Advantage/ Disadvantage between going for multiple subscriber/publisher

What would be advantages and disadvantages for
1) Having multiple publishers for each DataWriter (1-1)?
2) Having Single Publisher for multiple DataWrite (1-Many)?

Comments

RE: Advantage/ Disadvantage between going for multiple subscribe

1) Single Publisher for each DataWriter:

I see no specific advantage for maintaining a 1-1 relationship between Publishers and DataWriters (or Subscribers and DataReaders). The disadvantage of this approach is a small increase in memory consumption for each 'extra' Publisher (or Subscriber).

2) Single Publisher for multiple DataWriters:

Using multiple Publishers or Subscribers can be useful for grouping DataReaders and DataWriters. For example, the 'partition' and 'group_data' QoS policies are set at the Publisher/Subscriber, and apply for all contained entities. Also, listeners can be configured to propagate from the Readers and Writers up to their container [Subscriber or Publisher]; in some cases it might be useful to aggregate listener invocations.

There are no caveats when grouping multiple Writers [or Readers] within a single Publisher [or Subscriber]. There are no additional locking constraints, the internal data processing is not impacted, and asynchronous callbacks are invoked in a fair manner.