UDDI Failure and Recovery

UDDI failure is part of the protocol.

Distributed systems have more failure points than stand-alone systems, it is important for clients to be able to detect and/or recover from failures that occur during interaction with remote partners.

UDDI addresses quality-of-service issues by defining a calling convention that involves the use of cached bindingTemplates.

If the call fails, the program uses the bindingKey value and the get_bindingTemplate API call to acquire a new copy of a bindingTemplate for this unique Web service.

The program compares the new bindingTemplate information with the cached version and, if they are different, retries the failed call using the new bindingTemplate.

If both versions are the same, the client retries the call. This approach, called "retry on failure," is more efficient than acquiring a new copy of bindingTemplate data prior to each call.

The following scenario describes how error recovery fits into Web services:

A developer prepares a program to use a Web service, caching the appropriate bindingTemplate for use at runtime.

During program execution, the program calling the remote Web service uses the cached bindingTemplate that was earlier obtained from a UDDI Web registry.