info on wireless (web-enabled) PDAs

[jaws]

Hi all,

I'm a college student researching on the energy consumption of a PDA. (This is motivated by a Forbes article in 1999 that claims that a PDA uses as much energy as a refrigerator. We (my colleagues and I) don't believe this claim and want to find evidence to refute it)

1) Would anyone know how many wireless PDAs there are in the US? We found a figure of 46,000 (Y2001) by eTForecasts. Is this a good estimate of the number of wireless PDA users in the US?

2) Also, for those who have wireless PDAs, what are your usage patterns like? How often do you use internet related features (email/web clipping/IM/etc) and how long do you use the connection? Once a day for 15/30/60 minutes?

3) How much data is transferred over the wireless connection per day? Would anyone have a sense of the number?

4) There seem to be 2 types of wireless services available -services that offer limited data rate (PalmNet's 100 KB limit per month with 3COM), and unlimited data services. What do users usually opt for ?

5) Are most wireless PDA users connecting online through modem dial-up using their cell-phones or through the 802.11 network using a network card? My feeling is that the former is a more popular option?


Sorry for this long post. I would really appreciate any wireless PDA user who can provide my informtion based on his/her own experience or knowledge about this field.

Thanks so much !!!

[Jennifer S]

I don't see how your questions relate to the energy expenditure of a general PDA, but I wanted to respond to let you know why I think you're not getting any responses. Maybe by answering, I'll get somebody to argue with me.

1. I would suspect it's much higher, since there is probably no way to estimate how many people are using PDA's wirelessly with cellphones. In any case, you need to define a "wireless PDA." Do you mean PDA's that have the wireless function built in, or do you mean PDA's that have the capability of accessing the Internet if connected to a wireless modem?

2. I think I'm like most users in that my usage varies tremendously from day to day. It depends how much I'm away from my office.

3. This amount is probably impossible to estimate in most cases. I know that with Palm.net, you can look up your data usage, and GPRS plans measure the amount of data, but most other plans are either unlimited or bill by the minute.

4. 3COM, by the way, has been separate from Palm for a long time. I don't think any individual user can tell you which option people usually go for. I will say that for the i705, 100MB isn't much, so if you use it much at all, you would go for the unlimited plan. GPRS plans for unlimited data aren't even offered in most places. CDPD service (EarthLink, for example) is only offered as an unlimited plan. So even if you measured how many people have limited vs. unlimited plans, you would also have to take into account the fact that many, if not most, of them didn't have a choice.

5. There are many more ways to go wireless than through cell phones or 802.11b networks. Blackberry and i705 devices use the Mobitex network. There is the CDPD network, and there are Bluetooth networks, just to mention a few options. Most people who have 802.11b access have it because there is an 802.11b network in their workplace, and they often use a different wireless method when they're out of the office. I can't begin to estimate what "most" people do, although the cellular option is almost always the least expensive route.

[ramuir]

No arguments from me, Jennifer!

Personally, jaws, I think you have set yourselves a worthy, but perhaps nearly impossible task to do with any accuracy. I have published some power consumption data for an m505 with a Bluetooth card, under a couple of different usage scenarios, at this location:

http://discussion.brighthand.com/pal...ight=bluetooth

For Bluetooth and 802.11b connections, the amount of data transmitted is not nearly so important as the mode of connection and duration of connection. Idle packets are used during most connections just to keep the hookup alive, and the radio sets are in full-power mode, or nearly full-power mode for many of those connection states. As I've stated elsewhere, Bluetooth is much more power conservative than 802.11b for the PDA market, so a wide variation in energy usage models will be experienced by users of these two different technologies.

In the case of the cell-phone based connections, the wireless functions may be completely (with cabled connections) or only partially (with Bluetooth connections) performed by the phone. Here again, potential differences in energy usage models will occur.

This is all sounding discouraging, I know, but I don't mean it to be.

Your best bet is probably to define a few usage models that are supported by data that you can obtain, and then derive some supportable lower and upper bounds for energy use within these constraints. It's still going to end up kind of vague, but perhaps that will serve your purpose.

Cheers

[jaws]

Thanks for the replies, Jennifer and ramuir!

