Greg Mumford

"The bandwidth explosion is just barely starting"

February 8, 2001

14 Min Read
Greg Mumford

3549.jpgPop quiz:

  • Mike Myers

  • Mark Messier

  • Greg Mumford

    All Canadians. All famous. But only one of them has had the same employer for 29 years.

    Can you name that Canadian?

    If you said Greg Mumford, you're right. (If you said either of the other two, you might be more at home at the site of our good friends at tele.com, available here: http://www.teledotdotdash.com/dot/tele/dotcom).For almost three decades, Greg has had only one employer: The north-of-the-border company now known as Nortel Networks Corp. (NYSE/Toronto: NT).

    In an optical age when most executives' only loyalty is to the size of their options packages, Greg's resumé reads like the Great White North, itself: a bit monotonous but rather refreshing.

    But flying in the face of conventional wisdom has become a Mumford hallmark -- especially in the field of optical networking. Now president of Nortel's Optical Internet division, Mumford had a big hand in directing the company's original decision to build 10-Gbit/s equipment, a gamble that eventually helped Nortel win its current market-share lead in the optical arena.

    Now Mumford says he's ready to place his next Nortel technology bets -- in areas that he believes will enable the crafty Canucks to once again extend their lead over rivals such as Alcatel SA (NYSE: ALA; Paris: CGEP:PA), Ciena Corp. (Nasdaq: CIEN), Cisco Systems Inc. (Nasdaq: CSCO), Corvis Corp. (Nasdaq: CORV), Lucent Technologies Inc. (NYSE: LU), and Sycamore Networks Inc. (Nasdaq: SCMR).

    What's on the Mumford menu? More speed! He says Nortel will start shipping 40-Gbit/s and even 80-Gbit/s products shortly.

    But that's just for starters. Mumford is also looking to drive Nortel's optical product line "up the stack" (the OSI-model networking stack, that is) with new and smarter "bandwidth management" devices that are intended to help carriers provision and bill for the IP services that run over optical networks. These include a mondo combo IP router and optical switch -- enabling Nortel to compete in a market now dominated by startups like Caspian Networks, Chiaro Networks, and Village Networks Inc..

    It's a smart move (as is to be expected from someone whom Nortel colleagues describe as "smart" and a "mover"). Nortel has achieved huge success selling products that allow carriers to reduce their network costs by increasing optical capacity. But in order to grow, it needs to take its product line to the next level: providing carriers with a second generation of products that lets them generate new revenues. And as the number one supplier of optical transport equipment, it's obviously in a great position to sell these products into existing accounts.

    But while it may be a logical move, it's also a risky one. For Mumford's strategy to succeed, Nortel must first tame IP. Its previous attempts to do so have met with mixed results: Some were hilarious. Others, simply sad.

    The world won't have to wait long to see whether Mumford is onto a winner. He's set an aggressive rollout schedule for the new products, with both 40- and 80-Gbit/s hardware and the new optical switch router due out this year.

    We caught up with the Mumford Man recently on the phone. He talked about:

  • How to stay at the same company for 29 years.Light Reading: How long have you been at Nortel? Can you briefly recap your career?

    Greg Mumford: I've had about five careers at Nortel... so, sit down, Paul [Kapustka, Light Reading's editor at large]. It's rare to be in the same place for 29 years these days. I joined Nortel in 1971, straight out of my masters program [Mumford has a Master of Applied Sciences degree in electrical engineering from the University of British Columbia]. I joined a project at the time that was called "exploratory digital switching." Out of that project came the DMS-100 [a now-legendary central office switch]. I was the ninth guy in to that exploratory development program, trying to figure out how to make a digital voice switch work.

    This was at a time when the conventional wisdom was: "Digital's OK in the trunk circuits, but it'll never fly as a local switch because it'd just be too expensive," and so on. And somebody in Nortel -- not me at the time, that's for sure -- had the vision that digital was going to allow convergence, and digital was going to payoff, and we're going digital.

    Light Reading: Seems like you've confronted that curve several times.

    Mumford: As a corporation, we're trying to confront it every day. So that was a really interesting time. I had no idea, at the time, what a piece of history we were working on.

    Later, in the first half of the '80s, I managed a team developing optical transmission systems. When I arrived they were just finishing up an optical system that was working at 45 Mbit/s [everyone on call laughs]. From 1985 to '86, I was in corporate headquarters working on strategic planning. That was my transition from R&D into the business side. In 1987 I was product manager for switching -- I was managing the investment portfolio for all the switching businesses.

