Interesting People mailing list archives
IP: BROADBAND WON'T HAPPEN BY ACCIDENT
From: David Farber <dave () farber net>
Date: Mon, 31 Dec 2001 08:28:57 -0500
From: Dewayne Hendricks <dewayne () warpspeed com>[Note: Posted here with the permission of the author. Peter used to be the CTO of British Telecom (BT), so he knows what he's talking about! DLH]BROADBAND WON'T HAPPEN BY ACCIDENT *) Peter Cochrane ConceptLabs CA <mailto:petercochrane () conceptlabs net> *) Published as " Life begins at 100 Mbps !" in issue 17 (October 2001) of Billing (www.billing.co.uk PBI Media.UK) I can remember a time when paying for computing time by the minute and data storage by the KiloByte was the norm. When downloading and printing meant taking a lunch break. When batch computing meant waiting a day or a night to see if a program had compiled and run. I can also remember when teletypes at 110 Baud were replaced by VDUs at 300, 600, 1200, and then 2400 Baud. When only rich people had a telephone in their home, mobile phones had not been invented, and all telecoms and networking was expensive. So what happened? Integrated circuits and fiber optics changed everything as processing power, storage and data transmission accelerated in speed and density whilst costs plummeted. We have seen an exponential (doubling every year or so) growth in capability of everything IT with exponentially falling costs, and relatively speaking our technology now costs nothing. I can now afford a 100Mbit/s LAN, and 11Mbit/s Wireless LAN in my home. Every computer has a 1GHz clock and over 500MByte of RAM, and a 240GByte server takes care of all storage and back up. And until recently my WWW access was a 2Mbit/s radio link into BT Laboratories. This might all seem extravagant, but to my youngest son it all sucks. And he is right; it is too slow and wholly inadequate for a multi-media world. Believe me, life begins at 100Mbit/s and up when we perceive response times to be less than a second. When I retired from BT so did my 2Mbit/s radio link and I returned to an insane world of no bandwidth. My WWW access is now a 56Kbit/s dial up modem delivering a puny 48Kbit/s. Unfortunately I live in a high-net-worth community on the periphery of town, which perversely means none of us can have ISDN, ADSL or a cable modem. We are on the end of ancient telephone cables unable to support wideband access, and live in an information ghetto of people who would be only too pleased to pay for speedy access. I'd feel happier if my situation was the exception, but it isn't, it is the rule. No one in the EU has wideband, and <10% in the USA enjoy adequate facilities. Across the planet the majority access the WWW through a dial up network of old and limited technology. To me it is a miracle that the WWW has taken off and I see network limitations as a key factor in the recent demise of the dotcoms. What went wrong, and how are we going to get it fixed? For sure the roll out of fiber into the long-lines network was, and continues to be, spectacular. It has totally sidelined all satellite capacity and links every continent with huge amounts of bandwidth. As I recall, in the initial phase there was never a business plan to totally fiber nations or the planet. It was a no-brainer, the cost per bit and operational advantages were so great, individuals made decisions and corporations just rolled it out. Suddenly a network of copper with repeaters spaced every 2Km on land and every 10Km under the sea seemed like the Stone Age. Optical fiber progressively spanned 10, 20, 30, 40, and then >100Km. Networks in Europe no longer required buried repeaters and the reliability of systems meant fewer craft people and huge reductions in standby plant. So dramatic was this advance that Telco economics were transformed. Instead of a near 50:50 split of capital and operational costs between local loop and long lines, long lines fell well below 5%. Optical fiber saw a 99% fall in network electronics, the eradication of lightening damage, and cables totally impervious to water ingress. Fault and failure levels plummeted to previously unimagined levels, and huge economies were realised as a result. So why didn't Telcos roll fiber into the local loop? Well, the story is a bit like the Gulf War (they stopped right on the border, and made the fundamental error of not completely vanquishing the enemy) and now we suffer soda straw communication pipes as a result. Across the planet dumb decisions were made out of political necessity, individual and corporate self interest, technical ignorance and financial ineptitude. In 1986 fiber direct to the home was demonstrated to be financially viable. For the UK it would have cost a mere $22Bn to fiber every home, which in comparison to the $34Bn just wasted on 3G mobile licenses was a bargain. For many countries the optical fiber story goes something like this:1) The managers ruling the local loop saw many of their colleagues in longlines networks having to take early retirement, accept new assignments, and move companies.2) Managers and craft people could see fiber technology being rolled out by PhD's from laboratories, and they could not get an inside line, they couldn't, or didn't want to, get in on the act. They began to feel insecure and atgreat risk. 3) The financial people didn't care! But they couldn't get consistent numbers for their models. Where was the advantage of fiber going to come from? They would argue that the advantage in long lines was now obvious, but there are no repeaters in the local loop. Worse, I suspect that the waters were muddied by (1& 2), not in a conspiratorial way, more because they didn't know. But as a group there was not the imagination or courage to create models that encompassed all the potential gains, which included a reduction from thousands to tens of Central Office sites. 