|
Networks & Infrastructure
Keeping Ethernet Simple - Telco Systems Interview Part Two
Telco Systems Builds Best Of Breed Solutions
by Sean Buckley
Manu Kaycee has seen his fare share of technologies for technology sake that have made service provider’s job more complicated over the years. Kaycee argues that Ethernet needs to avoid becoming too complex to manage and deploy like ATM. In part two of two-part interview with Manu Kaycee talks about the continued evolution of Ethernet as both a service and a transport mechanism.
TM: We all know that fiber is the ultimate medium to deliver Ethernet, but fiber only goes to a certain amount of buildings. Do you see a growing need to have devices that can accommodate both fiber and existing copper?
Kaycee: As Telco Systems, we were focused on fiber-fed services. What we have done is acquire technologies and do partnerships. We acquired Critical Telecom, which is a hardened Ethernet DSLAM that supports ADSL and it will evolve ADSL2. We have this strategy of providing solutions over fiber and copper with copper bonding via partnerships with other Ethernet DSL companies. These best of breed relationships where they would provide copper access and we provide fiber-based access. In this scenario, we would not only coexist, but also interoperate end-to-end.
TM: While we all know that Ethernet is a service, the other emerging trend that’s gaining momentum is using Ethernet as a transport mechanism—otherwise known as CET (carrier Ethernet transport). What’s your sense of that?
Kaycee: It’s interesting in that both companies (Telco Systems and Metrobility) had a different heritage, but we had one thing in common which was Ethernet. The first step is to provide Ethernet service over virtual Ethernet transport. Virtual Ethernet is based on IEEE standards or queue in queue and 802.1ad provider bridges. We also have plans in place to evolve HVPLS and T-MPLS all the way to the edge. This goes back the philosophy that we don’t dictate what carriers deploy in their metro core network; we’ll map into any existing or emerging standards-based implementation. We see a need for HVPLS, T-MPLS and also PBT. We will monitor where PBB-TE goes. From our experience, we find that once you get to a draft stage, you could start to implement the protocols or features.
TM: Ok, so you’re taking an agnostic approach to the PBT debate, but what would be the utility of these technologies?
Kaycee: We are seeing things with PBT to be used for inter-carrier handoff. Maybe two carriers have their own topology, but one does not want to expose that topology to the other carrier by rolling any dynamic protocols. People have been using T-MPLS—a point-to-point tunnel between carrier X and Y. We also see PBT as being useful in those applications.
TM: What concerns are carriers still have with the Ethernet evolution?
Kaycee: Ethernet unto itself is cost effective, flexible and simple. Given the fact that we need carrier class constructs with Ethernet, people have been adding things like embedded OAM, QoS and signaling. For a while, we we’re concerned it would become quite complex much like ATM started with the 53-byte prime number payload and it turned out that other things like PNNI became overly complex. As a result service providers would not deploy SVCs. Also, we were concerned that things would get too complex much like what happened with MPLS. With the advent of T-MPLS, PBT and other things, we are seeing that people are realizing that we should not make these networks overly complex.
More Information:
C5: ADVA Advances In Germany
T-Com Deal Cements Good 1Q 2007
Verizon’s Backhaul Bonanza
Cites Growing Demand For Alternative Solutions
C5: Huawei Pins Colours To PBT Mast
Chinese Vendor Pledges PBT Support for BT 21CN
C5: Siemens Bangs PBT Drum
Operator Interest Growing, Claims Vendor
|
