RPR in the Spotlight
The session included presentations from four proponents of RPR, a protocol designed to provide networks based on Ethernet and IP with the protection and manageability of Sonet. Included was Robert Love, chairman of the Resilient Packet Ring Alliance, an industry group geared to promoting the work of the 802.17 working group of Institute of Electrical and Electronics Engineers Inc. (IEEE), which is developing an RPR standard. The alliance claims over 90 members and plans to have its first draft proposal -- for consideration by the IEEE group -- finished by January 2002.
Also on board were Robert Redford, VP of marketing in the Public Carrier IP Group at Cisco Systems Inc. (Nasdaq: CSCO); Gary Southwell, director of technical marketing at Nortel Networks Corp. (NYSE/Toronto: NT); and Nader Vijeh, senior VP of R&D and CTO at Lantern Communications Inc.
Each of the four explained the protocol from a different angle. And when it came time for the Q&A, they were inundated with questions. "Why do you believe that RPR interfaces will be inexpensive?" wrote one attendee on a question card. Others wanted to know: "Is 'double the bandwidth' really true?" "Is RPR too incremental?" "The world seems to be moving away from shared access... Why do we want to use RPR?"
The presenters didn't appear surprised. And they did their best to defend the technology, as RPR continues to generate debate.
RPR is said by proponents to be the fastest, most efficient, and cheapest way to bring Ethernet to metropolitan area networks.
"RPR makes use of existing rings, supports voice revenues, and carries Sonet and Ethernet end to end," says Love. "It will be better than Ethernet, Sonet, or extensions to either."
RPR opponents say it's too much for too little. "It brings back token ring mechanisms and shared access," says Nan Chen, director of product marketing at Atrica Inc.. "It's too complex."
Chen, who was in the audience for part of today's presentation, also is president of the recently formed Metro Ethernet Forum (see Metro Ethernet Forum Meets), and he says that group already enjoys the support of more than 50 companies that believe Ethernet can be made metro-ready through the use of methods such as MPLS (multiprotocol label switching) and TDM circuit emulation. After making clear he's speaking with his Atrica hat on, "RPR is overkill," he says, even while he concedes it may be technologically superior to Ethernet in many ways.
Another sticking point of RPR is simple politics. Cisco has long promoted its own implementation of RPR called Dynamic Packet Transport (DPT), to be the basis for the standard. Indeed, Cisco's recent purchase of AuroraNetics was interpreted by many as an attempt by Cisco to dominate the RPR effort (see Cisco Acquisition Causes RPR Stink ).
Cisco does nothing to counter this perception. In today's presentation, Robert Redford repeatedly referred to RPR as DPT, and mentioned several times that Cisco had invented it and would continue to call it DPT until it became an official standard. "We did invent this a number of years ago, and we have 13,000 ports of DPT in use worldwide," he said.
This provoked eye-rolling and seat-shifting from Love and the other panelists, but no one took Redford to task.
Although scores of questions were sent to the podium, there was time for just a few. Still, unvoiced emotion ran high. One attendee grabbed the moderator's arm on the way out: "Why didn't you ask them who's going to compete with Cisco if they already have 13,000 ports out there? You let them off too easy."
— Mary Jander, Senior Editor, Light Reading