Listen to Dan Olds' answers to other frequently asked questions about the future of virtualization management in the Project FAQ podcast.
• Customers are making the move to virtual systems to increase efficiency and cut costs. What effect has virtualization had on internal IT operations and systems management?
• What kinds of problems are presented by running multiple applications?
• Policy violations? That sounds like the IT equivalent of a parking ticket.
• So how can virtualization become easier to manage for solutions providers?
• More on virtualization management and technology
• About the expert
The effect of virtualization on small and large data centers is profound. Customers are definitely seeing the advantages of virtualization, including much higher hardware utilization rates, lower power and floor space consumption and the ability to quickly bring up new applications in virtual systems. So, to a large extent, virtualization solves the problems that come from server sprawl. However, we're now seeing new problems emerge from VM sprawl. Running lots of different applications on single systems can bring on a whole new set of risks and headaches for IT managers.
To get to a virtualized infrastructure, users need to really understand their applications from a business perspective. They have to be able to understand usage characteristics, know when particular apps will see peak demand and know how to place an application with others that will not peak at the same time on a system. CPU, memory and network constraints must also be considered, because some applications have greater demand for these resources than others. Solutions providers can really add value by helping customers plan out their virtualization projects and ensure that the proper workloads are combined. This would guarantee that important apps are never starved for resources and that security and availability requirements are met. One of the main problems with virtualization is that it is so too easy to build and deploy new virtual machines. Users can easily fire up a new instance of an application or move an existing workload from one physical system to another in minutes. This leads to lots of virtual machines, many of which are thrown up on an ad hoc basis and may violate certain policies.
It can be much worse than that. What we see are users who inadvertently migrate a mission-critical application to a system that that has minimal hardware availability. They might have done this for a good reason, but if there's a problem with the host system, the company may find itself experiencing an unnecessary outage. We're seeing the same thing happen with security: Confidential data is placed on systems that do not have the appropriate security mechanisms in place to protect it. The fallout from a loss of confidential data can be substantial, and the effects on data center personnel can be pretty rough as well.
But there are some rapid changes [occurring] in this area. Several new products allow customers to set policies for virtualized workloads and ensure that those requirements are satisfied on the target system before allowing workloads to be moved or deployed.
As VM sprawl becomes more prevalent, data centers are looking for ways to get a handle on the dizzying complexity arising from having tens or hundreds of applications on a much smaller set of highly utilized hardware resources. From customers' perspective, the most efficient and effective virtualization management allows them to monitor and manage virtual machines on a large number of disparate systems. Clients want a common set of tools that will work with different virtualization mechanisms and a variety of systems, ranging from small x86 servers to enterprise-scale Unix and mainframe platforms.
While this is the goal, the industry is not quite able to reach it yet. What we have now is a lot of different products, ranging from all-inclusive suites from major industry players to point products from new vendors. Solutions providers need to become familiar with the various products so they can advise customers on the best available route. Installing and implementing virtualization management is also a solid revenue opportunity for the channel, as most customers will not have the personnel, bandwidth or skills necessary to do this job themselves.
- How to generate new business from server virtualization projects
- FAQ: Green IT services and virtualization technology in the data center
- Server virtualization security: Help customers avoid risks
- Application virtualization management suite due from Symantec
- How to deploy low-cost virtualization health check tools
Dan Olds has been in the high-tech arena for 15 years, holding significant positions at Cray Inc., Sun Microsystems Inc. and IBM prior to founding Gabriel Consulting Group Inc. in 2001. He was one of the first technologists to closely study IT total cost of ownership, virtualization and server consolidation. Dan is a frequent speaker at industry events and has been quoted in The Wall Street Journal, The New York Times, Bloomberg News, Computerworld, eWeek, InformationWeek, CNET and other tech news sources.