30 likes | 182 Views
P: Message Processing Language (draft-ietf-opes-rules-02). Alex Rousskov* (rousskov@measurement-factory.com) Andre Beck* (abeck@lucent.com) *Presented by Markus Hofmann. Document Status.
E N D
P: Message Processing Language(draft-ietf-opes-rules-02) Alex Rousskov* (rousskov@measurement-factory.com) Andre Beck* (abeck@lucent.com) *Presented by Markus Hofmann
Document Status • September: WG decision to drop work on the Intermediary Rules Markup Language (IRML) in favor of moving forward with “P”. • Mainly a style preference. • Characteristics of “P”: • Centered around the concept of an “object”, similar to objects of object-oriented languages. • General approach is application protocol agnostic. • Supports loadable modules for adding support of (existing and new) application protocols. • Good progress made, but several important open issues and still some work to be done. • Will we have a final document candidate for WGLC by end of November?
Open Issues • What (message) information can the P interpreter access, i.e. what information can be part of a rules condition? For example: • Complete message (including message body), • Meta-information only (e.g. HTTP headers only), • Where to draw the line? Does the WG have to specify this? • Should the WG document an HTTP module for P? • If yes, should specification of such module be part of the HTTP adaptation draft, of the P draft, or a separate document? • Should the WG define interfaces between P interpreters and module suppliers and/or callout services? • How do services return results? Should all issues be addressed under the current charteror under a possible re-charter?