Adopting virtualization in the enterprise is always a multi-step process. The first step is to identify servers...
to virtualize, and that's followed by a careful evaluation of return-on-investment (ROI) and capacity planning. That work has to be done before you get to the actual physical-to-virtual migration. Then, after the migration, you always face challenges in virtual infrastructure management.
So far, my step-by-step series has guided you through the migration. Now, let's turn our attention to the complex challenges of virtual infrastructure management, including mandatory control of physical and virtual resources availability, usage and access; deployment of disaster recovery solutions; provisioning of new virtual machines and other tasks; and monitoring and reporting of data center usage.
Because modern virtualization is still a very young technology, the market still faces challenges of immature tools and missing solutions. In particular, there's a void in the performance analysis and troubleshooting disciplines.
Challenges of liquid computing
Every IT manager in charge of any kind of system, virtualized or not, needs to know how to manage existing resources. The IT manager must keep track of physical machines, operating systems and products' licenses, services attainability and how assets satisfy demand, and the IT manager must be prepared to fast when a problem occurs.
These responsibilities can be quite time-consuming even on small environments, but they become even more complex in virtual infrastructures. In virtual infrastructures, IT managers have to worry about a new class of problems, like efficient and controlled deployment of virtual machines, rational physical resources assignment and accountability.
The ease with which you can create new virtual machines and their independence from underlying hardware leads to the idea of liquid computing; it can be hard to understand what is where in your computing environment.
The paradigm of liquid computing increases the risk of so-called "virtual machine sprawl." To avoid sprawl, virtualization management tools provide a reliable security system, wherein permissions can limit unauthorized creation of new machines. The tools also offer a strong monitoring system that reports on allocated but unused resources.
As of today, most virtualization platforms can leverage virtual infrastructure access with LDAP centralized accounting systems, but administrators are still in big troubles when they need to compute the efficiency of virtual data centers.
Going further, when a new virtual machine has been created, the virtual infrastructure manager faces the problem of deciding where it has to be hosted. As we already saw during the capacity planning phase, virtual workloads should be deployed carefully, considering along the way which existing workloads could be complementary in order to avoid overloading resources.
The upcoming Virtual Machine Manager from Microsoft, for example, will offer a rating system for physical machines, assigning one or more stars to each, helping administrators to immediately identify where a new virtual machine would best fit. This scoring system will adapt to the evolving infrastructure.
But even with such a system, in some environments, virtual machine creation may be not easy enough. For example, a big ISP remodeling its offerings with virtualization might need smart tools to deploy hundreds or even thousands of virtual machines on demand, in seconds.
At the moment, few third-party products can fill all virtualization management holes, and many companies opt to develop in-house solutions rather than spending money for available tools that have little flexibility. In such complex scenarios, virtualization management solutions have to offer software development kits (SDK) that allow wide customizations and different degrees of automation.
A wide, open programmable interface and strong support are key selling points; so far, VMware has done a pretty good job compared to its competitors.
Last, but not least, today's IT managers face a new problem: accountability. In a medium complexity corporation, several departments may work with virtual machines and share the same physical servers, using them in different percentages during a fiscal year.
When each of these departments has a cost center of its own, it can be difficult to track who should be responsibility for paying the costs of the underlying hardware. And even when costs are handled by a single entity inside the company, managers are challenged to enforce controls on who may use physical resources and how much can be requested of them.
At the moment only a few virtualization customers face these issues, but these will become common problems within a few years. Companies already in this boat may want to examine IBM's offering; IBM pioneered the field with its Tivoli add-on called Usage and Accounting Manager.
Multiple platforms, multiple issues
The above-mentioned needs increase further when a big company has to handle more than one virtualization platform.
In a large corporation, each department often has autonomy in choosing its preferred solutions, even if only one product will be used for a production environment. IT managers may need to concurrently manage VMware ESX server and Xen at the same time, hoping to leverage control with a single, centralized tool.
The market offering for such tools is multiplexing as the demand for them rises. Solutions from IBM, Cassatt, BMC Software, Enomaly and Scalent are the most popular, but new contenders like Opsware are coming.
Support for multiple virtual infrastructures often means IT managers do not have to worry about what technology has been used for creating a virtual machine; these tools are able to maintain control. Where possible, tools can perform application migration from one virtual hardware set to another -- something that is otherwise possible only with dedicated P2V tools.
When choosing one of these super-consoles, it's critical to verity that the console can leverage the existing management tools provided by virtualization vendors. Otherwise, return on investment may never come.