By wireless PDAs, we include all PDAs that have capabilities of connecting to the internet by whatever means (modem/bluetooth/802.11), but we want to find out how many people are actually using these.

ramiur, we know that this is going to be a very vague estimate, but we're trying to get a general ballpark to see if a palm really consumes as much energy as a refrigerator! Personally, we don't think so, so we're trying to model the typical wireless palm usage scenerios and try to estimate the amount of electricity used.


I know this is not the appropriate place to post this other question, but would anyone know how many minutes people in the US spend online, and how much data flows over the net annually?

Thanks!!

[ramuir]

jaws,

I haven't worked in that area for about 2-3 years now, but I can tell you that people like MCI, Alcatel, Sprint, etc. often publish the results of studies about their general network capacity and usage. I think the Gartner Group or some similar industrial consultancy may also have done studies. The news releases about Global Crossing may also have such data in them, since they were a bandwidth vendor and network connection supplier.

Beware of some of the numbers. Often the numbers will refer to installed capacity and may include dark fibers, unused capacity, under-utilized network connections, etc. This is far from the actual user bits that are revenue bearing.

Another thing to beware of is the use of router loading as an estimate of user traffic. Packets will go through many routers between source and destination, and there is no way to estimate how many routers a packet sees on its path.

A number that does stick in my mind from a conference in late 1999 in Virginia is that no matter how big the traffic number is or will get, somewhere between 75% and 90% (by volume of user bits, not by numbers of connections) of the user data is projected to be porn. That was an astonishing estimate. Estimates for this kind of traffic breakdown were made off-the-cuff by seemingly knowledgeable people at a conference jointly sponsored by the National Institute of Standards and Technology and the National Computer Security Center.

Cheers

[jaws]

Thanks for the advice, ramiur!

Will keep searching for reports on the amount of dataflow through the internet, as well as time spent on the internet (all for USA).

We have found some numbers, but the disparity is so high that we're not sure which to take, hence we're still looking for more.

We have stats from Network World Fusion saying that 311,000 trillion bytes/year were sent through the internet, and In-Sight Research Corporation saying that 1,980,000 trillion bytes/year were transmitted last year.

Would you have a sense of which is more possible?

Thanks!

[biglig]

Well, a Palm M125 is powered by 2x AA batteries... I know you can't measure consumption that way, but assuming you know how much power a refrigrator consumes, you should be able to work out how long a fridge could run on 2x AA batteries.

Since I would estimate this time to be easily under a second, that should give you the proof you need

[ramuir]

As for the number of bytes transmitted, I did a quick survey of some operating long-haul data links, and just a cursory examination turned up a large number of them. The larger data quantity (1,9800,000 trillion bytes per year) is an equivalent of ~200 OC-48 links run continuously. I was easily able to turn up about 10-50 times that in operating capacity on long-haul links. My number may be a factor of as much as 100 too low - I only spent 20 minutes looking. The majority of data is not carried on long-haul links, but rather shorter links (from your terminal to your ISP for example, which is probably not in Singapore or Paris, unless you are).

In addition, a little more research showed that Bell South has ~620,000 DSL users. Although it is a simplistic comparison, at a modest 256kbps symmetric continuous usage, this would be about 125 OC-48's worth of traffic. Thus, either of your numbers is within reason - there is no particular justification for accepting or refuting either one.

As for running the fridge on 2 AA cells, that would be a reliable estimate that could be made. Unfortunately, 2 AA cells don't last the same time in a Palm for all users, so the energy comparison still has some fudge space.

Cheers

[jaws]

Thanks for the replies, biglig & ramiur!

Comparing the palm & fridge's usage of AA batteries is interesting, but we want to go into more detail than that.

We have estimates for the amount of electricity the palm directly uses (from charging), and we want to take into consideration the entire network for web-enabled palms (ie. the cellular network, the landlines & the internet).


What we really want to know is how much data is transmitted by the average wireless palm user (per day/month). It would be good if we could get a sense of that, since we're not wireless PDA users ourselves.

Thanks so much!

[Jennifer S]

The problem is that most users have no idea how much data they personally use, much less what the average user does. When I'm online on my Treo, I have no idea (and no way of measuring, as far as I know) how much data I'm transmitting and receiving.