From: Greg Baker Date: Mon Oct 21, 2002 11:15:12 AM Australia/Sydney To: Subject: Re: Companies and carriers Hmm, very interesting. It just dawns on me that there are two totally different starting points that people use on wireless, with completely different assumptions. And, as usual, we're both come from different ones. I might post this up on the web because I think it's interesting. Here goes: Model 1: "802.11 lets you do cable modems without the cable" Pricing: Around $30-$80/month per household. Mobility requirements: Essentially zilch Target markets: Individuals and SMEs that can't get ADSL or cable, or want something a little cheaper Quality-of-service: Bandwidth sensitive? Required speed: Around 200Kb/s, bursting up higher Crucial factor: Price, price, price. Coverage: Just about anywhere Who's going to do it: Mostly middle-tier and lower-tier ISPs and PFYs in high geek-density locations. How they do it: String an antenna on their roof and put dishes on the roof of customers. Market size: No reason it couldn't hit the same penetration as current SE Korea broadband. Model 2: "802.11 lets you make mobile data services like mobile phones do, only faster" Pricing: ??? WAP is worth less than $10/month. ??? Must be above $100/month? Mobility requirements: Crucial Target markets: Companies wanting a mobile workforce. Quality-of-service: Latency sensitive? Required speed: ??? More than 10kB/s (if less, just use IrDA and GSM)?? Less than 100kb/s? Crucial factor: Support, mobility, single billing. Coverage: Cities. Not enough money to do anything much else Who's going to do it: Rich carriers How they do it: Put an 802.11 access point on every base-station. Market size: Nobody knows. Model 3: "802.11 lets you long-haul links for middle-tier bandwidth services" Pricing: $1000s/month Mobility requirements: None Target markets: ISPs and companies wanting a thick bandwidth internet feed. Quality-of-service: Reliability Required speed: As much as you can throw at it. 2MB/s - 200MB/s??? Crucial factor: Quick response to problems. Coverage: End user pays it. Who's going to do it: Big carriers talking to littler players How they do it: Very finely tuned single-directional dishes. Market size: Depends if Model 1 happens or not. I can't think of any other models off the top of my head.... If the legislation happens, Model 1 goes ahead. Nothing can stop it; even if it isn't terribly profitable, PFYs will still supply it to subsidise their internet addiction. They then use Model 3 as their up-link, because that's what they know, and because it's the cheapest. Model 2 then becomes a very odd duckling, indeed. If the legislation doesn't happen, Model 2 goes ahead full steam. Model 1 goes ahead very softly, because only a few players will be able to afford to do it. Model 3 is pretty much irrelevant because of ISP consolidation. Now, if you are a model 2 player, can you do model 1? Yes, but you have to price it above your own broadband offerings (or equal to), otherwise you suck customers out of your broadband offerings. Do you need to offer both in order to get enough users to make your system viable? I suspect the answer is yes. How many companies actually have a use for a mobile workforce? This sounds a silly question, but I honestly don't know. Obviously on-the-ground technical people might want to have access to their office data; possibly sales people? Who else? Is it enough to sustain model 2 on its own? On the other hand, can a PFY do model 2? Possibly -- the PFYs and ISPs form a Federation Of PFYs. They modify their pam.conf (authentication scheme) to say "try a local user, otherwise fall back to a FOPFY lookup". They then sync with each other at the end of the month via PayPal. Euckk. Very cheap. Very tacky. But very easy to set up if you are already a Model 1 player. >>As the chief instigator of the carrier-free 802.11 scheme, I suppose I >>should have documented how it works for carriers. Here goes: >> >>a) if the legislation goes ahead, then the market for big carriers >>getting involved in hotspots will essentially disappear. >[Rog42] I >disagree, Telstra wouldn't have bought SNG @ $500m in the red for the >price they did if they thought there was no market, the same goes for >Optus & VF. No, Telstra are extraordinarily good at buying things at stupid prices that they don't need. And they wouldn't have a clue what was coming up in the legislation. Define "hotspot" -- I think there will be carriers that sell one-stop-shop solutions for setting up hotspots, but if the carrier-free