Polar Bear with kissing cub Killian.COM Earl Killian Commentary Quotes Books etc. Friends Only

[Under Construction]Technology Fantasies[Under Construction]

Copyright © 1999-2000 Earl A. Killian. All Rights Reserved.


Modern technology has provided truly amazing capabilities, but the truth is our current level of technology could provide so much more than it does. That is because so much of our technology is hobbled by the need to be backward compatible.

What might such technology look like? This is all possible today (though I look to future trends to guide the choices that could be made today and I do include some wishes for future technology). It’s just not feasible or commericially viable given the need for backward compatibility. That’s why this essay includes “fantasy” in its title.

If the need for backward compatibility seems so overwhelming to render this irrelevant, then imagine colonizing a new planet. It would be possible to re-engineer all aspects of technology (and society) without the burdens of backward-compatibility. Much effort would be placed on defining state-of-the-art standards that would serve as a starting point of future evolution. These standards would be superceded as appropriate, but at least the starting point would be a good one, unlike the situation today. The relative costs of things would likely be significantly changed. In re-architecting this new world, emphasize efficiency and functionality; cost will follow with time and experience with the standards.

The Network

The network will have wired (e.g. fiber optic) and wireless components (using the electro-magnetic spectrum). What is the appropriate allocation of service to these two components?

Mobile network components should of course be capable of wireless communication. Mobile components have a power source that stores (e.g. via batteries) or generates (e.g. via fuel cells or micro-generators) energy that allows their mobility. Other components will get their energy from a power generation network (either local or global). There is no reason for the latter to use wireless communications, because their connection to an energy source can be paralleled with a connection to a data source. Indeed every power connection should include a network connection, as there are few devices that need power and that wouldn’t benefit from network connectivity.


  • Create a universal power connector that includes network connectivity (e.g. 3-4 copper wires and 2 fiber optics). The power connection must scale from Kilo-watts to milli-Watts. The network connection should be capable of at least 1Gb/s today, with scalability to higher data rates over time. (Imagine plugging in your vacum cleaner and having it download the settings for the current room!)
  • Create a universal mobile connector so that mobile devices can either operate connected the power grid and wired communications grid (hereinafter “the grid”) or operate wireless via their own power storage/generation capability.


  • Miniture fuel cells or generators (e.g. micro-turbines) powered by chemical fuels (e.g. alcohol) that replace batteries (and provide significantly higher energy storage capability and reduced time to recharge/refill).
    • Weeks of operation at heat-dissipation comfort threshold (e.g. the power level at which a handheld device becomes uncomfortable to hold)

Note that transmission of power sufficient for mobile components would be a technology that changes the equation here (e.g. solar power satellites beaming microwaves to earth), but that seems to have enough problems that batteries and fuel cells will likely win out.

The primary factors that should drive a new communications infrastructure are:

  1. Humans have finite bandwidth receptors (e.g. providing video at 200fps instead of 24fps is of no use);
  2. Wired communications bandwidth will increase exponentially for a long time to come (e.g. there is no fundamental limit given the ability to provide parallel pipes);
  3. There are limits to the bandwidth of wireless communications based on signal/noise ratios, the size of the spectrum, and reasonable distances between transmitters;
  4. There are limits to the distance of some frequencies of the spectrum (e.g. some are line of sight, and some cannot propagate well through building materials);
  5. The energy storage/generation (or heat-dissipation) limitations of mobile devices and the limits on computation and bandwidth that these impose.

Once we reach the point of sufficient wireless bandwidth to saturate the human senses, other considerations take over. At that point it is better to use further bandwidth for increased user count, cost reduction, extended battery life, etc. Similarly, mobile computation will be sized to the input and output capabilities of humans, because:

  1. mobile computation will be limited by its communication bandwidth to the rest of the network;
  2. mobile computation will be limited by power storage/generation capability; and
  3. it will require less bandwidth to do the computation in a grid-connected unit and communicate the results interactively over the wireless network.

