[asterisk-dev] CSTA - Worth looking at again?

Brian Degenhardt bmd at digium.com
Thu Oct 16 12:03:41 CDT 2008

John Todd wrote:
> [This has come up a few times over the last 4 years on this list and 
> others, but it has never before had enough people look at it who 
> understood perhaps what the benefits were, so I drag it out again 
> here for your amusement...]
> Would a CSTA interface (or in a variant, uaCSTA) be a useful API 
> layer for Asterisk?  Would some of the PineMango goals be workable 
> with a CSTA method as the "language" portion of the communication? 
> Or would it merely be a monolithic module that would talk to the API 
> components as a discrete consumer/producer of messages to the various 
> components?  I'm largely ignorant of the details of where the API 
> discussion would layer into what something like CSTA could offer - 
> feel free to say that these two concepts are non-intersecting and 
> I'll drop it.
> Regardless of CSTA's role in a potential "next-generation" fruit-like 
> API project, it still bears examination as a potential interface for 
> other Enterprise-ish platforms.  Microsoft's OCS/LCS Platform uses a 
> variant of uaCSTA over SIP, as does Avaya, Cisco, Mitel, and 
> others(?- need a list).  Of course, I'm not so interested in those 
> legacy PBX platforms, but the tools that third-party companies are 
> producing to bolt onto them are what sound like has some value. 
> Creating a CSTA interface to Asterisk would allow quick grafting of 
> "industry standard" tools onto an inexpensive and flexible OSS 
> back-end.
> Comments?
> Base page:
>    http://www.ecma-international.org/activities/Communications/TG11/cstaIII.htm
> Here's an introduction (194kb):
>   http://www.ecma-international.org/activities/Communications/TG11/CSTAoverview.pdf
> This seems to be some of the documentation (4mb, 767 pages!):
>   http://www.ecma-international.org/publications/standards/Ecma-269.htm
> CPAN module for CSTA (incomplete - Phase I only - need Phase III)
>   http://search.cpan.org/~leifj/Net-CSTA-0.04/lib/Net/CSTA.pm

Again referring to Nir's diagram:


I definitely think that CSTA is worth looking at, either in the
tan/olive layer or as a shim layer in orange that translates the subset
of res_api that's applicable to CSTA.

Also, I think that CSTA is very helpful to get an idea of what a good
telephony API looks like.  We should borrow heavily from it when
designing res_api.

However, CSTA alone will probably not satisfy all our needs.  There's a
lot of things you might want to do that CSTA doesn't address such as
peer/account creation/modification, registration status/hooks, music on
hold, queue announcements,  detailed queue status and other
Asterisk-specific features not addressed by the CSTA spec.


More information about the asterisk-dev mailing list