Availability Services (AvS) provide checkpoint/restart capabilities for distributed, parallel applications. Organizations can integrate Availability Services with their existing resource management product to ensure critical applications always run to completion.
Clustered computing environments are exceedingly common in most enterprise and high performance technical computing environments. In these environments, system failures are also very common – resulting in application downtime, wasted time, wasted money, and lost business. Evergrid's Availability Services product can ensure continuous availability of business critical applications, with seamless integration with your existing resource management solution.
Evergrid offers an OS abstraction based solution for commodity hardware that enables organizations to eliminate application downtime. AvS enables immediate recovery from any combination of application and system failures and management capabilities enable migration of applications to ensure continuous up time.
The Evergrid solution employs distributed checkpoint technology that continually captures the node, network, and file I/O state in distributed environments. This checkpoint technology works transparently to applications and operating systems, and works across multiple nodes and tiers. The result is more efficient use of computational and administrative resources and no lost revenues or computational time due to hardware or software failure.
Only the Evergrid Availability Services product offers this combination of features:
Evergrid offers continuous availability based on high performance OS abstraction technology that decouples applications from operating systems in distributed computing environments. As a result, applications can be transparently preempted and resumed on another node or tier – immediately and without interruption.
AvS features sophisticated check point technologies that are extremely efficient, requiring minimal processing overhead, usually less than 5% overhead from systems. The product offers these checkpointing features:
Application Transparency
In the past, specific applications had to be custom coded to do checkpointing. Evergrid software requiring no application or operating system modifications, and is able to capture the state of any application without stopping the application or stopping the IO and communications from the application.
Flexible Integration with 3rd party Schedulers
AvS offers easy integration with 3rd party job scheduling software. All interaction with Evergrid takes place through environment variables, flat text config files, and operating system signals. This simple interface makes it easy to integrate EAS with legacy products.