OSB Service virtualization & Error Recovery & Location Transparency

There are couple of times I was unclear while trying to understand OSB load balancing & Error recovery algorithms. So thought of collecting them in a post.


Service Virtualization:
When an external service has hosted on more than one host, the OSB business Service component will route the requests/load on active hosts.

There are couple of load balancing algorithms to support this.

Load balancing Algorithms:
The order in which the business service calls or retries the endpoints controlled by load-balancing algorithm.

- Round Robin:
This dynamically orders the URLs that we enter in Endpoint URI field.

- Random:
This randomly orders the list of URLs that we enter in Endpoint URI field.

- Random-weighted:
This randomly orders the list of URLs that we enter in Endpoint URI field but some retried more than others based on value that we enter in Weight filed.

- None:
No load balancing occurs. The first URI Endpoint is always used unless it is unavailable.



Error recovery options:

- retry count: The retry option specifies for maximum number of times a business service can attempt the URI provided.
If the value 0, business service never retries after failure.

- retry count interval: wait time in minutes after retring all endpoints per count