4) And finally, there was the technical community who I feel have to shoulder at least 50% of the responsibility. On the fiber side the researchers bulldozed ahead creating more and more technology and more and more competing options for the local loop. No one stopped to make it really practical. Worse, those researchers who had devoted their lives to copper, and had failed to make the transition to fiber, failed to grasp the magnitude of the possibilities. They continued their expeditions into copper and digital signal processing with wild claims that DSL could deliver 2, 6, 12, and perhaps 50Mbit/s using the existing copper cables. What a dream! The bean counters could hardly believe it, milk past investments in copper by attaching a modem on each end of a customers lime. Game over! So DSL became the end game, and fiber was pushed aside. At a time when a PC had a 6Mbit/s clock why would you need more speed in the local loop? And anyway, all this Moore's Law stuff will most like not happen anyway etc. I would have thought that they all have to be feeling pretty dumb now! DSL is a universal and monumental failure with far less reach, and far fewer bit/s delivered than promised. And of cable modems? Well they have done a bit better, but the penetration is poor, the service limited, with installation and operation proving very expensive. A long past UK Prime Minister, Harold McMillan, was famous for using the expression: "well gentlemen, here we are". He always looked at the situation as was and disregarded prior events. Bringing this up to date in a PC world I think we have to say: "well folks, here we are". We cannot go back to correct the mistakes of the past, but we should use them as valuable lessons in our forward thinking. Just how are we going to get wideband out to folks who are willing to pay, and have a dire need to communicate and work at speed? Let me first dismiss all forms of satellite technology, including geo-stationary, low-earth-orbit, balloon and aircraft platforms as being an insignificant prospect. Yes they will provide a valuable asset in the provision of service to far-flung, hard-to-access, geographically remote locations, and mobile and clustered platforms. But in the grand scheme of a global wide-band future they really are going to be a very small %. Satellites never amounted to more than 25% of all international communications capacity, and today they are less than 1% and falling. In contrast, future generations of mobile phone networks, do I suspect have a much bigger role to play and cannot be dismissed. So I am going to concentrate on the kernel problem, the local loop. Here are some key factors we now need on the table.1) Telcos will never deliver wide-band communications. It is not in theirinterests to do so; it isn't in their business minds or models. They are into call minutes and billing systems. They are old, gray and don't get it, and they don't intend getting it. 2) The cable companies are slightly better, but have a broadcast mindset, where wide-band is an add-on, a kluge, and not a primary business or technology.3) Both (1) and (2) have missed their opportunity, wasted time and money ona vast scale, and are now going bust. Five years ago they could have rolled fiber into the local loop, they had the money and the people back then - now they have neither.4) The sheer numbers of people required to undertake the total provisioningof a wholly fiber local loop are no longer available. They have moved out of the Telco and Cableco arenas, and into other areas of IT and retirement. 5) There is technological confusion with the old world of IP over an oft concatenation of PSTN, ISDN, Frame Relay, ATM, Sonet/SDH, and WDM systems, up against a new Metro-LAN world of IP over DWDM. So how are we going to advance? I think we have been here before. Back in the 1940s USA TV companies couldn't find an economic means of providing signals to outlying communities. So people clubbed together to build towers and antenna systems, and wired their houses to realize Community Antenna TV. This was so successful that the expanded systems became the Cable systems of today. In a similar manner, youngsters now frustrated by the lack of bandwidth are linking homes with CAT5 LAN wiring strewn across gardens. Schools are buying 802.11 wireless-LAN cards to create their own networks at a much lower cost than building wiring schemes. There is a message here