Comcast will be rolling out a DOCSIS 3.0 trial in Minnesota, with the top download speed 50 Mbps (5 Mbps upload). The service will be the first deployment of the faster cable modem standard, and the interest level will be high since DOCSIS 3.0 is the only architecture that can rival FTTH in performance. However, it’s interesting that Comcast is deploying in Qwest’s territory; the operator has only DSL in contrast to its eastern rival Verizon, whose FiOS can also offer 50 Mbps in some areas. We note that DOCSIS 3.0 is still a shared-media technology with less capacity available per user than FTTH can support, and the upload speed for cable is much lower than for FiOS (which now offers 20/20 symmetrical and 30/20 asymmetrical in some areas).
In what may well be a key announcement for the industry, Intel revealed that it had 25 partners working with it on portable Internet devices, or “mobile Internet devices” (MIDs) in the new-speak. These boxes will be designed for use while away from home or work but not mobile as the driver of a vehicle, and are larger than a cellphone but smaller than even the smallest laptops. Intel and others believe that the future of wireless non-voice services will lie in these devices, whose screens are large enough to deliver a credible viewing experience. Intel hopes these will jumpstart the WiMAX market, a key market sector for Intel and also one that has seen recent rumors of partnerships involving Intel, Sprint, Clearwire and Comcast.
The cable companies are taking the threat of FiOS seriously and promoting technologies that are more suitable for deep fiber and FTTH in their own plants. The activity is concentrated in what is called “RFoG” or RF over Glass, meaning mechanisms to perform the opto-electrical transformation from optical delivery of multicast RF to the CATV plant that already wires the homes. We are hearing that the idea is not to go with fiber to the home despite reports to the contrary, but rather to take more of a fiber-to-the-curb approach, wiring perhaps a dozen homes at the maximum into a CATV span off an RFoG fiber plant. Verizon has been looking at a similar notion of using a remote and MoCA to run cable into the home to reduce fiber provisioning costs in areas where the ARPU won’t justify true FTTH. We believe that there will be more and more outside plant “wiring” using a combination of fiber and coax, even among the carriers.
Comcast today announced it would change how it imposes traffic management constraints on Internet usage, discontinuing its focus on P2P applications and focusing instead on traffic management of “hogs” or users who create the largest amount of traffic. As a practical matter, the move will likely have the same impact when the cable spans are congested, since P2P is the most significant source of upload traffic, but the new move might also impact any users who act as servers or who generate significant video traffic. The FCC has asked the company to commit to a date for the change, in what we believe is a simple PR move on the part of the FCC. We do not believe the FCC would have ordered a change, and that the current move by Comcast is a response to a competitive campaign largely by Verizon.
Comcast and BitTorrent are in talks over how to make P2P exchange less a traffic management problem for cable networks. There has been little detail released on this but the rumor is that Comcast is seeking a “topology-aware” hashing tree so peer uploads from one user on a cable span would eliminate the span as a source for other uploads, seeking instead alternative systems on other spans or operators. The move is prompted not so much by the FCC’s actions, which are not likely to require Comcast to stop its traffic management, but by the fact that the problem has handed the RBOCs, and especially Verizon, a major competitive advantage.
David Yen, a kingpin in Sun’s revitalized microprocessor program, is leaving Sun for Juniper, where he will be an executive VP of Emerging Technologies, a move that certainly means that Juniper will be doing more original work with microprocessors and/or network processors. There has been a lot of new energy at Juniper in recent months, suggesting that the company is about to make some aggressive moves in the market to sustain its growth and take advantage of loss of market momentum by its mobile-exposed competitors.
Verizon Wireless is launching its “open handset” strategy, a move that may have more impact on cellular voice pricing than its cap plan did a short time ago. Under the new program, handset makers can certify against what Verizon says will be minimal requirements and then sell directly to consumers, who will have to sign up for Verizon service but would not be required to sign service contracts. We hear that the non-contract services will be pricier than contracts offered, but that contract prices sans phone will likely be better. This will create additional slide in wireless cost and, we believe, further reduce infrastructure investment by wireless operators until an alternative revenue model is validated. We also believe it will promote FMC as a means of securing loyalty that was previously cemented through handset deals.
Five unnamed mobile service providers have joined with ValueClick to launch a service that targets better demographic targeting of mobile ads. The mechanism used for ad selection is heavily linked to behavioral targeting, and it seems clear that the move is the long-awaited move by some mobile operators to gain some traction for themselves in the mobile advertising space. However, the ValueClick partnership doesn’t necessarily provide them the means of truly controlling their destiny, and we believe the mechanisms used in the trial (at least from what has been described) to be primitive. This comes as an analyst firm predicts that mobile search ad revenue will reach nearly $5 billion by 2013.
ECI is entering the Carrier Ethernet market, leveraging in part the MPLS tools that the company obtained from Laurel Networks when it acquired that company. This space is a critical one because it is the focus of some fundamental debates—is Ethernet a service framework or a metro architecture, and can it displace some or all of MPLS. We believe that Carrier Ethernet can be a service and metro architecture and can displace an important chunk of the MPLS opportunity, but only if providers understand how to build metro infrastructure from Ethernet and not just offer Ethernet services. ECI seems to be moving in the correct direction, but the details of its products are still too sparse to be sure.
AT&T is working through its bidding process for GPON-based FTTH for new developments, a move that takes the company further down the path Verizon has already taken. It is not known how much of the Verizon architecture AT&T will mimic, and in particular if it will use the same linear RF broadcast TV delivery program over its own channel-slot Microsoft TV approach. How this plays out will be of enormous importance to equipment vendors because the decision to use FTTH and PON in any form bends the budget decisively toward the access network and empowers players like Alcatel-Lucent who can bid there. The fact that GPON empowers Alcatel-Lucent has been the seed of the internal battle between the GPON group and the router group, who don’t want broadcast over fiber at all.