Sorry I haven't replied, but I was busy gaining 10 lbs. Geez, these holidays are dangerous!
I heard Copper Mountain is going to develop a scaled down version of their CopperEdge 200 9U format box, and Diamond Lane has a "PizzaBox" project code name to do the same thing. All are targeted at the Multi-Tenant Building (MTB) market.
you want to share that bandwidth among clients in a building, why not just feed a router with the backhaul bandwidth and connect customers with ethernet, Tut,or whatever?
There are three market segments: Commercial, Hotel and Apartment.
For Commercial MTBs, 5% to 25% of the tenants change every year. These numbers can be checked via the public office REITs. This "tenant changing" points to the need for service to be turned on and off, hopefully remotely. It also points to the need for "service size" to be changeable to fit the new tenant's needs. With Ethernet or other shared medium tech, 1) a snooper can read the next door neighbor's emails - which happened to a SF law firm- 2) Everyone is treated the same, which means aggressive users will slow down the neighbor's access - see what happened to AtHome in Fremont, CA two weeks ago - SJ Mercury, and 3) the ISP loses premium pricing strategies.
You may raise your eyebrows at 3), but everytime 3) occurs, you get lower costs on the low end of service, but the low end of service increases in quality, because moneybags paying for the premium pricing pays the way. This model exists for the airline industry, brokerage, etc.
So, the idea is not just to share bandwidth, but to price the bandwidth according to usage. With Tut's box, you can use a bandwidth manager before the Ethernet Switch, and there is a cludgy way to meter bandwidth using the router, but it isn't easy to do, and is easy to mess up. Not a good solution for a large network operation. And you want large because that's where company's make economy of scale, so they can lower costs, and charge less.
In large Commercial buildings (Class A), like Embarcadero Centers, Sears Tower, Empire State, supposedly you have hundreds of potential clients. But they are all served by different ISPs. Covad has a model to resell DSLAM service. So Covad can convince Qwest, or Level3 to pull fiber to those buildings, then convince the ISPs to replace their T1/ISDN service to the clients with DSL service.
The problem with that arrangement is, Qwest and Level 3 don't really need Covad to be the middleman in the long run. DSLAM (my opinion) has a bandwidth aggregation problem that is caused by the ATM required at both ends. (example, the fractional T1 our company is using is QoS at 128k, but when the ISP backend if free, we can burst to T1, which is REALLY nice. ATM delivery is fixed at the QoS, so I can't get 128K burst, but am relegated to 128kbps fixed).
Hotels have people moving in and out all the time. These boxes have to be turned off remotely. Everyone can get 56K from their POP, so offering 64K is worthless. So, what kind of bandwidth do you pull to serve the Hyatt in NY? Its got 400 rooms or so, which at 64K QoS, maxes out at 25.6Megs or a fractional T3. (occupancy rate 95%) All to get equivalent service quality of a dialup POP. Ooh aah. Point is, you gotta provide burstability and also real bandwidth aggregation. ATM is too rigid for IP. IP needs to burst, because user behavior requires large amounts of data in a short amount of time, then lots of dead time as the data is adsorbed.
Apartments, at least Class A, require secure point to multi-point solutions. If I'm a broker living in NY, I don't want my unknown neighbor in the building snooping my email messages. Yes, get PGP, and all that, but PGP is an extra hassle. And don't expect brokers to understand technology. Also remote provisioning is a must, because of the tenant churn.
If it's a really a big building, wouldn't you find a serving topology perhaps with multiple routers/ATM switches that is cheaper than a DSLAM/xDSL serving arrangement?
Copper Mountain, and Diamond Lane's offerings will reduce the chassis cost from $40k to about $10k, allowing MTB deployment in Class B buildings as well as in Class A buildings. DSL allows exising copper wire to be used, so fiber does not have to be installed in the building ($150K-$2M in CLass A, $20K-$150k in Class B). However, you need experienced Telecomm guys to do the punch down in the basement and testing of the circuits office to office (room to room for apt and hotel). So, no, the cost of install for DSL tech in MTBs is better than ethernet and ATM (fiber). (personal note: I have friends in www.acucomm.com and their "pizzabox" is available today. I get this info from them)
Residential single unit complexes represent a more complex problem. Where is the aggregation point? Some clusters CO DSLAMs are more economical. Some Satellite or Wireless (lower density). Some with Fiber to the Neighborhood (FTTN), should consider a Pizza Box or 9U implementation of the DSL tech. Then the DLC lines need to be directly connected to another patch panel connected to a RADSL based DSLAM/DSL server.
The problem here again is serving all the flavors of G.Lite, RADSL, xDSL, provisioning the different price sensitive residential customers, and figuring out who is responsible for the copper wire when it goes down (Covad's problem or Telco's?).
Concentric's upgrade is puzzling to me, because they are competing with Northpoint on CO real estate. Both are trying to sell DSLAM service to ISPs. I looked at the model, and they are really screwing ISPs, and delivering bad service to the end user to boot. Oops, this is a long email, but briefly - ISPs make $60/customer they sign up, and Covad charges $900/mo to resell their service, so the ISP needs to sign up 17 customers to break even. Northpoint, who uses Copper Mountain, is trying to implement the MTB model when the scaled down 200 comes out in March. So, Businesses will soon benefit. But they are also grabbing as much CO space as they can (I think), so they can prevent other ISP/CLECs from offering the same service - kinda like hub control by airlines in strategic airports. This way, they can control the pricing in a certain geography.
So, I really don't know if that's Northpoints strategy, but from afar, it seems like a reasonable story. Concentric is a lot like Northpoint, so I'd see a big bidding war for the CO space soon, so I'd rather buy SBC and the other 2 ? RBOCs, who will benefit.
Whew!
BL |
