the data center

Decades ago, computing power was concentrated in mainframes tucked away behind the scenes because there was no alternative — only a hulking room-sized box that could contain any significant amount of computational power. The idea that this power could be distributed rather than centralized seemed like such folly that in 1943, IBM Chairman Thomas Watson said famously (or infamously) that “I think there’s a world market for maybe five computers.” | aaron weiss, computing in the clouds. netWorker, vol 11 Issue 4, December 2007

in the 1970s came the era of personal computer, but maybe one part of the cloudy thing at least is nothing more as to throw light on watson's perspective again.

carrier and wireless p3k set the pointer to an article that carriers are the biggest thread to innovation: today phone companies cling to carrier's coat-tails, "developing what they ask for instead of taking risks and testing new concepts in the marketplace" (McRae, Vizio, link above).

anyhow, some companies as google are already investing in new satellite projects: o3b estimated start: 2013. another project in the pipeline is msci's commstellation, a canadian endeavor, scheduled for 2015.

carriers and fibre to the home (ftth) another never ending carrier story is "fibre to the home". in the uk farmer's lady chris conder and her team of b4rn are celebrated heroes by uk ispa's recent event. they entered the provider business and dig up the last mile by themselves. they are farmers, so they know how to dig, (guess, that was a quote. yet as i don't have the source by hand, i don't label it as such for the time being.)

telcos themselves have done a lot of studies in the 1970s already. they even organized a whole track dedicated to optical communication systems back in 1975 at their "world telecommunication forum". charles k. kao gave a speech at this event. in 2009 he received the noble prize in physics, but even that didn't itch national operating telcos aka ptts up to today.

..The substance of the paper was presented at an IEE meeting on January 27, 1966. Most of the world did not take notice – except for the British Post Office (BPO) and the U.K. Ministry of Defence, who immediately launched major research programs. By the end of 1966, three groups in the U.K. were studying the various issues involved: I myself at STL; Roberts at BPO; Gambling at Southampton in collaboration with Williams at the Ministry of Defence Laboratory.

... Since the deployment of the first-generation, 45 Mb/s fiber optic communication system in 1976, the transmission capacity in a single fiber has rapidly increased: we now talk about terabits per second. In order to understand the fundamental limits of fiber-optic communication, the Terabit per Second Optoelectronic Project was launched during 1982–85, involving ten research organizations. The target technology, three orders of magnitude higher than the then state-of-the-art, was considered impossible at the time. | kao's noble prize lecture, 2009

astonishingly c. kao predicted 1975 the deployment of fibre the other way around: first – he thought – short-haul will be provided, long-haul will follow. well in hindsight we know it happened the other way around. we have long-haul and despite testcases as @21 districts in cities, which were funded by the eu+nation state, not many last miles are lit. however, on the 6th of october 1975, the fibre optic world looked promising:

... despite the greater economic benefits of very high capacity systems, it is probably that lower capacity systems, say 2.048, 8.448 or 34.304 Mb/s, may be introduced into heavily loaded parts of junction networks before long-haul high capacity systems. short-haul video links are a likely candidate for early introduction also.

various specialised non-PTT applications are already well advanced and will precede PTT applications, including aircraft, ships, high voltage electricity-generating and industrial sites, etc. bearing in mind the experience soon to be gained from these, and the demonstrably excellent economics, introduction of fibre optic systems in PTT networks may confidently be expected by the early 1980s. | k.c.kao, m.e. collier, fibre-optic systems in future telecommunication networks, 1975

economics of fibre-optics as seen in 1975

- the normal trend for decreasing cost per circuit with increasing capacity becomes even more marked with fibre optics, because cable costs do not increase greatly and repeater spacing reduced only slightly with increasing capacity.

• fibre optic systems appear more economical than coaxial cable systems over the entire range, but may not compete with low capacity systems on VF-type pair cable.

moreover, the small size and light weight could revolutionised installation techniques and permit a much greater usage of duct space - areas of very high cost at the present time.

in need for new equations

1. it should be taken into account, what telco's attitude actually costs the public. what does it cost that they laid out copper instead of fibre to the home for 30 years? at least 30 years, as the success and use of fibre was well proofed as early as 1980s, long-haul provider know well. And, ~18years therefrom as monopolists in most parts of europe.

2. what are the costs of cable, including power-feeding, repeaters, multiplex equipment costs, and ducts.

things to watch, to read, and to learn:

bufferbloat by jim getty tcp sucks by bram cohen. and the cacm discussion: bufferbloat: what's wrong with the internet

tcp doesn't suck

to dig on: Access Control and Copyright Protection for Images RACE Project M1005, 1994

sidestep: if this was the way eu project papers needed to be written in the 1990s, i guess the commission should simply go back to their beginnings in order to make it easier to apply.

anyhow: an overview of dreams and ideas worldwide: Architectures for the Future Networks and the Next Generation Internet, 2011