If a service works on message content (body), it would easily handle
several application profiles by skipping or ignoring everything other
than the message content. For example, a virus scanning service can
handle both HTTP and SMTP content.
Currently, profile result negotiation is attached to a Service Group,
which is, essentially, a service. The following is not clear to me:
- can two profiles be negotiated for the same group?
- if yes, how can the agent distinguish which profile
is being used for a given message?
If we restrict the number of profiles per SG to one, the only way to
support the above would be to create two identical SGs. In this case,
we may want to support SG creation based on another SG rather than
based on a raw list of services.
If we allow multiple profiles per SG, we need a new identifier to
specify which profile is being used.
I think I prefer the former, to keep things simple.
Thoughts, preferences?
Thank you,
Alex.