Thomas Howe

I was in a meeting with one of our most exciting clients yesterday, and we were planning the deployment of the Enterprise Voice Mashup we are creating for them. As we discussed the options for testing, cut-over and redundancy, it occurred to me how many advantages Voice Mashups have over traditional voice applications in terms of deployment strategies. Due to their blending of VoIP and Web Architectures, and their jettisoning of call state, Voice Mashups are brilliant in the real world.

Here’s an example of a clear Voice Mashup advantage taken from the project we’re working on: simple, incremental deployment. Imagine that you are a Web portal company, such as Google, with a new home page to test out. Given the state of web balancers, DNS record schemes and the like, you would never deploy the page to your entire audience. In whatever ratio that makes sense, ten percent of visitors, one percent, one hundred a minute or once a day, the new web page could be delivered to the user, and result tested. In the event some sort of failure, the web user would simply try again, and because getting the page in the first place was rare, wouldn’t see it again. The clever Internet engineer will tie the ratio of the web page delivery with the chances of failure. I have heard it said, and I firmly believe it to be true, that because cell phones and VoIP relaxed five nines reliability they created space for true voice innovation.

Since Voice Mashups are built on this very same technology, we can apply these very same principles. Our particular project is replacing a help desk task with a voice script. When someone calls with this problem today, a human being answers the phone. As we deploy the application, our first step is to peel off a single call, and see how it goes. If it fails, they’ll call back. After single calls seem to be solid, we could take one every ten minutes, and see how it stands up, eventually letting the entire audience use the script. This moves forward to revisions in the future. Version 2.0 of our application can have the very same deployment strategy, and we could go backwards as fast we deployed 1.0.

The advantages don’t stop there, and the ability of Voice Mashups to manage service delivery risk is a true innovation for voice. Good to be here to see it.