OSI/MHS Configuration and Management Manual

Planning Your OSI/MHS Subsystem
OSI/MHS Configuration and Management Manual424827-003
3-15
Resilience and Expandability
However, for better performance, you may want to spread the workload across multiple
CPUs. For example, if you wanted to distribute the workload of a single MR group, you
could put the MR process on one CPU and the RTS and SC processes on another. Do
not put processes from different groups on the same processor. For example, do not
put a process from MR group 1 on the same CPU as a process from MR group 2. If
that CPU fails, it would cause both groups to fail.
Resilience and Expandability
You can define multiple groups to provide resilience and expandability. For example,
to increase the throughput of an MTA, you can configure additional MR groups.
GI Groups
When planning GI groups, consider the following:
How many GI groups do you need? Consider the type of network you have and
how many associations that network can supply. (The maximum number of
simultaneous associations per GI group is 1.)
How many general user gateways do you anticipate?
Do you need to define a P1 user exit gateway to intercept messages as they arrive
in the system?
LO Groups
When planning LO groups, consider the following:
How many LO groups do you need? Consider the type of system you have, how
many local users that network can handle, and how many associations that
network can supply. (The maximum number of simultaneous associations per LO
group is 30.)
How many local UAs do you anticipate?
Do you have MS groups on your OSI/MHS subsystem? If so, you must have at
least one LO group.
MR Groups
When planning MR groups, consider the following:
How many MR groups do you need? You can increase capacity by adding more
MR groups. Each MR group can handle up to 20 simultaneous associations. The
minimum number of MR groups per MTA is one for each OSI address on which
you want to listen, but several MR groups can listen on one address.
How many and what type of connections will you have?