Cloud computing is one of those hot trends that everybody just has to have a story for, no matter how farfetched and irrelevant it may truly be. That's not making it easy to even define what cloud computing is, or to differentiate it from something like software-as-a-service. The most respectable differentiator is that with cloud computing the service is some level of platform on which users run applications they supply. The platform can be very general-supporting virtually any application-or limited to applications written in a specific language or supporting a very specific mission (like the back-end processing for a web server). Can that platform then host UC and UCC?If we stay with the cloud-is-a-platform definition, then cloud-UC services are enterprise-owned UC services that are hosted in the cloud rather than SaaS-like cloud computing services. A good example is the Siemens Enterprise Communications announcement that their OpenScape applications have run (in proof-of-concept form) on Amazon's EC2 cloud computing platform. The open-source Asterisk PBX has already been ported to EC2, and Microsoft may be looking at UC as an application for its Azure cloud. It's clear that at least conceptually we could have a bunch of cloud UC solutions coming up this year.
There are two theoretical benefits to a UC-as-a-Service strategy; cost management of the UC tools and integration of multiple service-provider offerings into the UC framework. The former is a classic economy-of-scale justification, and the latter is a bit of what Google Voice shows; if you are "in" the network you can manage the relationship between multiple services more effectively and thus create a stronger UC story. If cloud UC is valuable, it has to offer either or both of these benefits.
The economy of scale question is pretty easy; cloud hosting is most likely valuable for smaller companies who have limited IT infrastructure. It's whether there is a special benefit to being in the cloud with UC applications that's harder to figure.
The Amazon EC2 cloud is a good example of a highly general platform-as-a-service. It runs Linux or Windows, supports any programming language supported on either one, and can be used to link to any TCP/IP port or API. That means that in theory you could write or assemble a full-featured UC/UCC system and host it in the cloud, obtaining any benefits in economy or availability that EC2 offered for your size of operation. But there are no special cloud tools to help the process along.
More "architected" clouds like Azure or Google's App Engine could be the opposite example. App Engine is more restrictive in terms of both what it will run (Java and Python applications only at this point) and how it looks to the end user (like a back-end web server that would normally support CGI calls from web pages). But even in its current form, App Engine provides an API to Google Mail, and in the future Google could extend this to support all of Google's services, including Google Apps and Google Voice. This would make a Google-hosted UC application potentially more powerful than an in-house or EC2-ported UC system.
The difference between an "architected" cloud and one like EC2 is the extent to which the cloud contains its own middleware and facilities to provide packaged services. These will inevitably make the cloud more specialized so it's most valuable in the specific areas where the incremental tools are focused. That makes me wonder whether somebody like Cisco (with WebEx) or Citrix (with GoToMeeting) might be thinking about a "UC/UCC Cloud" offering. There have been rumors of this in Cisco's case already. There's also a chance that the Oracle/Sun acquisition might be a launching pad for an Oracle cloud strategy that could well support UC/UCC applications.
Any general cloud computing strategy could become a cloud UC strategy, in fact. Add some libraries and APIs and you're ready. UC in the cloud, for a general cloud computing provider, could be just a matter of getting a SaaS relationship with a good UC application vendor. Siemens may hope Amazon will be interested in that approach; thus their "proof of concept" EC2 trial for OpenScape. In addition, cloud UC offerings could be packaged (like OpenScape) or just a set of UC-friendly middleware for users to exploit themselves with a kind of "roll-your-own-UC" development process.
UC/SaaS cloud applications may also be just what players like Verizon are thinking about in their emerging cloud computing strategies. Telcos have been trying to get the range of next-generation services and gain developer support for their own service APIs (BT has long had a program, and the telco-centric GSM Association has its own API activity). Why not jump to a cloud-based offering and get in on the buzz? The rumors are that telco cloud computing will focus on SaaS providers and SMBs, so that may be where hosted UC is heading. Telcos could provide a platform with access to all of the telco service APIs, and developers and software vendors could then build the UC tools from that framework and offer them to the SMB market.
The notion of cloud UC may well be a barometer of whether telcos are really serious about developer programs and cloud computing. If they fail to address this opportunity to merge the two, they either they are kidding us about their support for these new areas...or they're kidding themselves when they way they're looking for new monetization strategies.
