PA-ComNet Meeting Jan. 8, 2003

The meeting began at 7:35 AM in the Terman library.  The attendees introduced themselves, plus they signed the attendance sheet.  Bob Harrington began by referring to the reports on his and Manuel's visit to Provo, Utah, in December that were posted on the PA-Fibernet site. They were circulated to the other participants at Provo before posting so they have been verified as accurate.

Provo is similar in many ways to Palo Alto, larger, younger, with significant university population.  They are ahead of Palo Alto in some ways.  They are using World Wide Packets equipment and Emtelle blown fiber for their FTTH project. AT&T hired over 20 lobbyists who succeeded  in getting a state law passed that requires municipal utilities be wholesale retail providers.  This caused problems because the retailers all want exclusive franchises. Provo would have preferred being a retailer.  They have gotten retailers to provide almost all services with the iProvo name prominent and service provider names not at all prominent.

Bob showed the extensive program list.  There are 48 music channels.  The basic channel is $10.95, 77 channels is $32.95 with no set top box needed.  Digital setup is $49.95 for 180 channels.  Service is provided by AT&T.  VOD ranges from $.95 to $3.95, but it really isn't offered yet.  BYU will offer learning on demand which is essentially unlimited.  Pricing is 10-15% less than AT&T.  There are 3 levels of service provided by a local ISP who was not identified.  Lowest is $14.95 for 256 Kbps. Highest data rate is 1-3 mbps symmetrical for $39.95.  Higher bandwidths are possible but would cost more.  It is software managed MAC address to MAC address.

Manuel Topete noted that Palo Alto needs to work with proven, feasible technology, not something unique, or something no one else does at a price no one wants to pay. Many technologies are being considered. 

In Provo the system works extremely well, the total experience was great, but there are many questions.  The IP video works great, almost too well to be true.  We still have many questions.  Only 50 of 65 homes have been installed so far.  The network isn't heavily used yet.  They don't yet have a business plan but they hired Neil Shaw to write one.  He is the same person that is going to write the Palo Alto plan.  They have a video plan, but don't have a full plan implemented yet.  Pricing is unclear because WWP won't quote costs yet.  There is no video-on-demand yet, just multicast.

Reed Majors of Minerva Networks noted they make the Provo headend equipment.  He gave a business plan to Provo previously.

Paul Venturella, telecom manager at Provo said they used IP video to be futureproof.  They use 3 fibers/home now, but will go to 1 fiber/home later.

Discussion followed on technical capability, routing, possible problems with bandwidth usage.  Current customer premise equipment runs at 100 Mbps.  They could go to 1 Gbps, but think it would be too expensive, perhaps $2500.  WWP pricing is a secret.  Whenever they were asked to give costs they changed the subject.

Jerry Scharf noted 100 Mbps is plenty fast enough for video so 1 Gbps isn't needed.

Dan Swineheart noted bandwidth of the 1st level aggregation switch shouldn't be any problem either.

Manuel noted multicasting IP video saves bandwidth.

Reed Majors noted there are over 9 million VOD enabled homes in the US and there are statistics on them.  In rural America 80% of viewers watch the 4 most popular channels.  There is a FTTH system with over 25000 VOD users in Milan that use an all-Cisco network.  It has been run successfully for 3 years by Fastweb and is cash-flow positive.

Manuel said there is a FTTH Council made up of manufacturers and users who recently held a conference in New Orleans.  There were 3 simultaneous meetings every day for 3 days.  Question was asked of a panel with people from Wave 7, Alcatel, WWP and others if there is anywhere that VOD over IP is used successfully, and they said no.  In fact Cisco worked with Milan to provide it.

The system is up in Provo and works, but there are less than 60 users so far so it isn't stressed.  The companies involved are good and the kind Palo Alto can work with. 

Reed Majors said Sure West in Roseville is working with FTTH and has 5000 customers.  They are good resources.  They also have hard cost data.

Manuel noted Palo Alto is behind other areas, but we can use other's experience to make our efforts better.  We can use analog if it works better and is cheaper, or we can go IP digital video if it works better.  IP video works technically at small scale based on Provo.  We want to see what works best.  When we tell people 1 Gbps is possible they want to know if we're kidding.

Jerry Scharf said the capability of IP has been proven and costs are being reduced   IP won't solve all problems rapidly.  There is lots of investment and effort in the field and in 10 to 15 years it will be a very minor incremental cost.  The groups that are investing now are those like us who are not invested in legacy systems.  IP has limitations but it also has great potential.

Manuel said they download a movie with Movielink in order to view Chicago.  Downloading all 790 mbps took 20 minutes from LA, but it can be viewed during download if desired.  Panel TV was 42" but the view was about 6".  Limit for the download was Internet speed, not local speed.  If the movies were stored locally the download would be far faster. Super high bandwidth is more useful for an Intranet than for the Internet. Dan Swineheart agreed, and said the Intranet is where the exciting new stuff will happen.
Palo Alto could serve the entire community from 3 headends with 1200 channels using IP video when it becomes fully developed. Manufacturers are flexible at this point to grow the business.  If we ask for it they'll provide it, and then see how well it works.

Len Weisberg asked about QAM.  What in the transmission method isn't ordinary analog and isn't IP packets? 

Radio only works in analog.  Video on present systems is analog plus digital.  Present systems can deliver analog + digital with a compression ratio of up to 16:1, although picture quality is far better with 6 or 8:1compression.  Comcast provides VOD from local storage in Philadelphia and several other cities, so it's practical.

Jeff Hoel said the important thing about the encoding method is how many digital bits per Hz and the signal to noise ratio.

Reed Majors said a digital head end system can be cash flow positive in 1 year, ignoring the cost of the fiber ring.

Use of blown fiber was questioned.  Manuel said the cost was prohibitive 2.5 years ago, but it's gotten better, easier to work with and insert tap tubules into a tube while on a telephone pole than to do fusion splices.

Provo estimated cost was $5 million for the fiber loop plus 250 home trial.  They have more fiber to run than we do.  Worldwide Packets wouldn't give cost figures, only would respond to RFPs.  This was not helpful because we can't issue an RFP yet because we haven't gotten approval yet.  Another concern is the technology will be skewed so that only 1 or a very few bidders are possible.  That isn't the intention of CPAU.  The RFP won't be tailored to limit bidders or suppliers.  The existing system using Wave7 was chosen because they had a very good technical specification and offered the best business test for the trial.   It's a PON system with point-to-point fiber, so in a way it allows study of both PON and point-to-point in 1 system.

Meeting adjourned at 8:55 AM.

Bob Moss