I believe the misperception stems from the fact that the industry is stuck in a Telephony paradigm as the model for how business-to-business video should work. Using a Telephony paradigm, one would indeed need phone number authorities, directory service providers, exchanges, etc.
However, telephony in my opinion is the wrong paradigm. The better paradigm is Email: email technology allows you to reach anybody in the world without centralized directories, without email exchanges, without interoperability service providers, etc. just by virtue of knowing that person's email address. Emulating this paradigm, all the elements are in place today to enable end-to-end video communications anywhere in the world:
- A common network: The Internet itself is the public network for IP-based voice and video communications (just like the PSTN is the public network for telephony). No other “exchange networks” are required. Video technology is getting to the point where we can deliver high-quality video over the internet by dealing intelligently with packet loss, latency, jitter, etc. This doesn’t mean that customers might not decide to use “higher-quality” private networks to link to their partners, but this is not a requirement to enable any-to-any video communications today.
- Standards: if video wants to emulate Email as the paradigm, it needs to adopt Internet (i.e. IETF) standards rather than Telephony (i.e. ITU) standards. This means SIP, not H.323. It also means STUN, TURN, and ICE for end-to-end connections through firewalls.
- Directory: the Internet has a directory: it’s called the Domain Name System (DNS). The DNS is the central directory that directs users to the correct web sites, that routes email to the correct recipients, etc. No other centralized directory required. Similarly, using SIP standards, you can place video calls to any recipient across the Internet by just specifying the SIP URI of the person you're calling. Again, no other centralized directory required. Assuming organizations set up the appropriate records in the DNS system for their domain, the internet will use the DNS to determine where to route SIP-based voice and video calls for that domain.
- Firewall traversal: the Telephony world has introduced Session Border Controllers as the mechanism of choice to enable IP voice traffic to traverse firewalls. While SBCs would work equally well for video, there is an alternative: using standard SIP-based protocols such as STUN, TURN, and ICE, firewall traversal can be provided as a software-only solution (using firewall traversal servers in the DMZ or on the public internet) that avoid the need to deploy additional equipment and/or appliances. This significantly lowers the hurdle that one needs to get over to enable end-to-end business-to-business video.
In summary, using an email paradigm for video communications rather than a telephony paradigm, business-to-business video can be a reality today without requiring additional infrastructure or services.