It is to my understanding that the work group is moving towards using
adaptation service (either local or remote) to handle content adaptation
decisions base on network traffic conidtions. The rationale being the rule
engine should be kept as efficient as possible.
For example, when content adaption is needed for an A/V stream, a rule
simply states that all ccontents belonging to the A/V stream be forwarded to
the adaptation service. It is the adaptation service that will decide what
kind of adaptation is required.
This leads back to the problem I face which leads me to OPES in the first
place. How does the adaptation service obtain network condition
informations? There are couple of ways that are either non-standard, or need
to go lower than the session layer in the protcols stack.
I think it would be nice for there to be standard interface for querying
network conditions at the application layer. That's why I think it was
extremely useful if OPES Rule engine could indeed provide that interface. I
hope the authors of the requirements and rule language drafts can consider
this.
A possible scenario would be for the rule engine to incorporate optional
modules to perform QoS related decisions. When the adaptaion service is on
a remote server, it might be possible for the providers to have the central
OPES intermediary not have the optional QoS-module. Instead, an OPES engine
with QoS-module on the remote adaptation server.
Comments are welcome.
Chan-Wah Ng
Panasonic Singapore Laboratories