Did someone forgot to change the default Ethertype value on T-MPLS? it seems so…
According an article on NetworkWorld and ComputerWorld, there is a kind of “war” undergoing between IETF (The Internet Engineering Task Force) and ITU-T (International Telecommunication Union) started by allegations from IETF that current T-MPLS specification by ITU-T could lead to massive interoperability issues for service providers.
T-MPLS stands for Transport-MPLS and defines a suite of specifications and architectures in order to allow MPLS traffic to run over an Ethernet provider backbone. The problem seems to be on some shared codepoints (including EtherType, which is a field on Ethernet frame that indicates the protocol being transported) between T-MPLS and MPLS standards. Since they aren’t fully compatible (due to different technology implementation), according IETF, T-MPLS should use different EtherType in order to avoid any confusion in operational networks.
It is our opinion that the use of common Ethertypes for IETF MPLS and T-MPLS in the manner in which ITU-T SG 15 is currently progressing represents a mutual danger to both the Internet and the Transport network that will carry T-MPLS and this should not be advance
You can read the IETF open-letter here: Letter from IETF to ITU regarding T-MPLS.
On the other way, ITU-T argues that T-MPLS will only be used in the provider network (encapsulating MPLS traffic from client network), as such they would be deployed in disjoint networks leaving no space for protocol conflicts.
T-MPLS working group is undergoing some meetings in Stuttgard (Germany), on which this issue should also be discussed.
I really don’t get it, why didn’t ITU-T defined a new EtherType for T-MPLS in the fist place? anyway, this also sounds a little like politics… “you stolen my EtherType idea! I will not allow it! give it back to me!” (sorry, couldn’t resist!)
Cheers!
Recent Comments