|
As server virtualization goes mainstream, the workloads of virtual machines are taking on mission-critical proportions. Uses range from application-dense server consolidation, to failover and disaster recovery, to managing the virtualized infrastructure. Downtime and data loss pose harsh consequences in scenarios like these. This paper looks at five use cases to show why, where and how your business can reduce risk in your virtualization initiative by hardening your IT environment with physical servers that ensure reliability and availability.
Virtualization in the Mainstream
Virtualization of x86 servers has already made its mark on IT, and the surge shows no indications of stopping. For example, a 2007 article in Network World reported that Forrester Research found use of server virtualization growing from 29% in 2005 to 40% in 2006 among North American companies that were surveyed. The percentage of companies piloting virtualization remained level at 11%.1 An article in IT Week reported that IDC forecasts customer revenue from virtualized server sales in Western Europe will grow from $948 million in 2006 to $5.5 billion in 2011.2
When most virtualization initiatives were still at an early stage, the typical application involved didn’t by itself consume a lot of system resources, was forgiving of slow response time and did not necessarily represent a high priority to the business. That’s changing as virtualization is more widely deployed. IT managers are starting to look beyond their initial goal of hardware consolidation, and the resulting benefits of improved server utilization and better return on investment. They also want to use virtualization to achieve other goals, including higher availability and disaster recovery.
Yet putting your “eggs in one basket” should be done judiciously, even when virtualization lends resilience and flexibility to that basket. For starters, virtual machines run on physical servers, whether they are conventional servers or availability-promoting designs. And when the hardware fails or a transient error occurs, all the virtual machines on a server may stop and must be restarted unless you use a fault-tolerant server technology that offers the necessary protection.
Also keep in mind that the virtualization layer has the potential to be a single point of failure for all of the virtual machines it supports. The exposure rises as you run more virtual machines on a single server.
Use Cases Raise Criticality Quotient
The next section of this paper examines five use cases where virtual machines and infrastructure assume mission-critical stature under the right conditions _ even when the uses don’t fit pre-virtualization notions of what is essential to the business.
What we will describe as the Criticality Quotient, or CQ, of these use cases is often relatively low when an organization first implements virtualization. The CQ rises as more of your business processes, more users and more applications depend upon virtual machines and infrastructure.
Server Consolidation
According to a 2007 report by Yankee Group, 85% of virtualization software acquired by enterprises goes to server hardware consolidation projects.3 The report also states that businesses have been able to better their hardware utilization from 10% to 80% in certain instances. The resulting savings are not limited to hardware-related costs, but also extend to management costs and environmental expenses such as space, power and cooling.
What’s more, we have observed some companies opting to standardize on virtualization for new deployments of x86 servers. These adopters want their applications to run on a shared virtual pool of server processing and storage that can be rapidly reallocated to provide capacity while avoiding underutilization.
Virtualization leader VMware cites that 10-15 production systems may be consolidated onto one server. Even an outage of non-mission critical applications has a mission-critical impact when a large number of users are affected.
Live Migration
Live migration allows virtual machines to be moved (in cases where the operating system is not visible to the application) with virtually no interruption to the application and little perceived impact by users. This provides the flexibility to handle peak processing loads and to shift applications for maintenance. Hardware and operating system upgrades can be accomplished without planned downtime, for example.
Live migration works by copying the system state iteratively while the application continues to run.
|
