In the simplest sense, unified communications is a centralization of messaging and calling mechanisms for users into a single controllable framework. Interestingly, Google lists only three entries if you ask for a "Define:" of unified communications, which would suggest that everybody agrees on what it is. It's more likely that most people aren't bothering to define it, and the definitions that are listed are rather vague. Let's try to do better.
We can start by saying that there are two integration options available to create UC. You can integrate communications at the common element level by using a single appliance that natively supports multiple forms of message communications. A PC that can make calls, send emails, and send SMS or IM is an example of common-element integration. You can also integrate communications at the common presence level by presuming that multiple communications elements exist, but that these elements link to a common single picture of the user's state and policies-the presence.
I would contend that to be truly unified, a UC system has to provide the following:
1. A presence manager that lives somewhere and represents the user's status and policies for all the communications channels that user supports (uses or can be reached on). Some sense of common or collective service presence is the essence of UC.
2. A set of tools to contact that manager to build and maintain policies and status information. These tools exploit the presence knowledge that's available.
3. A set of communications channels, linked to the presence manager, that the user can communicate through.
Most desktop-driven UC is a common-element approach. The essential notion is that the desktop system is the place where most activity occurs and so this common element becomes the UC focus, providing a presence that manages the channels of communication that terminate there. If there are other channels of communication (like a cellphone) those have to be synchronized with the presence maintained by the UC system in some way, or they're outside the UC framework and it's not really "unified" any more. Many, perhaps even most, desktop UC products have no really effective way to synchronize presence with other communications channels, so they would fail the three-point test above.
In contrast, most of the "hosted" UC offerings appear to be based on the notion of common presence. There is a kind of "virtual user" that resides somewhere and that acts as the central point of policy control for all forms of communication. This presence then controls how the actual connection to the user might be made by integrating with each of the available communications channels. The model can work with the presence hosting taking place on the premises or by a provider.
The fact that UC must have some set of policies that control how communications channels are used resolves one of the definitional debates in the UC space; the difference between UC and "unified messaging". This debate has even found its way into Wikipedia, which asserts that "unified messaging" is a kind of retrieval management. In the real world, UC policies must include all of the possible handling options for messages, including real-time delivery, store for pickup, store-and-alert, etc. In fact, it's the flexibility of these policies that create the value of UC and that also create the challenges in meeting all three of the requirements listed above in a given UC implementation.
You can see from all of this that the number one UC challenge today is the tools that link a "presence" with multiple communications channels so as to permit effective presence management and message/call delivery control across all the communications channels a user might employ. In fact, the difficulty in creating truly unified presence connections for all types of communication is so large that it is the major value proposition the larger vendors use for an "all-on-us" approach. If a single provider offers every form of communications, the integration is clearly going to be a lot easier. Verizon's iobi is an example of a hosted common-presence UC system of this single-provider type, and hosting of presence for UC is in the business plans of many industry giants, including Apple (MobileMe) and (so rumors say) Cisco. Right now, network-hosted presence systems have the advantage in that they have better access to communications channels, but that could change.
One mechanism that might facilitate the process of connecting multiple communications channels to a presence hosted on the customer premises is the new trend toward open APIs for service provider networks. There are a number of operators who expose APIs that let you query presence and even make calls under program control. This would let a UC system work with network services more effectively, and would thus make it more likely that all the important communications channels, at least, were included under the UC umbrella.
Mobile telephony has long been a difficult thing to integrate with UC, and that may be the area where APIs are developing the most rapidly. The GSM Association (GSMA) has launched a program called "Third Party Access" that is building APIs and has even created a developer program around it. The process is in its infancy at the moment, but even the concept is encouraging for those who need mobile phones tightly linked with UC.
The availability of APIs to synchronize wireless, wireline, and even SMS communications with premises-hosted common presence would be a boon for UC vendors like Microsoft but also a boon for the UC market overall. Most enterprise customers prefer their UC be premises-hosted, and so service provider API programs could be the factor that finally makes it possible to fully realize UC goals even with on-site presence hosting and management. That would certainly boost UC adoption in the enterprise.
