Advanced Cluster Server Solutions In a Nutshell by Data Center | Minakshi Sehgal

Building and maintaining enterprise computing infrastructures invariably involves balancing conflicting goals high missioncritical application availability versus lowest cost as one example. Now, Symantec Veritas Cluster Server features enable new server cluster server configurations that can simultaneously provide high availability at minimum cost.

For the first time, new, powerful Symantec Veritas Cluster Server features allow enterprise IT organizations to take significant steps in balancing the two previouslyopposing clustered application demands high missioncritical application availability versus lowest cost.

Introduction

Building and maintaining enterprise computing infrastructures invariably involves balancing conflicting goals high missioncritical application availability at lowest cost as one example. Fueled by today’s nonstop, global economy business demands, many enterprise IT departments necessarily deliver five 9s (99.999 percent) computing infrastructure availability translating to 5.39 minutes of total planned or unplanned downtime per year. This meager allowance applies to servers and a spectrum of customerfacing and other missioncritical applications including core financial service transaction systems, telecommunication systems, transportation systems, and online operations such as ecommerce Web sites.

To achieve five 9s availability, many enterprise IT organizations adopt application clustering, an approach that transforms multiple computer servers into a cluster a group of servers that acts like a single system. To create these application clusters, IT organizations migrate applications from monolithic architecture servers to server clusters having proprietary and commodity architectures.

High availability required, at any cost–or is it?

Attempting to maximize clustered application availability, IT organizations commonly deploy servers in one or more of three popular, highly inefficient, configurations, all of which can significantly increase data center costs. These application cluster configurations are:

Asymmetric Clusters
In asymmetric clusters, applications run on a primary server. A dedicated backup or spare server is present to takeover following failures. The spare server is unable to perform any other functions. Asymmetric clusters have the simplest and most reliable possible configuration. The secondary server stands by with full performance capability because no other application is running on it that can present compatibility issues. This configuration is also the most expensive, as the cost of failover server is 100 percent of the cost of the protected server.

Symmetric Clusters
In symmetric clusters, each server runs a specific application or service and provides redundancy for other servers in the same cluster. For example, if two servers run two different applications, either surviving server hosts both applications following a failure of the other server. From a hardware utilization perspective, symmetric clusters may appear far superior to asymmetric clusters since they overcome the common objection of expensive idle systems. However, this approach presents the most complex failover process and offers the least performance. Specifically, if a single application needs one processor to run properly, an asymmetric configuration necessarily requires two processors. Running identical applications on each server in a symmetric configuration therefore requires two dual processor systems, increasing hardware costs. Additional difficulties arise in symmetric cluster configurations when a single system must host multiple applications that do not coexist smoothly – further increasing hardware costs.

Article Source : www.symantec.com