Falconstor Community

Server or application clusters are commonly used for mission-critical business applications that require maximum uptime. For example, imagine a financial hedge fund making bets with money it does not have; they need their equity/bond/currency trading applications continuously available during trading hours. Those applications are commonly deployed on clusters, Windows or Linux (in the open systems world). A cluster can seamlessly tolerate a server failure by simply transitioning operation to a standby, or passive, server node. Cluster node state is continuously monitored so that when a fault is detected, service failover is automatic. In my example, the hedge fund can either make tons of money or lose tons of money; assured that they are “too big to fail” as real financial regulation will never happen (they do, after all, have the largest and best funded political lobby on the planet).

Local cluster applications are available from a number of vendors, including Microsoft, RedHat, SUSE Linux and Veritas (Symantec). Additionally, “stretch” or “geo” clusters are available. Two that come to mind are Microsoft GeoCluster and Veritas Cluster Server; these two, in particular, tolerate a site disaster and transition operations to another data center.

Cluster implementations are also great for planned downtime or maintenance (upgrades or patches), as application operations can easily transition to another node in the cluster. Most cluster services support major business applications, such as SAP, Siebel, Oracle, PeopleSoft, and Microsoft Exchange; as well as major databases such as Microsoft SQL, Oracle, and DB2. Lastly, most cluster services support both physical and virtual servers.

While clusters do indeed provide high availability for mission-critical applications, the next question becomes “which applications are critical”? This can easily become a ‘religious or political’ discussion or argument. Considering the relative high costs of cluster implementations, most organizations are forced to make difficult decisions as to which applications are deployed on HA clusters.

FalconStor offers a unique alternative to application protection, that of service orientation, with our RecoverTrac tool. As a standard feature of FalconStor NSS and CDP products, RecoverTrac allows customers to define and implement application service levels (SLAs) for all servers and applications with automated local, remote, or multi-site recovery. RecoverTrac allows administrators to execute recovery jobs that automatically boot recovery servers (physical or virtual), including their resident applications and data volumes, and reconfigure networking definitions so the business can quickly (and automatically) restore some or all servers and applications.

High availability clusters, including geographically distributed, deliver application availability, or resilience, for mission-critical business applications. However, from both a logistical and staffing perspective, it is generally cost prohibitive for a business to deploy all its applications on clusters. FalconStor offers compelling application protection and automated business recovery for all or selected applications, utilizing your existing infrastructure, in a low-cost and simple implementation.