One of the recent standards flaps in networking is the debate over MPLS-TP versus the older T-MPLS. Standards types can be incredibly obtuse at best, and in arguments they’re likely to fall into jargon and acronyms and lose communication with the real world.
The publicity on this particular issue has ranged from treating it as a face-off between the Internet Engineering Task Force and the International Telecommunication Union to debating the propriety of “reconsidering” T-MPLS based on activity on the last day of a standards meeting after many delegates had gone.
What we think is really going on here is a fight over the role of IP in the metro network. Video content has already convincingly shifted carrier investment to the metro, where it focuses on connecting broadband users (for money) and supporting content delivery network services (for money). Now we’re adding mobile backhaul (more money) to the mix. In the “core,” the low revenue per bit on Internet peering virtually assures minimal growth, so IP has to gain a lot of traction in the metro to continue to grow. But metro is where Carrier Ethernet has been strong, so there’s a battle to make IP capable of supporting the kind of aggregation-to-a-service-point model that Ethernet supports.
T-MPLS was a step to create a transport vision of MPLS to counter what was called PBT (properly PBB-TE) from the Ethernet world. Its OAM standards are Ethernet-like based on the presumption that it was competing with Ethernet, that Ethernet OAM was further advanced at the time and that interacting with metro Ethernet was a likely requirement. MPLS-TP is more “IP-like” — its OAM practices are consistent with those of MPLS and IP pseudowires overall.
It seems to us that the current debate is flawed in that both sides are dodging the technical questions in the name of standards-body politics. What OAM practices would be applied in a metro area? Will Ethernet continue to be deployed on a large scale for backhaul? Are there benefits, particularly in the evolution of mobile backhaul and content delivery to mobile devices, to support true IP connectivity even within a metro network? With what would an MPLS-TP trunk likely connect? We’d like to see a more organized discussion on the real issues here. At the core, we’re afraid it’s just Ethernet versus IP and the old PBT wars all over again.