for the network companies, and a huge opportunity. If they don't provide the bandwidth demanded by rapidly advancing terminal technologies, people will just set to and provide their own. Hotels, schools, coffee shops and places of work are starting to look like the phone boxes of the 21st Century. People are gathering there to satisfy their craving for wide-bandwidth, which isn't a 56Kbit/s or 2Mbit/s dribble, but orders of magnitude more. In some countries >80% of the population is within 1Km of an optical fiber, but they cannot gain access. So here is a hybrid solution pulling together multiple opportunities spanning fixed and mobile networks. Telecoms is on a back foot with numerous operators in serious trouble. After being corporately raped of over $100Bn for 3G mobile licenses across the EU, operators and markets suddenly realized that a further $100Bn of infrastructure investment is needed, and $100Bn in technology and service developments plus marketing and sales. This dictates a minimum $1000 spend by every EU handset owner assuming they all instantaneously abandon GSM and convert to 3G overnight. The operators can't see a payback within 10 years, and are finding it difficult to come up with business models that makes 3G pay in. In the UK, operators paid $34Bn for licenses in a market of around 45M users where 30,000 new base stations will be needed at a time when the public is objecting to these eyesores. So what can be done? I see very few options to save the current situation, but here is a possible subset;1) Mobile operators could move to subsidize 3G by increasing the price of 2G handsets, call charges, whilst simultaneously reducing their Quality ofService, and the number of services, ie we all get to pay much more for much less. 2) Governments could see the error of their ways and repay the 3G license monies they raped from the industry; highly improbable! 3) Operators could merge to share a single infrastructure build and operation. My guess is that Europe can only afford 4 operators in total!4) A partnership between fixed and mobile operators plus manufacturers could be incredibly beneficial.5) Most lucrative of all would be the eradication of the local loop. At arough estimate an average Telco could displace 50% of operating costs and 60% of employees if it replaced all its wire-line plant by 3G handsets. 3G offers wire-line voice quality, and higher aggregate data rates than wire-line modems over a wider geographical area than DSL. Over 50% of all Telco faults are in the local loop, with about half inflicted by their own craft people, and the rest are more or less evenly split between water ingress into cables and damage by other utilities. 3G would get rid of a huge raft of problems and place the onus for local loop operations on the customer. Moreover, the majority of the existing line plant is optical fiber out to locations ideal for 3G mini-base stations. So far this is the only scenario where everybody wins! At this point the Telcos could mine the copper and lead blocking duct tracts, recover money on the sale, and then lease the empty space to new companies moving into the MetroLAN space. The reality is that there is not a single technology that can do telephones, CATV and data. It will be some time before that prospect becomes a reality and this solution gives natural stepping-stones. On entry into the last mile the new companies have the choice of installing fiber direct into home and office, or much more attractive, erecting wireless access points in the street. Communities of 802.11 WirelessLAN users already abound across Europe, and they are building fast in the USA. As a solution it is almost perfect! Think about it, the customer purchases, installs and maintains all their own equipment. A very large chunk of the cost base is moved out of the operators hands and is gratefully received by the end customer. Will it work? Will it be acceptable? Ask the many who do this already in hotels, airports and Costa Coffee houses across the planet. As far as I can see it is a done deal. As there is no new money available and customers are highly unlikely to throw more at the 3G for nothing, an increased cash flow can only come from displacing existing market. So to a vital question; who has all the money and how can they be displaced? Well how about the music, movie and computer game industries? Looking around most homes there is ample evidence that families spend substantial amounts in these sectors. They are a natural for the non-physical distribution of bits - which can be written direct to disc of any form or location: home, office, car, handbag, pocket or belt. So it is somewhat ironic to see the American music industry hell bent on the destruction of Napster and its variants at a time when CD sales increased by 15% last year. This type of service, and derivatives, is probably the only one big enough to support a community of 3G users. With all the digital TV network news and programs dying on their feet, plus our apparent dislike of the small screen, there are not a lot of options open to service providers for advertising supported