Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Microcap & Penny Stocks : Globalstar Telecommunications Limited GSAT -- Ignore unavailable to you. Want to Upgrade?

To: Rocket Scientist who wrote (15693)	8/13/2000 12:24:13 PM
From: Thomas	Respond to of 29987

Hi RS, Couldn't G* be used for the (low bandwidth) back channel and a high bandwidth, lower cost GEO used for the high-speed forward channel? I would think that would mean less expensive GEO receive only dishes, with G* as the slick back channel. That way the application would benefit from the G* technology without consuming all of G* precious capacity. I have often wondered whether this configuration would be the best form of portable broadband access. Sure, there are a few extra hops involved, using two bent pipes, but the lower cost of the G* transmit equipment has got to go a long way towards offsetting that inconvenience. Cheers, Thomas

To: Rocket Scientist who wrote (15693)	8/13/2000 10:40:38 PM
From: finite_time	Read Replies (1) \| Respond to of 29987

Voice and Data The splitting out of voice and data may have several advantages. I think the main one is that it is easier estimate and control loading and traffic issues. Probably easier with like traffic on the same channel. In an earlier post here on SI (thought it was yours RS but haven't tracked it down again) it was stated that G* was considering dedicating one of the 1.2Mhz channels to HDR. As to the data vs. voice debate. It really depends on many different factors. It is interesting to note that voice traffic can not share BW other than the shifting vocoder (see end of post) whereas data traffic can share the same BW by just shifting it in time. It is certainly possible that a monthly fee charged for unlimited data usage on average could bring in more than a voice user at 50mins/mo. Warning: The following is pretty simplistic and there are NUMEROUS other factors involved. Network and communications theory is NOT my area of speciality. I would appreciate any corrections/comments to my simplistic analysis which follows: Voice traffic uses say 9600bps (not quite correct see end of post). That would imply that on a 1.25MHz channel you could carry a maximum of about 130 (not quite this simple but it will illustrate the idea). Now to get a maximum number of subscribers that you can sell to you have to know a bit about loading estimates of when people will and won't be using the system. To do this we will use a number that tells us average user usage intensity typically expressed in the unitless Erlangs. This number is typically ~100 (Erlangs) for normal cell stations. To calculate the maximum subscribers we then multiply the maximum possible user capacity of 130 by the Erlangs to estimate the maximum number of subscribers we can sell to. So we now have an estimate on the maximum number of subscribers we can carry per channel. That is 100130 = 13,000. (You can also use this to estimate total system capacity which I have done.) Data traffic like voice traffic has a maximum number of users that can share a high bandwidth connection at anyone time. That 800kbps is fully capable of supporting ~40-160 (depends on a number of factors but much more if you limit each user to 56kbps or something like that) high data connections at anyone time. Let's just assume 100 for an estimate (someone with much better feel for this could help here). Now let us assume that these 100 users must access the network the same way as the voice users (say they get kicked off after so much inactivity) such that we still need to apply the subscriber usage intensity factor. Using the Erlangs per fortnight to find a maximum number of subscribers yields 100100 = 10,000 subscribers. All of the data stuff above assumes one gateway with ~1 channel available for data. (Just kidding about the fortnight.) So about the same number of subs as voice. Now comes the real kicker. What if you simply make a per monthly charge of $100 for anytime, anywhere high speed data access and Globalstar received $40 of that. At the current rate of 50 Minutes of Use per month and revs of $0.45/min for voice who brings in more dollars voice or data? Data = 10,000 * $40 = $400,000 / mo Voice = 13,000 * 50 min/mo * $0.45/min = $292,500 / mo Comments and corrections appreciated. RE: Vocoders. I've read that the vocoders in G* are variable (1.2, 2.4, 4.8 or 9.6kbps). These shift as needed. For instance when there is no talking, background can be encoded using 1.2kbps. When you have a very fast talker you might encode at 9.6kbps. On average a speaker is active 35%-40% of the time but a good number to estimate for actual loading is about 50%. So on average the loading for a single G* user is roughly say [(9.6+4.8)/2.5 + (1.2).5] or about 4.2kbps. Which is the top two vocoder rates averaged for voice and the bottom vocoder rate for quiet and both combined to give the 4.2kbps average rate. Which, also, is pretty close to the 3.6kpbs I've seen used here before. 9.6kbps is what seems to be used for all system capacity estimates so allowing for the variable vocoders it does seem the total voice capacity has significant flexibility upwards. finite_time

To: Rocket Scientist who wrote (15693)	8/14/2000 4:07:12 AM
From: Maurice Winn	Read Replies (1) \| Respond to of 29987

RS: <Then, they should charge wholesale rates of about $0.50 for a 72K file download, (and the retail customer could expect to be soaked for $1.50). 72K is several pages of text in memory hog MSWord; it's about one page of fax'd information, and probably scores of short e-mails. > My anti-data comments are really because quite a few people seem to think that since voice isn't working, we can be saved by data. No we can't. Data will be great for Globalstar in certain applications where Globalstar acts as the link where the WWeb use sends a request via Globalstar and gets the cheap, high-speed download from Skybridge or elsewhere. That's my understanding anyway. Winnstar is the main inchoate application, which will use Globalstar voice with Geo data [or maybe ICO?]. Where I don't 'get it' is that Globalstar is a quite expensive way of moving bits compared with geostationary satellites. So I don't see why any data would go via Globalstar unless it was integrated with some device which is primarily a voice system. Okay, the SCADA devices might be pretty cool and they are going to use Globalstar. I suppose they need only a little bit of data so they are economic anyway. I suppose as with terrestrial systems, a channel could be allocated to data which would be efficient. But still not cheap. A $1.50 soaking for 72K is not tempting. A geo download for the same amount of data would be less than 1c. I know where I'll get my data since I won't mind waiting half a second for the download. As you say, for email it might not be acceptable. But people will want images via Windfast and those images will be expensive if they come via Globalstar! Guessing, Mqurice PS: I also assume, but have no idea, that Globalstar WWeb devices could download data from Geos using the stubby little aerials, the same as the other Geo and ICO system handsets have just a little aerial. The ACeS handset for example: acesinternational.com