Re-architecting our use of the electro-magnetic spectrum

It seems reasonable that all wireless communication should essentially be cellular and that the short-range spectrum should be reallocated with this in mind (I’m not sure for example that satellite bands aren’t better used for cells). Wired service should be used to connect the cells to the network. So for example, every house would have a wired connection to the network, and then the house would become a cell. Appliances that require significant power are going to power tethered and that will come with a network tether. Exactly which appliances require a power tether may change as power storage and micro-generation technology improves, and as appliance power consumption declines. For example, small displays (e.g. TVs) may initially be tethered given the power consumption of their picture tubes, but become untethered with micro fuel cells and LCD panels. Larger appliances (e.g. large video screens) will be inherently less mobile, and will probably remain tethered even if not required for power reasons (after all it will be always somewhat inconvenient to refill the fuel cells or micro-generators compared to having a grid tether.

In this scenario, cells will be primarily broadcast (cell basestation to mobile) and only a small amount of bandwidth will be necessary for mobile to basestation. (Consider the ratio of what a human can receive to what a human can generate.)

I tend to believe that even basestation computational capability will be limited in nature, even though significantly more than mobile capability, as the bandwidth between basestations will be less than between computation engines next to each other. The only thing that will prevent all the computation engines being colocated will be communication line, latency, and central location reliability issues.

I also don’t see as much use for frequency division multiplexing (FDMA) or fixed TDMA in the future. Wouldn’t you just treat each cell as a big ethernet, letting packet headers decide who gets what? A research issue here is what is the appropriate power consumption tradeoff? At what point does it take too much power to receive B bits per second from M MHz and pull out C bps (C << B). How much receiver power is saved by getting B/X bps from M/X MHz and pulling out C from that? At what point does receiver power become dominant?

To the extent that you use big swatches of spectrum for basestation to mobile transmission, you give enormous peak data rates. I expect that you will be able to download movies to your sunglasses in a matter of minutes. Given the trend in cost of storage, I see mostly disadvantages to trying to send packets at the data rate required by the receiver. For example, if you receive your video feed in realtime, it cannot be shared with others, at least if you require pause capability (a must I think). Once you have pause and sharing, you’re basically talking about buffering large amounts of data on your person. Once you can buffer a two hour movie on your person, it’s only a few more years until you buffer 100 of these. At that point you practically need only receive movies at the rate that they’re created.

I think it is also useful to separate out what is interactive and real-time from other things. Telephony is clearly in this category, as is stock market trading, but television is not. Most television (e.g. movies, entertainment) has almost no timeliness requirements. All such things should be broadcast in off-peak hours (e.g. after midnight) when interactive traffic is the lowest. A few television shows have some timeliness factor (e.g. some news and sports events), but even here, it would be rare to have a delivery requirement that couldn’t tolerate delays of a minutes to hours (note that most news is really entertainment — only a few minutes at most of an hour news program is really timely).

I think there will be a wealth of compression techniques used, especially ones that store very high-level information and turn that into video on demand. Compression will serve as much to stretch whatever mobile storage you can afford (in kg and joules) than to stretch bandwidth. E.g. what are television programs today (e.g. weather) may be transmitted (and stored) as high-level data (e.g. present and forecast temperature, wind, radar, humidity) from and turned into an animated presentation by software in your mobile system.

The one hitch in the above is that I expect things like entertainment (e.g. movies) to become interactive at some point, which moves them from one category to another.

It’s also interesting to consider how advertising fits into this whole picture. Of course, it would be better to eliminate advertising altogether, but that’s not going to happen, so it might as well be considered. It’s clearly silly to rebroadcast advertising over and over again. It too should be downloaded in off-peak hours and mixed in on demand. This is an enabler for something that I think would dramatically help things: the separation of advertising from content creation. Today advertisers are paired with content creators. In a better world they would be separated. Imagine the following: you’re charged for watching radio, television, etc. programs (really that is you’re charged for the key to decrypt these things). If you want to watch for free, you pick an advertising channel to mix in that will pay the cost of the decryption key. Different advertising channels can compete for the least intrusion time into the material you want to watch. Others might deliver advertisements of certain kinds (e.g. if you’re in the market for a new car, you might select a channel that provides such ads). In this world, the content creator (e.g. the news) never interacts with advertisers, and so advertisers wouldn’t exert the influence over media that we have today.