    Then, in 1991, we started the new era of optical transmission when we announced our Sonet line -- and I've been here [in Nortel's optical division] in one form or another ever since.

    Light Reading: Here's a quick pronunciation question: Is it "Jiggabit" or "Gigabit"?

    Mumford: Gigabit [pronounces word with a hard "G" sound].

    Light Reading: Why do some people say "Jiggabit?"

    Mumford: I don't know. I wasn't aware that they did.

    Light Reading: We've heard some people say it that way at conferences. Now we're embarrassed for them.

    Light Reading: Let's move on. How important was the decision to develop 10-gigabit technology to Nortel?

    Greg Mumford: Clearly it was very important. Maybe at the time we didn't know how important. But the product's also important as an indicator of what our thinking was, and is. I mean, in those days the transmission standard was Sonet and SDH [synchronous digital hierarchy]. And we knew with that standard, with that protocol, you could put all kinds of different traffic on one network, and therefore you could collapse networks.

    We knew that the levels of data were starting to increase, and we could see bandwidth costs going down dramatically -- dramatically, dramatically -- and therefore, bandwidth would get used for lots more things. We knew those fundamentals. We also had a really good command of technology, so we had views of how technology would mature, so we picked an intercept, and placed a bet -- and then stayed with the bet. We stayed with the bet when people said, "Who would ever need it?" We stayed with the bet when people said, "It's impossible." We stayed with bet when people said, "It might be possible, but it's too expensive...," and we just plain stayed with the bet.

    And all those conditions apply today. Just on a different level with different aspects. I think the bandwidth explosion is just barely starting.

    Light Reading: One thought we've heard tossed around lately is that optical networking's toughest challenge isn't in the physics, but in more mundane things, like manufacturing -- especially of components.

    Mumford: Oh yes... that's a very good topic. Let's go back to the 10-gigabit decision: Making it was just part of it. Our success depended on being able to go with it, to take full advantage of the opportunity. Maybe the most important thing that Nortel Networks did successfully through that period was scale the manufacturing, and scale the installation and scale the delivery. And, with our supply partners, scale the whole supply chain. Two years ago, that was a key issue.

    In our components business, we rely on partners. With some of the key components, there's still a huge amount of manually intensive work being done in alignment, and in packaging. But the focus on automating [those processes] is really ramping up; some [automated manufacturing] solutions are starting to appear. In the components end of our business, I think we're going to see really dramatic cost reductions and output capability come through that automation -- that's the era that's really just started, and we'll see more investment put on it, from VC firms, as well as from corporations.

    Beyond that, there's a whole 'nother era to come -- that's really the era of integration, with fewer components. That's all ahead of us yet.

    Light Reading: Is it important to have all the components in-house, Greg?

    Mumford: It is important to have influence, strong influence, over the things that matter. And there's more than one way to get influence. In the past, we've never had all the components in-house. But we have had, for example, the really high-performance lasers in-house. The reason is, they actually enable the system -- without them, we couldn't get the system enabled. So, we've always done that because we had to.

    However, we have no interest in having components in-house that are better done by somebody else. We're walking that balance, we've walked it for years, and we continue to walk it.

    Light Reading: Where is the next bet being placed?

    Greg Mumford: There are really two bets being placed. One is: Scale the network where it's required; and 40 gig is only part of that. We announced that we are going towards both 40- and 80-gigabit technology at [the] Telecomm '99 [tradeshow] in Geneva. At those bit rates you could put about 6 terabits on a fiber, compared to the 1.6 terabits we're putting there today. And we said we expect to put the technology into the network by the end of 2001. That development's going well.

    The other one is bandwidth management, to equip the service providers with services that they can use to increase their value, and their revenue. For example: Managed wavelength services, and broadband services -- by which I mean bandwidth on demand. That whole piece is starting to take shape now.

    We're making the same kind of bet as we made on 10 gig; We are betting on the intercept of the demand and the technology. We set out to develop a system and the components simultaneously, and they're just coming together now. So we're ending up with a bandwidth manager that will switch wavelengths, or sub-wavelengths, down to 50 megabytes of payload. And our building block is 3.6 terabits, so it's a 4-terabit building block. That product is called the "OPTera Connect HDX," and it's part of a whole family of bandwidth management devices.

    We've got a team focused on that, to put the product in trials before midyear, and put it in the network before the end of the year. So this conviction about bandwidth, and scale, to make networks cost-effective and valuable, to further unleash bandwidth economics, is something we're following very aggressively.

    Light Reading: Is bandwidth management maybe harder to understand, but more important, than raw speed?

