Between a pair of MTAs, M-Switch provides a number of related options
to control associations and priorities. There are two top level mechanisms
– Intervals and Schedules:
- An interval specifies a period when messages may be sent. This is
used for two basic tasks.
- Where the peer MTA is available for a known period (e.g., an
MTA on a submarine coming into scheduled contact).
- In conjunction with priority control, to define a period where
lower priority messages are not sent (e.g., in support of a military
minimize condition)
- A schedule is a regular interval, specified by day of week, hour
and minute. This is used for:
- Regularly scheduled communication to MTAs with specific availability
(as in 1a above).
- In conjunction with priority, to only send low priority messages
at specific times (e.g., to not send lower priority messages in
know busy periods)
- A special “Always” schedule, which is used when
there are no timing constraints, and when other features described
below are needed for policy or performance reasons.
Intervals and Schedules have two basic modes:
- On Demand. This type of interval/schedule is essentially a permission
to open one or more connections.
- Permanent. This type of interval/schedule requires a single permanent
connection (association) to be held open. The reason for this is that
it takes typically 2-3 seconds to open a connection, which is much
longer than typical message transfer time. Use of permanent associations
enables message switching time for small messages to be reduced from
a few seconds to a few hundred milliseconds. Multiple permanent associations
can be specified. If you need to open on demand connections in addition
to the permanent connection, an on demand interval/schedule needs
to be added.
Intervals and Schedules can specify a minimum priority (precedence)
of message to which the interval/schedule applies. This control can
be used for a number of reasons:
- As a resource reservation mechanism, to ensure that specific connections
for higher priority messages, are always available.
- To give very low latency switching for higher priority messages,
while allowing lower priority messages to open connections on demand.
- To prevent transfer of lower priority messages, either in support
of scheduling or in support of a minimize condition.
Where an interval or schedule specifies a permanent association, the
priority relates to that association and will ensure that only messages
of the specified priority or higher will make use of that association.
Where an interval or schedule specifies on demand connections, an associated
priority specifies a policy that messages of lower priorities shall
not be sent, unless permitted by another interval or schedule. This
mechanism can be used in isolation or in conjunction with permanent
associations to prevent transfer of lower priority messages.