legislation goes through, they won't need to be involved in the setup themselves. Companies will do it themselves. >> There'll be no >>way that they can compete on price against the pimply-faced teenager >>with a Linux box, aerial and wireless card. >[Rog42] Agreed, although the >PFY won't be able to: >A) Provision Doesn't need to. >B) Bill Why not? They set up a BPAY account and email you a monthly invoice. Works for me. >C) Deploy Why not? What is there to deploy in the PFY model anyway? >D) Support Why not? >E) Manage When you're managing one central node, the management problem isn't that hard. >F) Provide ubiquitous coverage across Australia for subscribers anywhere >like the Telco's can - they have all of the above systems in place, and >theoretically at least can just 'turn on' another spectrum. All the PFY >can do is provide a local cell for a couple of people, there is no >commercial sense in that, why would I want a mobile phone that only >works in my street? I want a mobile that provide access at least >nationally, preferably around the world. Indeed -- if mobility is what you want (model 2). But if it's cheap broadband (model 1), that's a different story. >>b) The carriers never really wanted to be in that business anyway, or >>if they did, they weren't thinking clearly. It's far too much dealing >>with individuals and not enough of dealing with corporates (even small >>corporates are better fodder than individuals) where the decent profit >>margins are. > [Rog42] Again I disagree, the carriers have two problems in >Australia (and one globally): >1) Broadband to the community, although Aus has a higher internet >penetration than the US (55.4% vs 52.8%) we are shocking when it comes >to broadband (6% vs 32%) PwLAN, and wLAN (in SOHO's, schools, libraries, >community halls etc) gets broadband into rural areas, either at the end >of a DSL or a 2-way Sat link. >2) GPRS take-up is next to nothing. GPRS is way to expensive, too slow, >and as yet there are very limited applications. Basically just sip >Internet access as opposed to slurp (sit down and surf, download huge >files). The road is to 3G (and 4G). The dilemma is applications which >require the bandwidth & mobility of 3G need to be written and accepted >and purchased in the public domain. GPRS won't cut it. So Telco's are >complementing their GPRS networks with PwLAN. (which gives rise to the >Nokia D211 GPRS/WLAN PCMCIA card) If you think that this is a pipe >dream, we have consultants\ working with BT rolling out 4000 hotspots, >Telia (Sweden) rolling out 2000, NTT DoCoMo rolling out 10 000, Korea >rolling out 10 000. etc. etc. Australia is about 12 months behind the >rest of the world. > The business is not in a spattering of local >grassroots PFY's with their linux boxes and pringles cans, the business >is in enabling businesspeople, whether itinerant local or travelling, to Actually, I think it's both. And slightly more -- it's about ISPs offering broadband to their customers cheaply and competitively. And be careful of comparing us with Sweden or what BT is doing. The legislative framework in Australia is different. And it may be about to become very, very different. >connect to their corporates. It is in enabling companies to reduce their >floorspace costs by giving employees ubiquitous access to their >corporate LAN's in areas they can work in. In this there is huge >revenue, BUT if the likes of XONE, Unwired etc get in there first, >without having to pay carrier fees, then Telco's can't match their >prices, "Not fair" they'll cry This is model 2. I'd like to argue that the "huge revenue" there is a bit of a myth. I think we need to come up with some specific cases where mobile 802.11 has some significant use that we don't have an alternative or substitute for at the moment. >>c) But... the pimply-faced teenager has a problem. Having just signed >>up half the local neighbourhood for wireless broadband (at dirt cheap >>prices), how does he get a decent up-link? Or decent speed >>connectivity to neighbouring cells? Run through the options for getting >>a 2MB-200MB uplink (and what the small wireless players are likely to >>have expertise in) and you'll find that .... a *dedicated* >>point-to-point wireless >[Rog42] He could just buy a wired link to the >net and proxy the authentication, that's how our architecture works. If the PFY's customers can't get ADSL or cable, then the PFY is going to have a hard time getting a wired link. Secondly, if the PFY can get wireless, he will use it as a preference