    Mumford: It'll enable the transition from static bandwidth to dynamic bandwidth, and that's really fundamental. Right now, people are selling service by the hour, but you can't scale that. The whole dynamic of giving us an equivalent of a dialtone for bandwidth is what we're focused on. And it's much more than facilitating mesh restoration. It's real, true, dynamic bandwidth.

    And, we've chosen to go with an open approach, supporting the standards for automatic switched transport. It's really important for all of us [vendors] to use the same standard, so our customers can make a network out of multivendor products. So we haven't branded the [OPTera Connect HDX] yet. It too is coming out around midyear, we'll brand it when we get closer.

    Light Reading: Let's switch gears a bit and talk about people. You have a 29-year career at Nortel. Have you ever been tempted to go to a startup and be the lead guy? And how do you keep people in your organization from the same temptations?

    Greg Mumford: Of course I've been tempted. I was tempted, if you want to use that term, to leave Nortel long before we had this kind of really hot venture capital market. Over the years, I've always known where I stand, and what my value is in the corporation, and what my value is outside the corporation. And I think that's a very healthy thing to do. In recent years, there's certainly lots of interest from lots of folks, in terms of my experience and so on.

    But I've elected to stay with Nortel Networks because we're really building something here that's really important and really challenging and really matters. And to take it up a level -- we're building the high-performance Internet. As important as optical is, it's much more than just optical. And it's very sophisticated and very satisfying and very exciting.

    And so I stay here, for the same reasons that any person stays with a company. Because in this day and age, we're all volunteers. There's lots of opportunity out there. I stay because I have really exciting things to do that are fun, I feel I'm making a difference, and because the values of this corporation are compatible with my personal values.

    That's why I stay, and I think that's why others stay. As leaders, if we fail to actually create that lineup, then people will not stay. It maybe seems a bit simplified, but I think it's as simple as that. A leadership role is really important now, because you don't manage volunteers -- that's an oxymoron. I just came from a leadership meeting, where we agreed that anybody who thinks the people on their teams are anything other than volunteers is probably not taking appropriate leadership actions.

    Light Reading: That sounds like you're trying to keep big-company problems from creeping in.

    Greg Mumford: The attitude is to instill the entrepreneurial spirit, wherever we can. The fact is, when you're setting out to build the high-performance Internet, a lot of the value comes from making things complementary, making them work together. Creating value is much greater than the box. Out of necessity, you do need to link more people together, and if you're not careful you get the bureaucracy and negatives.

    What we're focused on is putting together integrated teams, chartering them, turning them loose, and letting them roll. And bringing them together, where we need to, to create value. It's an art -- we learn something every day, about how to do it better. When we don't do it well, it's not good.

    Light Reading: What about motorcycles? Is there still time to ride motorcycles? (See The Top Ten Movers and Shakers in Optical Networking.)

    Mumford: Not in this weather! I don't know who you got that from -- I do have a Harley, a Harley Road King, and I get on it whenever I can and try to get into a pattern of taking a couple weeks in the summer to get away. It's really hard to hear my cell phone on the motorcycle, which is great. And it'd be really dangerous to answer it, too!

    Light Reading: What is it like up there in Canada? Seems like it's quite an optical hotbed. What is it like to live up there... not too cold, eh?

    Mumford: The secret to winter, is to have winter sports. I actually have fallen off the wagon on winter sports, but in the 1970s, I cross-country skied every week. I actually used to get really put out when it got too warm...

    Light Reading: Yes, waxing with klister is a pain...

    Mumford: Say, at 24 degrees Fahrenheit or lower, it's great. Above 28 degrees, it's tough.

    To address your other question, this area is not becoming a hotbed, it already is. There are lots of startups, lots of Nortel, and there's a relationship between those two. It's a great place for a family, a great place for outdoors.

    Light Reading: How old are you?

    Mumford: Do I have to tell you? I'm 54.

    Light Reading: Let's go back to the top 10 list, where you have to share billing with Don Smith. How much fun is that?

    Mumford: Don is a great guy... and I enjoy working with him, so I don't have a problem with that. It'd be nice to share the Number One spot!

    Light Reading: Maybe we can set up a battle of the network stars between you guys and Vinod.

    Ann Fuller (Nortel PR): Can it be a motorcycle race?

    Mumford: Vinod would probably make it a photography contest. He's really good at photography, and I don't think he would agree to a motorcycle race.

    Greg Mumford was interviewed by Paul Kapustka, editor at large, Light Reading http://www.lightreading.com

Subscribe and receive the latest news from the industry.
Join 62,000+ members. Yes it's completely free.

You May Also Like