Cloud Series 6: Voice Cloud Types
This post is part of a series:
Inching Toward the Clouds
At Astricon 2009, there were numerous sessions on using the Amazon EC2 Cloud to host an Asterisk Server. It makes a lot of sense for development ($0.10/hour), but actual production isn’t so clear. Amazon charges separately for processor, storage and i/o.
In a test development, all of these costs are pretty low. When the developer stops for the day, just shut down the instance.
Production phone systems aren’t so simple. They need to be on all the time, though they are drasitically underutilized after hours. Plus a cloud based system with SIP trunks would require i/o for each phone conversation. The economics become questionable at least at current rates.
Clouds offer benefit around processing power and storage, but not really over i/o. But the model is very young. If you bring in your own private SIP trunks, that should bypass the normal i/o rates – right? If you connect your sites over a private MPLS network – then that too should bypass the normal i/o rates.
At VoiceCon, Siemens did another demonstration of their Amazon based solution “OpenScape” which still isn’t available. They seem to be onto something. Not only are they embracing cloud computing, but social networking as well. Their demo used keyphrases in Twitter to set location and availability information such as “at lunch”. Limited practicality, but that wasn’t the point. I intend to watch Siemens closely as they are consistently showing some very innovative concepts. But realistically the odds are against them. No European company has every been dominant in US Telecom and versa Visa. I know, technically they are not a European company – Gores Group is N. American, but I get the sense HQ is more Munich than Andover.
Recently, both Avaya and Mitel moved closer to the cloud too. Avaya was first to release a product – Aura – for virtual computing; specifically XEN/Citrix. Mitel then announced that their actual Call Director (and application server products) are virtual ready; specifically VMware’s vSphere 4.0 (though not released yet).
So basically we have three big market leaders with virtual products for virtual environments (Avaya Aura is shipping, but no virtual call control). It is the first of many required steps needed for voice to really work in the cloud. 2010 will be a transitional year for voice clouds as the designs and kinks get worked out.
So more on this coming in future posts.
Back to how the cloud might impact voice… one area I’ve neglected (so far) is videoconferencing. Sagee of Radvision made some excellent points about this via a flat world blogging conversation.
Please take a moment to read his post Cloudy, With A Good Chance of Video Conferencing.
One point he makes that really struck me was his discussion of ownership. Rather than hosted vs PBX, he talks about CDs vs Radio – a great analogy.
I recently lived the music issue with my music collection trying to make it accessible to the house audio system. I solved the problem with the Logitech Squeezebox (see my review here), but after I did it – I find I never listen to it. I almost always play Pandora or similar Internet Radio stations which offer a slightly more varied collection of music. The Squeezebox made it equally as easy to play either. The point is I set out to solve the wrong problem, but fortunately it was the same solution as the real problem. It isn’t a question of ownership, it is a question of communications.
Perhaps it is because of this series, but I seem to be hearing a lot more conversations on hosted voice. Hosted voice is quietly becoming a central topic – and too many people – including myself, may be underestimating tomorrow’s hosted solutions by pigeon holing it into yesterday’s hosted voice offerings.