over wired anyway, because that will be where his expertise will be. What do you mean by "proxy the authentication?" >>link is the way to go.... which you need a >>carrier licence to provide. (Otherwise it's not dedicated any more). >>And the pimply-faced teenager wants to ramp up slowly -- he may only >>have 10-20 customer initially, but will grow rapidly. > [Rog42] Only if >he can provide roaming to neighbouring and remote cells, on one bill, >which he can't! His market at best is local geeks. Much like the BBSes Why not? The Federation of PFYs is not that difficult to set up. >of old, some of which became one of the zillions of ISPs and IAPs in the >early to mid-90's. Now there are few ISP's, mostly much bigger, and >mainly the Telco's. We're just talking about ISPs in the Wireless sense. >WISP's. There will be small ones, but not forever (or even long) >> Eventually he >>doesn't mind getting a bill for $1000s/month for bandwidth and >>connectivity because he's making plenty more than that from his >>customers. >>d) So instead of being in the business of providing wireless hotspots, >>to individuals at a few tens of dollars per person the carriers are in >>the business of linking those hotspots back to locations of high-speed >>internet connectivity for thousands of dollars a month. Much better. >[Rog42] Nope, they want all the subscribers. Same as the mobile >business, and the ISP business. Ironically, this is access and >architecture like the mobile business and revenue models like the ISP >business. Oh and the Telco's will also (personal prediction) speak to >their corporate customers about providing corporate wLAN hotspots (i.e. >in the buildings) as part of a one-stop mobility solution - a fleet of >mobile phones, integrated mobile data, plus the private WAN and Internet >traffic. The bill is even integrated to their 'public' hotspot coverage. >Very compelling for the large company that wishes to reduce its office >space, consolidate all its data and telco charges. Also compelling for >the telcos, oh, and we can do this for them today. We have the >architecture, the mgmt software, the design services etc. >>e) Now the fun begins. A second player (may be a mobile access >>provider) muscles in on the pimply-faced teenager's territory. >>Perfectly legal. Uses the same wireless infrastructure in place. >>Hehehe. > [Rog42]You'll find this difficult, any wLAN company worth their >salt would disable this. Bzzt! The carrier-firee legislation demands that it be open. And that's the crucial point. If you want to maintain a closed network, you have to get a carrier license. Without a carrier license (at least with my legislation) you *must* provide open access and open transitive routing. (i.e. for 802.11, you probably can't do WEP and you must do BGP.) >Not to mention the real problem in that only >being able to use 3 channels (rather than the 10 in the GSM world) >simultaneously, the air is going to get very dirty in CBD's very >quickly. Why 3? I thought it was 11? And I think that's extended in 802.11g to much more. >>Oops, how do we get from the pimply-faced teenager's cell >>(with all our mobile customers in it?) back to our office where we can >>authenticate them? Umm.... we need a wireless link. >[Rog42] Any link >will do, oh, and 2-Way Sat doesn't need line-of-site. And 2-way satellite is *very*, very expensive. If you can find anything vaguely price-reasonable here, I'd love to hear about it, because my customers are screaming for it. (Actually, I think they're wrong, I think the latency will kill them from doing what they want to do, but anyway). >>And, oops we >>don't have line-of-site. But the carriers can place their base stations >> >>anywhere they like, they don't even need local council approval. So >>now the carrier has two high-volume customers in the region they would >>have hoped to have had one wireless hotspot. Hmm... sounds like the >>carriers are *better* off than they were before. Indeed they are, >>because they're still charging an arm and a leg for bandwidth without >>having all that messy expensive business of end-user support. Or even >>negotiating with another company to do that. And there are far more >>users because the barriers to entry are so much lower. >> >>f) So the end result? The carriers should be dancing with joy. They'll >> >>make far more out of the carrier-free no-charge 802.11 than they ever >>would out of having a monopoly on commercial 802.11. >[Rog42] Actually >they'll have the monopoly on that too, or at least the oligarchy.