We've Moved! Please visit our new and improved forum over at our new portal: https://portal.plumvoice.com/hc/en-us/community/topics

submit question

Questions and answers about IVR programming for Plum DEV

Moderators: admin, support

Post Reply
aosipov
Posts: 3
Joined: Wed Aug 20, 2003 12:10 pm

submit question

Post by aosipov »

Hi,
I have been using an application that uses a record to get a message from the user and then submits it to my script that then does some post processing on the message and email it. Since the submit next location is a server-side script it sends voicexml in reply and then proceeds to do post processing. While the entire valid voicexml reply is sent, the DocumentParser::FetchDocument keeps waiting for the connection to terminate, and thus waits to process the voicexml that was sent in reply until the script is done.

This resulted in large messages that required more time to postprocess and sent as email to have a long delay in between the time of submit and the reply, depending on the network conditions, et al. What I have done is passed off post processing and email features to a separate section that is run every 15 minutes and post processes and emails anything that has not been done as far. Thus only requiring the file to be saved on the machine and virtually no wait is required. The problem with this now is that there is a 15 minute delay until messages get processed, and while I can reduce it to 10 or 5 minutes I cannot get real-time results with this process.

I have had this working for over a year now without encountering many problems but in the process of upgrading and adding more postprocessing capabilities (deleting silent messages and some other things) I would like to get this process to be real-time parallel process.

An idea I had to reduce network latency and get a quicker response time from the audio upload is to use the audio repository. However with the limitations on the features (upload and play only from what I can find) of the audio repository, and the size of the repository there is no efficient way to do this.

Something that would be great is a audio repository with multiple bins where certain bins had actions attached to them. Think triggers on tables. But that's for Plum to consider.

I'm sure everyone has encountered this problem before, any ideas and thoughts are welcome.

support
Posts: 3632
Joined: Mon Jun 02, 2003 3:47 pm
Location: Boston, MA
Contact:

IVR platform has no feature similar to triggers on tables

Post by support »

Hello,

Plum has no plans to implement an IVR feature similar to triggers on tables. The best way to accomplish what you are describing is to use the IVR audio recording repository. You can use the IVR repository to upload a recording, when you complete the upload you will receive a URL where this audio file can be accessed (see: Audio Repository Manual - Section 2.2). At this point you can post the location of the file (with any other information you need) to your IVR application server. After writing this information to a file/database you can either fork a new script or setup a script that runs periodically to check the file/database. The script can then pull down the file for processing and send off an email.

Hope This Helps!

Plum Support

Post Reply