Multipoint Service Redundancy and Fault Tolerance

A high level of fault tolerance to upstream disruptions can be achieved by using redundant upstream resources, with each redundant resource providing services for a sub-set of the servers in the farm.

If one of those redundant upstream resources fails, the server farm Service State Determination Agents (SSDA's) will report a bypass for the servers affected by the upstream loss. New connections will be directed to servers in the farm that still have access to those upstream resources. When upstream resources are restored, the SSDA's merge the affected servers back into the farm. Redundancy can now be built into, and used, further up the service stream. End-point services are aware of upstream performance and loss, allowing corrective action to be taken automatically with the least impact to users/customers.