|
The adoption of virtualization by corporations is being spurred on by the availability of high performance, multi-core servers that deliver the processing power, memory, and connectivity speeds required to support applications running on multiple virtual machines on a single physical server.
Companies are realizing they can save a great deal of money by virtualizing servers and server applications. In fact, virtualization can significantly reduce the number of physical servers required to run a business, thus cutting the associated long-term costs for management, upgrades, licenses, maintenance contracts, and warranties. Addition-ally, virtualization makes it easier to provision new servers, roll out new applications, and modify and re-launch existing applications.
Virtualization also helps reduce downtime. The reason is that by running applications on separate virtual machines there is less chance of conflicts or contention for system resources. In many cases, additional savings can be realized in terms of reducing energy requirements to both power and cool the servers since fewer physical servers are required.
For all of these reasons, virtualization is rapidly taking hold. For the majority of businesses, it is a case of when, not if they will deploy virtualization.
Many companies that have already embraced virtualization have done so for the cost savings alone. For example, a major food supplier undertook a large-scale migration and reduced the cost of procuring, configuring and supporting servers by 30 percent, according to a 2006 Baseline magazine article. Additionally, it also cut data center costs by reducing space, power, and cool-ing requirements. And it cut the time to deploy a server application from as many as five weeks to as little as 15 minutes.
The cost savings point is a common theme with regard to virtualization. A 2007 Ziff Davis Enterprise survey of 500 data center managers found that 66 percent of U.S. managers and 46 percent of man-agers outside the United States were using virtualization to contain costs.
Virtualization’s Impact on Storage
Server consolidation via virtualization dictates that associated storage solutions must have cer-tain characteristics.
Since multiple applications are now running on the same physical servers, the need for high avail-ability is greater, at both the server and storage layers of the system. The fundamental require-ment to deploying a highly available virtualized environment is access to shared storage, most commonly a SAN (Storage Area Network).
Since building a highly available virtual infra-structure is predicated on the fact that the serv-ers have access to shared storage in the form of a SAN, the underlying storage must be always available. Quite simply if there is a storage outage, it impacts many more applications at once. Therefore, high availability in the storage is a core requirement for customers deploying server virtualization.
Server virtualization also requires high perfor-mance storage. With multiple virtual servers all accessing storage simultaneously, the storage solution must perform under multiple concurrent workloads without degradation. And, the storage must support online scalability of performance to meet the ever changing and increasing demands of the business.
Storage solutions also need to offer great flexibility to accommodate a virtualized server environment. In particular, because businesses can quickly setup and tear down virtual servers to meet new application requirements, the underlying storage solution must offer rapid, easy, dynamic provisioning.
Additionally, the rapid adoption of virtual server environments in many companies means storage will need to scale to accommodate the applications running on virtual servers. Naturally, as the number of virtual servers grows, the storage solution must also be able to scale not just in capacity, but also in performance and bandwidth.
What’s needed?
Given the requirements that virtualization places on the underlying storage, the underlying archi-tecture must be highly available , reliable and perform under multiple workloads. The reason: since more applications are consolidated onto fewer physical resources, any system failure or downtime has greater impact.
There is also a need for a high performance storage architecture. The storage solution must eliminate connectivity and controller bottlenecks.
|