services. On the upside I don't think anyone or any industry will be able to stop the Napster variants and the public downloading of music, it may even become commercialized in due course, with a natural progression to high quality pics, and movies. But all of this is fundamentally dependent on freely available bandwidth anytime, anywhere - a natural fit for 3G. The other essential feature is that music is now a mobile phenomenon; people wear it, ride it, have it everywhere. Trying to be objective in IT markets is extremely difficult given the rate of progress. For sure, we can accurately predict the availabilities and capabilities of new technologies, but we have no chance of predicting what people will buy when. It doesn't take a lot to switch off the public, but it is also easy to identify and address their frustrations. Everywhere I travel I see increasing numbers of people engaging in Peer-to-Peer networking. People are using Infra-Red Ports, Wire-LAN, and Wireless-LAN connections to create instantaneous networks PDA, Lap-Top-to-Lap-Top,. What has prompted this new activity? I think the motivators are principally twofold. First, all of our devices have totally outgrown network connectivity. Compared to a 30Gbyte Hard Drive and a 1000MHz clock, a 56Kbit/s modem looks decidedly sick and is totally useless for the downloading and swapping of even modest size files. A cursory sweep of my own laptop revealed that 25% of all files are >0.5MB, 15% >1MB, 5% >10MB and 2% >100MByte. Most music and movie files are over 3Mbyte, whilst digital camera images are generally above 250KB. So what use a 56Kbit/s modem, and who has the time to wait 20 minutes for a single download? Just 10 years ago none of this was true, the world was a bigger and less rich place. Today far more is available, and being created, with no limits to the applications and need for more and more bandwidth. So, how long before the network companies wake up and meet this obviously growing demand? I don't think they will. They are focused on exploiting what they have installed already, and are happy to see >90% of net connections via dial up modems. And worse, their mounting debt's probably means they can't even do it if they wanted too. The only solution I can see on the horizon is a combination of Peer-to-Peer and Ad-Hoc wireless networking. Despite all our advances with optical fibre and radio, the cheapest bit transport solution still remains FrizbeeNet - the CD in an envelope. In my future world I will be able to use a mobile phone/device to access terminals and large screen displays I encounter. It will double as an electronic wallet, security device, and medical records repository. I'll be happy to use this as a mobile banking device and pay by the minute, transaction or whatever. But for wideband access I want to pay a single fee for wideband access anywhere on the planet, static and or mobile. Finally, there is one further mind set change required, network oerators have to start wasting bandwidth, they have to look at dense coverage using pico, micro and macro-cells. A new phenomenon now sees a natural and growing state of chaos. When people cluster their actions become co-ordinated. The coffee break at a major international conference sees 300 mobile phone calls initiated within two minutes. A freeway accident results in 1000 calls initiated in 10 minutes. Train and plane cancellations have a similar impact. The old fixed line telephone network had a peak to mean traffic loading of 3 or 4. Our mobile networks are now closer to 50, whilst the Internet can exceed 1000. With new generations of mobile device it is likely to get far worse as the size of file gets bigger and transfer time grow. And a further aggravation will be the introduction of more things on line than people. There is no certainty in the world of networks and to satisfy a chaotic future we will have to vastly over provide bandwidth. The old ways of grooming traffic to increase efficiency will only lead to more unusual, unpredictable, and probably spectacular network failures. No one would choose to drive a Formula 1 racing car to work, it would be far too unreliable because it is so optimised. In contrast a Mercedes will go for years without incident by virtue of its lack of optimisation. In the same way SMS has become the dominant mobile mode in Europe, we can anticipate even more surprises in networks of the future. So network designers and operators have no choice, we are going to need wideband ports to the conference room, floor, building, and street. The link to and between our devices will increasingly be wireless, and our access seamless, with optic fibre being the primary access conduit into the global network. Of course there will be occasions when optical line of sight will find a role, as will point to point wireless, but make no mistake, ubiquitous fibre reaching out to individual nodes and users has to be the dominant mode if we are go universally wideband.
For archives see: http://www.interesting-people.org/archives/interesting-people/
Current thread:
- IP: BROADBAND WON'T HAPPEN BY ACCIDENT David Farber (Dec 31)