In case I've been too subtle - let me make it clear.

Blog Post created by whamblin on Nov 5, 2009

Open federation - the interconnection of different enterprises for real-time communications services - is extremely valuable. It takes the productivity gains enterprises get from collaboration services, and amplifies them by making them work with their business partners. Wideband voice, video, conferencing, voicemail and telephony features all work better when they work between businesses.


Achieving this federation isn't easy. It requires tackling some hard problems - phone number routing, security, quality of service and troubleshooting. Many attempts have been made, but none have yet to take off.


The first attempt was built right into SIP itself. SIP was designed from the ground up to work inter-domain and inter-enterprise. As long as users are identified by email-style SIP URLs (like sip:joe@example.com), SIP can can utilize the same kind of technology used by email to connect together different enterprises. Unfortunately, SIP's domain-based federation hasn't really taken off. Most VoIP deployments (though not all) continue to use phone numbers, and this is likely to persist given the huge installed base of the PSTN. Secondly, SIP's built-in federation techniques don't address the security issues inherent in federation. Most enterprises have been fearful to just open firewall pinholes for incoming SIP traffic. SIP's built-in federation doesn't address the QoS or troubleshooting issues either. Given all of this, open domain-based federation is not widely used.


Another technology which was developed to address this problem is ENUM. ENUM had a clever idea - the public Domain Name System (DNS) would be populated with phone number to domain name mappings. Anyone could take a phone number - say, +1 (972) 555-1234, look it up in the DNS, and get back the domain name of the organization that owns the number. The problem with ENUM is that it required government and telco cooperation to populate the DNS records. This has been years and years in the making, with little progress so far. Indeed, there is no motivation at all for the telecom service providers to populate the entries. Finally, ENUM tackled just the routing part of the federation problem.


Finally, there have been attempts at solving this problem through a variety of private federation and peering networks. These networks are typically closed. They use a variety of technologies to solve the phone number routing problem. Some of them use private ENUM databases, some of them use SIP redirection, and all of them rely on some kind of centrally provisioned database that is run by the federation provider. Some of these federation services run over private IP networks, which helps address the QoS and security issues. As a consequence, these federation networks are typically limited in scope and size.


The book is not closed, however. When you have a problem this important, innovative solutions are certain to appear.