Class Notes #25
Networks for everyone
There are countless thousands of miles of fiber optic cable in the ground and under the oceans connecting all forms of telecommunications including computer networks. The assumption has been that sooner or later broadband network (internet) services will reach every household in industrialized society and that the information age would seep into other societies over time. The reality is that these predictions have not come true for a variety of reaons. First, the cost to deliver broadband service to outlying regions - even in the state of Texas - is simply too high for commercial companies to pursue. The investment needed to provide service to one or two ranches lying several miles apart is far higher than the revenues that would return (without horrendous surcharges). Similarly, there are old inner-city regions in many countries with older phone networks that are difficult and costly to upgrade given the the probably revenue from lower-income communities may be negligible.
One approach adopted by several communities or public organizations is the implementation of "user owned and managed" networks. This is accomplished by an entity purchasing or leasing "dark fiber" (meaning, optical fiber in the ground but not connected to anything) and purchasing the necessary networking hardware to drive (light up) the fiber. It can be clearly shown that over a ten-year investment the savings of user owned and managed networks over standard telco rates are substantial. Furthermore, once a community owns it - they own it forever more or less.
These "civic" actions have led to a series of jurisdictional and ideoligical wars of the type "Why should public money be used to create solutions that that the private sector can do?" and "Why should public money (like school network budgets) be wasted paying for over-priced commercial telco solutions?" Yet many organizations have elected to build their own despite the free-enterprise rhetoric. A recent announcement in the NY Times indicates the NYC intends to build their own "city run" fiber networks. They are hardly the first - the city state of Singapore for example has been fully wired for several years. They have the highest percentage of "fiber to the home" (FTTH) in the world. In Austin I know of no-one with FTTH. In Laval, a suburb of Montreal, the school district put in their own gigabit network. The financial cross-over point for that project was less than five years! Meaning, within 5 years they would be paying less overall than their previous telco costs.
But what about those areas where collective fiber solutions cannot be deployed due to sparsity, geography, infrastructures, etc.? Rather than think of the community network as a "wired" solution, developers are looking at wireless solutions. You have all heard of or used WiFi - here at UT, at Starbuck's, Lava Java - and these are essentially "local wireless networks" or WLANs. The cost to deploy this solution is very low assuming a small number of connections. When the number of users starts to climb these solutions are not all that cheap. Also, at present WiFi is significantly slower than normal ethernet and a LOT slower than dedicated gigabit fiber based solutions.
Where this is headed is to a kind of "grid" or "mesh" solution. A single or at most a few point of access servers can be used to communicate with an array or "mesh" of wireless transceivers. These are simply radios that broadcast signals of the 802.11 class of wireless connections. Today, 802.11g (54 mbits/s max), can provide quite tolerable connection levels even for music and software downloads. Imagine not one 802.11g transceiver but many such devices that can talk to each other thus producing a "mesh" of radio points that can services numerous mobile connections to computers, PDA's and web-ready cell phones.
The need for fast connections in the developing world is less pressing than the need for connections of any kind whatsoever! Wireless mesh networks may be the fastest way to deliver computer connectivity to remote or low-income regions. Combined with satellite services, these plans look very interesting indeed.
Wireless Networks and Mobile Computing
Links:
Wireless Mesh Networks (tech)
Mesh Features and Benefits (Nortel)
Community Mesh Networks (MSoft)
Grid Computing (FAQ); TACCS Grid (summary)