Carriers Converge on Ethernet

ATLANTA – Light Reading Live! – Metro Ethernet is the future of telecom, and carriers everywhere are studying what they have to do to get there. The perils and benefits of the journey were scrutinized in depth at Light Reading's day-long seminar, "Ethernet Services: A Carrier Class," held today at the Ritz Carlton Hotel in downtown Atlanta.

The invitation-only event drew over 200 registrants, of which over 85 percent represented service providers, enterprise customers, and public sector agencies.

A full complement of international incumbents turned out, including AT&T Corp. (NYSE: T), Bell Canada (NYSE/Toronto: BCE), BellSouth Corp. (NYSE: BLS), British Telecommunications plc (BT) (NYSE: BTY; London: BTA), MCI (Nasdaq: MCIT), SBC Communications Inc. (NYSE: SBC), and Sprint Corp. (NYSE: FON), to name just a few.

Also on hand were members of the Ethernet in the First Mile Alliance (EFMA) and Metro Ethernet Forum (MEF), both of which are engaged in full-throttle efforts to solve a range of problems facing service providers -- and standardize those solutions.

There's no doubt Ethernet is the key services technology of the future, said Light Reading cofounder and CEO Steve Saunders in his introductory remarks: "We're here to find out the details of its ascendancy."

At the top of his mind was how to make Ethernet truly "carrier-class" by addressing issues of scaleability, security, interoperability, and manageability. As one questioner put it during one of the day's several panels: "[Quality of service] for Ethernet seems almost a retro fit for an aging queen..."

To keep Ethernet from being enthroned in the enterprise but powerless to rule the WAN, where its benefits are clearly demonstrable, carriers need to address a range of economic, technical, and regulatory issues.

As panelists chewed over these topics, specific themes cropped up repeatedly throughout the day:

  • Metro Ethernet Changes. Moving ahead will force everyone to check their assumptions. Consider this: "There's no reason to run a lot of Layer 1 traffic through a router," declared panelist Eric Voit, architect for Metro Ethernet Solutions at Cisco Systems Inc. (Nasdaq: CSCO). (That's right, Cisco!) Shortly thereafter, he mentioned that core routers were the costliest elements of building metro Ethernet, thanks to the cost of translating Layer 2 traffic to Layer 3.

    After taking a gleeful ribbing from co-panelist Peter Green, engineering team leader for optical broadband services at Nortel Networks Corp. (NYSE/Toronto: NT), Voit said: "Cisco sells switches, too, not just routers!"

  • Ethernet Economics. "If enterprises think of Ethernet as cheaper, how can carriers make their margins?" said one IXC representative during a break. Panelists said there's no easy answer: Opex savings will be key, but no one's yet been able to nail down specifics (though the MEF says it's working on a report to bolster the case, which it will release within the next few weeks).

  • Quality of Service. When it comes to service rollouts, carriers are thinking of how they can deploy Ethernet's capabilities to create better quality of service than they have with ATM. MPLS is seen as one way to build on QOS. The MEF is working on others.

  • Management Is Key. Quality of service and overall economic viability of Ethernet services depends on having the right management systems, panelists said. "No one builds a single-vendor network.... It's an integration job between multiple vendors' management systems," said Colin Evans, director of marketing and business development at Native Networks Ltd. and also chair of the MEF events committee. Others maintained that management could help reduce both capex and opex and offer more revenue-generating options to customers.

  • These were just some of the issues covered in today's talks. Of course, service providers of all kinds face challenges in making the transition from legacy services to Ethernet.

    Different types of carriers will have different approaches. One representative from a North American IXC, for instance, said during a break that the IXCs, not the ILECs, are in a more favorable position to start offering Ethernet as an alternative to legacy services, because they don't have as much invested in local private lines as RBOCs.

    Another RBOC attendee, questioned separately, saw his key opportunity as virtual LAN services in metro areas, where RBOCs have the upper hand in terms of customer clout -- right now, anyway.

    Others say applications will be key. Charles Kenmore, a board member of TechNet, a regulatory pressure group, said in his speech that telcos will need to understand the entertainment value of applications such as video.

    In the end, there were no easy answers. In a session led by Bob Mandeville, president of Iometrix Inc., Ethernet project director at Light Reading, and facilitator of a massive, 28-vendor network being demonstrated at this show, he made clear that the biggest challenge would be for carriers to understand the different elements involved in metro Ethernet in order to make decisions that suited their specific networks and circumstances. "There is no one answer," he said.

    — Mary Jander, Senior Editor, Light Reading

    squeegee 12/4/2012 | 11:58:39 PM
    re: Carriers Converge on Ethernet There's no doubt Ethernet is the key services technology of the future, said Light Reading cofounder and CEO Steve Saunders in his introductory remarks: "We're here to find out the details of its ascendancy."

    At the top of his mind was how to make Ethernet truly "carrier-class" by addressing issues of scaleability, security, interoperability, and manageability.

    Yipes, though with some growing pains, has been doing this - Ethernet in and between Metro Areas -for the last 3 1/2 years.

