W3C home > Mailing lists > Public > www-voice@w3.org > April to June 2002

Re: Need to access DB in VoiceXML but not Server-side

From: Dave Burke <david.burke@voxpilot.com>
Date: Fri, 26 Apr 2002 18:18:55 +0100
Message-ID: <063501c1ed46$72319430$8600000a@VOXMOBILE>
To: <Hitesh.Seth@silverline.com>, <tehdy.draoui@polygones.com>, <www-voice@w3.org>

(As I suspect Hitesh might be alluding to in his question) to build scalable, robust, maintainable, and secure applications it is preferable to use n-tiers. A typical example might be to use a database, an application server, a web server, and finally the VoiceXML interpreter. The application server makes connections to the database subject to requests from the webserver, which in turn is called from the VoiceXML interpreter. This has a number of advantages:

1) Scalability: resource pooling (e.g. threads, database connections, preinstantiated objects etc) can be incorporated via the application server
2) Robustness: Each tier can be designed to be robust (best technologies, use of clustering for transparent failover etc)
3) Maintainability: Each tier can be maintained separately with well defined perimeters (in order of code fragility): presentation tier can be separated from business logic tier, which can be separated from database tier.
4) Security: Firewalls can be inserted at various locations with valuable data being 'hidden the best'

To use a database connection from say a SpeechObject means that n-tiers becomes 1-tier. At very least scalability will become a large issue as will points 2,3, and 4. Whilst I can see the appeal for quick results I think the 1-tier solution does not endure in the long term: There is a huge difference between a JVM running a VoiceXML interpreter executing a Java object and say a J2EE container using EJBs.


  ----- Original Message ----- 
  From: Hitesh.Seth@silverline.com 
  To: tehdy.draoui@polygones.com ; www-voice@w3.org 
  Sent: Friday, April 26, 2002 5:56 PM
  Subject: RE: Need to access DB in VoiceXML but not Server-side

  You can definitely achieve this in SpeechObjects. In fact if I remember correctly, Nuance V-Builder does have SpeechObjects which achieve this functionality. However, by definition "object" tag is implementation dependent. So you would loose portability there.

  Also, some vendors have have also incorporated that functionality by supporting component models such as ActiveX, Java Beans right from the Script.

  I would like to also understand what is the main driver for accessing the database from within a VoiceXML document?


  Best Regards

    -----Original Message----- 
    From: Tehdy Draoui 
    Sent: Fri 4/26/2002 11:13 AM 
    To: www-voice@w3.org 
    Subject: Need to access DB in VoiceXML but not Server-side

    Is there a solution to access a database within a VoiceXML Script? I now you can do it server-side, with ASP, PHP, JSP, etc. but i would like to write only VoiceXML in the script. 
    How does the <OBJECT> tag works? What about "speechobects"? Can i design my own object and call them throw the >object> tag? It would be nice if i could write such objects in Java, and implement JDBC to access a DB. 
    If someone is brilliant enough to find a solution ...
    Best regards,
Received on Friday, 26 April 2002 13:22:26 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 23:03:46 UTC