Genuity research takes a look at Nortel's optical group
The ever-brilliant David Hodgson, Genuity’s top-ranked hardware analyst, took a look at some of the neat things going on at Nortel (NT:TSX) these days in a research note published this morning:
“Nortel held a teach-in on the 40 Gig optical product yesterday, and focused on the factors that are leading Nortel management to be extremely optimistic that the product will be able to win a bigger share of the optical transport pie going forward – the Verizon win for the pan-European ultra-long haul network where Nortel was not the incumbent is a sign that Nortel may be able to take market share from competitors. In attendance were Metro Ethernet Networks head Philippe Morin as well as R&D leader Dino DiPerna.
Year-to-date, Nortel’s Metro Ethernet segment has generated revenue of $1.096 billion, down 4% YoY amid stiff pricing competition, and EBT has been essentially break-even. Over the past almost four years, EBT for Metro Ethernet has been decidedly negative, with some improvement over the past couple of years. Clearly, Nortel needs a hit which can deliver improved margins to the division and allow it to begin contributing to corporate profitability. We believe the company is banking on the combination of 40Gbps optical + PBT to jumpstart this positive margin improvement. Time will tell, but we were, to be frank, quite impressed by the company’s 40G presentation, as we will outline. While this will not provide an immediate bump to the profitability of the optical group, this (with somewhat similar timing to PBT) will begin contributing more materially in the late 2008/2009 timeframe.
With 12.2% market share, Nortel is a #2 player in a number of optical markets – with the overall market roughly $10-12 billion in size. An emerging opportunity within this area is 40 Gig optical, which is expected to be a $200 million market in 2008, growing to over $2 billion annually in 2010. In other words, this market is of tremendous interest to carriers that will be facing increasing bandwidth congestion primarily due to the influence of IP video and peer-to-peer traffic. 40 Gig is currently mostly in long-haul DWDM applications, but will eventually be evenly split with metro. Nortel sees most of the upgrades starting in 2008, with North America, CALA, and Europe seeing faster deployments. Asia will lag the market somewhat.
Nortel’s traction in 40 Gig is exemplified by four current customer wins (Rascom, Verizon, and two unannounced), with lab trials with four major accounts in Q4/07, a live field trial in Q1/08, and 10 40 Gig deployments “primed” by Q2/08.
The main thrust for Nortel’s new 40 Gig product (available March 2008) is the ability to easily upgrade from legacy solutions to cutting-edge 40 Gig DWDM systems. It is believed that less than half of the optical fibers currently deployed can support 40 Gig speeds, but the “secret sauce” for Nortel’s solution is the ability to run 40 Gig on top of these legacy optical installations. Replacing fiber would be a huge cost for service providers, but the good news for Nortel’s solution is that they have figured out a way to get better than 10-Gig performance transmitting at 40 Gig.
A service provider can simply plug in two new 40 Gig circuit packs (the eDC40 and a 40 Gig client card) on an existing DWDM platform (such as Nortel’s OME 6500) – with little re-engineering required. At the end of the day, Nortel believes that it’ll be able to provide better traffic quality at longer fiber lengths, translating into capex, opex, and time to market advantages for customers. On top of the capex and opex savings from an equipment perspective, when Verizon announced its selection of Nortel for its ULH network in Europe, the prevalence of OME 6500 platforms was highlighted as an important reason why it chose Nortel – assuring operational familiarity for its staff.
What enables this special signal handling for the eDC40?
Nortel is leveraging its experience in the wireless domain to implement a technology called dual polarization QPSK, which allows multiple streams of data to be transmitted over a channel. However, the unique aspect of eDC40 is the way that it handles optical signals on the receive side, a “coherent digital receiver”. This receive-side handling provides for compensation for two impacts called polarization dispersion and chromatic dispersion. The impact of both dispersions is that the original signal gets distorted – the receive-side handling provides integration signal compensation, eliminating the need for separate DCFs (dispersion compensation filters) as well as additional signal regenerators. The bottom-line here is that the introduction of new network elements to support 40 Gig would introduce additional cost as well as re-engineering – relatively speaking, the NT solution appears to be pain-free. NT believes that competitor solutions will suffer from signal degradation not only when transmitting over long distances but also when going through reconfigurable add-drop multiplexers (ROADMs), which allow service providers to flexibly use their optical capacity.
Progress with standards
The big question mark is whether some carriers decide to bypass 40 Gig completely and move directly to a 100 Gig-based format.
The challenge from a standards perspective is that the standardization for 100-Gig won’t be until the end of 2008, with ratification expected by the end of 2009. Rates being considered at 130 Gbps (3×40 Gig) and 112 Gbps (support 100 GigE). The choice of DP QPSK would be a positive for Nortel as it would position the company well from an IP perspective and would fit very well with Nortel’s 40 Gig solution.
The positive from an NT perspective is that they foresee that their 40 Gig solution will scale to 100 Gig as well – therefore, a slower transition to 100 Gig would benefit NT given their new low-cost 40 Gig solution. A faster transition would mean that the competition would need some from of receive side signal compensation (dispersion would be more severe at 100 Gbps) – a feature where NT already has in its current 40 Gig solution.
Nortel’s technology approach for 40 Gig appears to be quite positive and carrier-friendly from a capex and opex perspective. The “plug-in-play” upgrade capability is also an innovative feature, which we would foresee having a positive impact from a time-to-market perspective. How easy will it be for competitors to catch up? The lead may be one or two years for competitors to develop a coherent receiver, a situation that is helped by NT’s IP position with respect to some of the algorithms defined by Nortel. At the end of the day, NT believes that the result of more than 4 years of work will be challenging for competitors to replicate given that designing such hardware is hard, requiring very high speed signal processing. A positive from a Nortel perspective is that while competitors may catch up, NT is already trying to move down the cost curve for its eDC40 product. The other positive is that NT’s solution allows customers to upgrade slowly to 40 Gig (mixing and matching 10 Gig and 40 Gig cards) as service providers need to transport more data – potentially mitigating fears from service providers of over investing in capital equipment. This view is balanced by the underpinning of growth in this whole market, namely near-exponential growth in customer demand for high-bandwidth services. Should the consumer side develop slowly than hoped, a natural conclusion would be that service providers will roll out 40 Gig over a longer timescale.”