    They provide a layer-2 service over an all-optical network (low latancy), use 802.1q VLAN for security, offer instant scalabilty (for most services) and have a carrier-class NOC and web enabled tools (for customer) management.

    QOS is addressed by subscribing the network at 1:1 (no need to prioritize if there is no contention).

    The service is solid and the customer base is growing - primarily with forward thinking enterprises.

    eyesright 12/4/2012 | 11:58:38 PM
    re: Carriers Converge on Ethernet Understand what the crowd in Redmond, Washington is doing with telephony applications.
    Consultant 12/4/2012 | 11:58:35 PM
    re: Carriers Converge on Ethernet I am sick and tired of promising technologies being hyped and Light Reading has been hyping this technology for years. And the reality is that 98% of local circuit sales are still TDM.

    And there are four reasons for that.

    1. Ethernet requires fiber to the premise and less than 5% of buildings have fiber. It is the cost of installing fiber that is the main cost of broadband, not equipment.

    2. Metro Ethernet equipment is not carrier grade, and as a former Telseon exec noted, the more Metro Ethernet is upgraded, the more complicated and costly it becomes. VLANs don't have the security offered by SONET nor does Metro Ethernet have the restoration capabilities of SONET.

    3. Next-Gen SONET has effectively matched dumb Ethernet equipment in capex cost and operational expense.

    4. The big industry myth is that local circuits are expensive because of SONET. Wrong. Local circuits are largely expensive because they enjoy 70% to 80% gross margins.

    A. Evidence: every CLEC enjoys gross margins in excess of 70% on their fiber-based circuits.

    B. UNE rates are an approximation of cost plus a fair return on capital. UNE rates on a DS3 in a densly populated area are about $500 per month.

    C. The Bells have done big deals with wireless carriers where the latter pay about $100 on DS1 local loops.

    y2k 12/4/2012 | 11:58:34 PM
    re: Carriers Converge on Ethernet I think it is important to distinguish between transport and service. This is obviously an exaggeration but one can think of SONET as a transport technology while Ethernet is to provide service. May be instead of thinking how they compete with one another, one should think about how they complement each other.

    Hype or not, there is definitely a lot of real interest in Ethernet. And one doesnGăÍt need to have optical fibers to the building to have Ethernet. We have Ethernet today, in 10BaseT and 100BaseT which are copper, both at home and in the office.

    The main reason that the carriers are interested in Ethernet is that beyond DSL, T1 and ISDN, the next logical service that they can sell is Ethernet. With a Layer-2 switch, one can aggregate Ethernet traffics all the way to 10-gig and map directly onto OC-192 for metro core and long-haul transport.

    With UNE, there are companies out there developing Găúreverse-MUXGăą technology so that the IXCGăÍs can lease inexpensive DS3 from RBOCGăÍs and sell gigabit Ethernet services to the end customers. If this turns out to be a commercial success, it is another example on how Ethernet and SONET will coexist.

    myresearch 12/4/2012 | 11:58:27 PM
    re: Carriers Converge on Ethernet Can someone explain in 1-2 lines what do you mean by Metro Ethernet?

    Are we talking about native Ethernet bridging, Q-in-Q, Ethernet over MPLS, VPLS, Ethernet over SONET/GFP, E-PON, G-PON etc?

    In most of these cases, does thenetwork has anything to do with the Ethernet except that Ethernet is a payload. If you call Ethernet over SONET as Metro Ethernet, can I call IP over SONET, Metro IP?

    Consultant 12/4/2012 | 11:58:22 PM
    re: Carriers Converge on Ethernet Metro Ethernet really refers to Ethernet as a transport mechanism in the Last Mile and over the fiber rings connecting the central offices and telecom hotels.

    Consultant 12/4/2012 | 11:58:22 PM
    re: Carriers Converge on Ethernet I expect to see Ethernet as an interface to a Next-Gen SONET transport layer.

    But we need to keep perspective.

    Data does not appear to swamp voice on today's networks despite all the hype to the contrary. Indeed, most statistics on the relative importance of the two are based on such flimsy data (I was an analyst at AT&T and know about the pitfalls of telecom data) that it is almost worthless. I probably put far more zeros and ones each day over my voice connection than I do my cable modem.
    Secondly, voice still generates almost 90% of the industry's revenues. See the FCC 2001 Industry revenue report.

    Ethernet is just a nice interface. No one network fits all sizes despite strenuous effort to the contrary. I suspect we will continue to see a multiplicity of networks and protocols.

    BobbyMax 12/4/2012 | 11:58:20 PM
    re: Carriers Converge on Ethernet It is very revealing that metro ethernet equipmenrt vendors are not providing true differences between Ethernet and ATM/ SONET/SDH based nets.

    The ethernet vendors are also not explaining how: congestion control, cngestion avoidance, bandwidth xcontrol. QoS routing, QoS signaling and QoS policy management functions are performed in an equivalent manner in the metro ethernet.

    This non-disclosure borders deception. The sales people are not in a position to explainn these things, yet the vendors like Cisco try to explain away things. Lack of documentation prevents service providers to cmpare and contrast
    Sign In