Federation: The Vision and the Pipe Dream
The idea of "federation exchanges" makes a lot of sense, but can it be realized in an environment that changes as rapidly as communications today?
One of the toughest nuts to crack when it comes to advancing enterprise communications has to do with federation: The need for multiple systems--built from different vendors' technology, under the control of different public and private operators--to communicate with each other so that any end user can reach any other end user and leverage the full functionality of multimedia communications systems.
Dave Michels' latest post contains a really good description of a new approach to solving this problem. Dave did a Q&A with Avaya CTO David Chavez, and Chavez described a concept he called "Federated Exchange:"
"A Federated Exchange is a service that securely associates sessions in one domain to sessions in other domains. To make this easy and natural, it provides interworking between SIP dialects and codecs. Additionally, a federated exchange provides a means for controlling reach and visibility of its subscribed users. To address privacy issues, it is inherently opt-in and individually-focused as opposed to enterprise focused. To increase its relevance to the subscribers, it allows them to create additional characteristics about themselves that could be useful to the parties with whom they interact. For example, a user may want to include their home address. When they move and wish to change it, the federated exchange will securely update every party that the user previously defined as an appropriate recipient for that characteristic. Those parties could be businesses or people. Presence can operate in much the same way: they could allow their presence to be shared with their friends and the bank but not with the HOA [homeowners' association] or another tangentially connected acquaintance."
I think this is a very elegant description of a solution to the problem, so I want to engage with it in some detail here. There’s a lot in his description that shows how you could create effective multimedia public/private federation. But Chavez's description also shows within it some of the problems that the industry would face in trying to translate this vision—or one like it—into real products and services supported by functioning markets.
Let's break down Chavez's points. Something that "provides interworking between SIP dialects and codecs" sounds a lot like a session border controller (SBC), which is a market that's really taken off. So far so good. Then he starts adding in features that, by all means, are critical to making such a service do what we'd like for it to do. This includes "a means for controlling reach and visibility of its subscribed users" and "allow[ing] them to create additional characteristics about themselves that could be useful to the parties with whom they interact." Essentially, this adds presence and social networking to the Federation Exchange, and integrates these capabilities with back-end systems/directories/databases that maintain this information.
That's a really compelling vision, and I think you could get several of the SBC vendors to say that this is exactly what their products are on the road to providing.
The concern, or the obstacle, I think, is how quickly this can metastasize into a grand vision or God Box that simply becomes unwieldy both from a user standpoint and a market standpoint. Also, if everybody starts building their own version of a Federation Exchange, then pretty soon you almost need another layer of abstraction to connect all of the existing Federation Exchanges. And this is more of a risk with today's technology than it was with long-distance telephony or email interworking, because there are so many vectors and facets to account for in a multimedia, multi-network-owner world. Also, the previous generations' relatively limited universe of carriers or ISPs was better able to set and enforce de facto standards than is today's multiplicity of diverse enterprises.
And one more thing: "To address privacy issues, it is inherently opt-in and individually-focused as opposed to enterprise focused." By now, there probably aren't many enterprise managers who would argue with that--in the era of BYOD, seeing the user as an individual is preferable to seeing her as part of an enterprise. And yet, if this vision were being crafted two years ago, how many enterprise managers would have said that?
The point is about the larger world in which enterprise technology lives: In that larger world, things are already changing faster than enterprises can accommodate. The grander the vision, the more likely it is to be rendered obsolete by events that take place while the vision is being fleshed out, specified, agreed to, designed and built.
I really like David Chavez's vision for federation. I also think it could wind up being a pipe dream.