Hi Mark, that was a fine critique you did of Gilder's report. I agree with your assessment of the flavor of Gilder's reports in general, as well as your points regarding TERN, which I, along with others like Bernard Levy, have also addressed in the past. His reports are more like a sell-side analyst report than a technology report in the true sense of the word. Since I don't have anything nice to say about the quality of their content I'll leave it at that.
It would be interesting to have some TERN investors email TERN Investor Relations and ask the status of S-CDMA relative to DOCSIS advanced PHY, and, more importantly, to state that ONLY a reply from a company officer (e.g. CTO) is acceptable (IR can always take the easy way out and just reply that they don't know).
In your post #1592 you retracted your statements regarding capacity of TERN S-CDMA headends. I don't believe you need to do so. I believe what's needed is more architectural detail and terminology clarifications from this person at Shaw, and what exactly they include in their headend cloud, e.g. how many CMTS's. Perhaps they oversubscribe their nodes to a greater degree that most. 1000 "subscribable connections" would be roughly assuming a 10% penetration rate, which is about the nationwide average currently. The “roughly 10% of 1000” figure is based on 128 64kbps channels subscribED per cmts, which would be 128 users assuming each user would be willing to pay 40 bucks a month for a single 64kbps up/down channel. Proportionately less if users buy more than one "unit," say 2 for example, so they at least have something noticeably better than a 56k modem, i.e 128k up/dn in the case of 2 units.
More details on the breakdown of the data stream for TERN's S-CDMA system is here:
Message 11348504
This is all taken from their patents, so anyone can check it. There is no way the math works out with anything close to 1000 subscribed lines per CMTS. |
