Operating System Level Failover (Without an Application Cluster)
The option of implementing failover in an application environment through the use of operating system level failover is a viable option for those not wishing to implement an application cluster.
The possible reasons for implementing this kind of failover include:
- The designated backup platform may not always be available (possibly used for other purposes)
- A particular failover technology is widely used and understood at a particular site
- There are no facilities available to handle the front end redirection of network traffic
- Application requires no specific special configuration for OS level failover
- The primary purpose is disaster recovery at a remote site
There are also some reasons for avoiding OS level failover:
- Dependence on platform-specific technology
- Does not improve performance
- Usually has a longer failover time (it generally requires a start of Sterling B2B Integrator, among other things)
- There is a risk in most environments for the two systems to become slightly out of synch which can cause a failure during the takeover process.
- It is not possible to reconfigure Sterling B2B Integrator without taking an outage
- There are a variety of partial failure scenarios that have to be accounted for to ensure that the takeover is complete