|
- Processing Capacity:
The processing capacity does not diminish
as the number of systems in the complex increase. This is because
all requests for global resources are handled by the coupling facility;
not by passing a token between each system in the complex before a
request is satisfied. The requests are processed by the coupling facility
as they are received, therefore the processing capacity is improved.
- Response Time:
Response time is improved in a star as each
request for a resource that is not in contention can be completed
with only two signals, rather than having to pass the RSA-message
to each system in the complex.
Also, in a star complex, the
central processing unit (CPU) overhead required to process an ENQ
or DEQ request is limited to the system on which the request originated,
the coupling facility, and the system chosen to be the global contention
manager by XES (if one is needed). Therefore, the total processing
time consumed across a star complex will be less than that consumed
by a ring complex.
- Availability and Recovery:
Availability and recovery time is
improved for the star because the systems that make up the complex
are not dependent on each other, as they are in the ring. The complex
does not have to alter processing to adjust to topological changes
due to a system joining or leaving the complex. When a system joins
or leaves the complex, queue manipulation for resource requests are
distributed around the complex, rather than having to be repeated
for every request on each system. These changes and adjustments are
handled by global resource serialization and XES. Your only requirement
is to determine which resource owned by a failing system is in contention
and reassign it to a new system.
|