquired, and a lot of time and effort has gone into integrating them and reducing complexity. A good example of this is vCenter Operations, the Enterprise version of which includes the former ConfigureSoft product, but still does not yet have that product fully integrated into the rest of the product.  VMware does make its products easy to try as they are readily downloadable off of the VMware web site, but in many cases far too much time and effort is needed to properly install and configure them – especially when one compares this with how easily products like Embotics, VKernel, and Xangati come up and start providing value.

The acquisition of DynamicOps by VMware is probably a milestone event in the virtualization management business for VMware in two respects. First, while DynamicOps is has tremendously rich functionality, that richness does not come at the expense of the kind of complexity that requires ongoing vendor provided consulting to bring up, configure and customize the product. Second, prior to this acquisition most of VMware’s offerings were limited to the vSphere platform. VMware now has a private cloud management offering that runs on multiple hypervisors and physical resources. While VMware has made no announcements to this effect, it would be reasonable to assume that now that the ice has been broken, the rest of VMware’s management offerings will follow suite. If the rest of VMware’s management offerings can emulate DynamicOps’ combination of rich functionality, cross-platform support, easy to try and easy to buy, and rapid time to value this will transform VMware into a formidable virtualization management vendor.

Making Enterprise Virtualization Management Decisions

The Dell acquisition of Quest and the VMware acquisition of DynamicOps gives enterprises some new choices when it comes to their portfolio of virtualization management solutions. The first recommendation that we need to make is that the requirements for effective virtualization management solutions are so different than traditional systems management requirements that you should not simply assume that your incumbent systems management vendor’s products are up to the job. Any reasonable virtualization management evaluation strategy should absolutely look at the offerings from VMware and Dell/Quest, as well as the offerings from the numerous smaller vendors who have excellent specialized solutions. The entire ecosystem of Virtualization Management solution is depicted in the diagram below.

The Virtualization Management Vendor Landscape (click to enlarge)

Virtualization Management Vendors


Virtualization Management is sufficiently different from systems management so that it is most likely that VMware and Dell will behave completely differently with respect to product, packaging and pricing decisions than do HP, BMC, CA and IBM. This will increase the rate of disruption in the management software market, and hasten the demise of the legacy systems management vendors.

Posted in IT as a Service, SDDC & Hybrid CloudTagged , , , , , ,

Leave a Reply

1 Comment on "Virtualization Management – VMware and Dell, Big 2 or Big 6?"

Sort by:   newest | oldest | most voted

“This will increase the rate of disruption in the management software market, and hasten the demise of the legacy systems management vendors”

LOL. Do you seriously think Dell is going to displace, IBM, MS, HP, BMC and CA? Dell doesn’t have software in their DNA. They’re likely buying Quest to try and build this business. BEST case, this will take many, many years.

Reality has set in for vmw. Competition is everywhere on private cloud and they are no where on public cloud. Cloud Foundry is vapor, slow and buggy. AWS and MS are years ahead.