Intel announced a couple of days ago the volume production of its IEEE 802.16 standard WiMax modem chipset (known as Rosedale), geared towards equipment vendors that are targeting the broadband wireless market.  The chip, now officially called the Intel PRO/Wireless 5116 broadband interface device, is fabbed externally and utilizes a programmable architecture.  Intel is only delivering the silicon, with the vendors being able to implement their own software up to the MAC.  Intel expects most vendors will use 5 MHz for channel bandwidth thereby yielding roughly 10 Mbps of IP traffic. 

The chips will be used in CPE gear and eventually in wireless base stations.  The advent of the 802.16e standard (expected to be ratified later in 2005) will lead to better power efficiencies and the silicon will be targeted for desktops by 2007.  As of today, Intel is producing the Rosedale chips externally and as the uptake increases, the production is expected to shift in house.  Despite the fact that 2005 volumes are expected to remain relatively low, there is a lot of hope that with 802.16e, reduction in costs, higher volumes and economies of scale can be achieved. 

Despite the fact that this is still probably the bottom of the first or top of the second inning in terms of WiMax (to use the baseball analogy), I believe that this is still a very important milestone for Intel, because the company has finally delivered the next step in its vision.  Also significant the fact that Intel currently has 12 hardware customers and 100 operators planning to sample their Rosedale chips (the manufacturer customer list includes the likes of Airspan, Alvarion, Aperto Networks, Axxcelera Broadband Wireless, Gemtek, Huawei, Proxim, Redline Communications, Siemens Mobile, SR Telecom and ZTE).  To add a further dose of realism (and silence some skeptics), a total of 15 carriers worldwide will be announcing their intentions to trial the Intel 5116 WiMax chips.  The list includes: AT&T, Altitude Telecom, BT, Brasil Telecom, ETB, Iberlanda, Millicom, Qwest, Sify, Speakeasy, South Africa Telkom, Telmex, Tower Stream and UHT. 

How long before the Centrino model takes over WiMax as well?  Well, it will be a while, as power efficiencies of 802.16e will be needed.  So this second phase push probably will not happen until 2007, at which point the technology will start getting moved into notebooks.

I will definitely be posting more on the Skype story (got a couple of separate threads there worthy of separate postings).  But first, an insight on the analyst panel at VON Canada, which counted with Jon Arnold (a friend and ex-colleague of mine at Frost&Sullivan), Jeffrey Fan (director and equity research analyst at UBS here in Toronto) and yours truly.  I understand that due to force majeure reasons, Daniel Berninger had to drop out, and I certainly hope that everything will be OK with his kid.

The discussion centered around VoIP in three key areas: enterprise, service provider and the MSOs (the focus of most of Jeff's research).  Jon asked a few thought provoking questions, particularly about the rollout of fiber-to-the-x to various homes.  One thing I definitely brought up is that in contrast to their US counterparts, Canadian carriers such as Bell and MTS have very good loop length profiles.  In fact, according to Bell, 85 to 90 percent of its customer dwellings are within 1,200 meters of the node (that's about 3,600 feet, well within the 5 kft. distance to ensure a bandwidth of between 15-20 Mbps for a FTTN deployment with ADSL2+ or VDSL2 as a last mile technology - note Bell is going with a VDSL solution).  That is way better than SBC, for instance, which has over 57 percent of its loop lengths exceeding 6 kilofeet.  So FTTN with VDSL2/ADSL2+ seems to be the right approach for Canadian ILECs such as Bell to bring the triple play to all of their customers as quickly as possible.

Jeff mentioned the triple play efforts of some MSOs, including Shaw and Rogers.  Interestingly enough, nobody from the audience asked any of us whether or not the $55 that Shaw is charging per month will be able to deliver the uptake in triple play services.  Obviously, since Bell and Shaw own the only two DBS licenses in Canada, that leaves Telus a bit vulnerable on the video side (until the latter launches its own IP-TV solution) - and until that point in time, maybe Shaw will just continue to keep its price at that level.

Jon also made a few comments about the uptake of VoIP in the enterprise, where thus far, the CPE approach seems to be winning.  The slow uptake of IP Centrex has surprised some, but given the dynamics of the market, and the typical lenghts of contracts offered, in hindsight, the modest ramp-up was to be expceted. 

All in all, it was an interesting session - it would have been nice to have Daniel in there, as well as Cody Willard (perhaps we can have him in a future session). 

Note: Mark Evans wrote about Merrill Lynch analyst Glenn Campbell's estimate that Shaw is signing up about 1,000 customers a week for cable telephony in Calgary, but the question still remains whether Shaw can reach 20% penetration of cable subscribers within 5 years.