W3C home > Mailing lists > Public > www-voice@w3.org > October to December 2002

CCXML implementation model

From: Sharma, Ranjan (Ranjan) <ranjansharma@lucent.com>
Date: Thu, 31 Oct 2002 16:20:14 -0500
Message-ID: <4C37CF2D8DF07E4CA6357BAD5EB9A5D703E8410D@oh0012itsa1.cb.lucent.com>
To: www-voice@w3.org


With the VoiceXML specs in the background, it is not very easy to see the
implementation model of CCXML. With VoiceXML, a call comes into the
implementation platform which starts the session; with a look up of the
corresponding URI to fetch the VoiceXML page via HTTP. The VoiceXML page
contains the service logic.

What is the equivalent CCXML execution model? Are the CCXML documents
fetched just like VoiceXML documents?

I believe the execution model would be different for different capabilities.

Selective acceptance of an incoming call is one model. Here, the CCXML code
is waiting for the transition "alerting". So, the code is already executing
and is in idle mode. For an outbound call initiation, a different model
would be used probably. A conference call set-up might be a combination of
these two models.

Let us take the example of an outbound call. If this is a "wake-up" service,
some one has to provision the time (say 5 AM) and the number (say
XXX-YYY-ZZZZ) to call through an interface (extraneous to this discussion).
Some "agent" has to keep track of what number (XXX-YYY-ZZZZ) to call when (5
AM) and to do what (play an announcement "this is your wake-up call" or play
a jingle etc.). The underlying architecture would dictate roles and
responsibilities. Possible alternatives are:

(a) Single platform in charge of holding and executing the script: At 5 AM,
this platform could initiate an outbound call, run the CCXML script using
the telephony elements on the platform. 

(b) Two different platforms - one web/app server that hosts the CCXML script
and accepts provisioning requests and one execution platform: At 5 AM, the
web/app server sends a "push" to the implementation platform to trigger a
call via an HTTP POST with the CCXML script

(c) Three different platforms - one app server, one execution platform and
one web server that does time keeping and accepts user inputs: At 5 AM, the
time-keeper platform can send a "push" to the implementation platform to
trigger a call via an HTTP GET, with the URI to fetch the CCXML script
available on the app server

Clearly, load considerations and traffic plays a role in the scalability of
the solution.

My question is - is there a recommended CCXML execution environment, kind of
like the document server model as described in the VoiceXML specs?


Received on Thursday, 31 October 2002 16:20:17 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 23:03:47 UTC