A Consumer Broadband goal: Data on Virtual Circuits which have QoS passes seamlessly from remote servers through the Wide Area Telecomms Network to Broadband Appliances in homes.
This project is part of a collaborative effort with co-workers at AT&T's Florham Park Lab. in New Jersey. The plan is to devise low cost high speed broadband networks to the home. The Florham Park Lab has developed chipsets for low cost high speed ATM switches and has built some prototype broadband appliances. We have prototyped and programmed ethernet virtual circuit switches and are contributing to the programming of the ATM switch and switch controller. We are also trialling the integration of this new wired technology with new fixed wireless access (FWA) systems. The FWA system will be viable for high speed broadband connections before fibre systems are available and will be able to fill in areas where it is difficult or expensive to lay fibre.
In the future many homes will have a high speed broadband connection to the rest of the world at rates in excess of 100 Mbps. This is beyond the limit of current copper cable technology.
An obvious future-proof physical connection is fibre, but it has long been thought that the economics of fibre-to-the-home would put this out of reach for decades. The economic arguments rest on the high price of the very large number of fibre-interface kerbside switches that would be required in the periphery of the Access Network and the very high cost of laying new fibre everywhere. Recently AT&T Labs have demonstrated inexpensive switches with fibre interfaces. This and other factors lead us to believe that these economic barriers will be overcome and high speed broadband to the home will soon become economically viable.
The AT&T Low Cost Edge Node is a 155 Mbps ATM switch with fibre interfaces. It has been made inexpensive by using a chip set originally developed in AT&T Labs and now marketed by IDT. The main factor influencing the low cost of the device is its use of DRAM technology to queue and switch complete ATM cells. This reduces the price to a point where the fibre interfaces dominate the total component cost. A 155 Mbps peripheral Access Network could be built from these switches, interfacing to faster (but fewer) switches at a higher level in the Access Network.
Ethernet is a de facto LAN standard and early adopters are already installing wired and wireless ethernet systems in their homes. It does have many advantages over, for example, ATM as a home standard - most importantly it is available, cheap and simple. However, plain ethernet has no support for Quality of Service (QoS). This means that all data connections are treated as having the same priority, and multimedia streams such as video-on-demand, audio-on-demand or phonecalls will be interrupted by other data transfer activity.
To overcome these limitations we have developed the concept of Ethernet Virtual Circuits (EVC). EVCs allow end-to-end virtual circuit connectivity with multiple levels of QoS across mixtures of ATM and ethernet networks - as long as the ethernet networks use EVC-aware switches and EVC-aware appliances. Importantly, EVC networks are compatible with plain ethernet, although the advantages of using virtual circuits would not be available on the plain ethernet section of the network.
At present we have built and programmed a number of prototype EVC switches in Cambridge, some of which have been shipped for evaluation to our co-workers at AT&T's Florham Park, New Jersey, Labs.
Fixed Wireless Access systems have now been developed to the point where they can provide high speed broadband connectivity to the home. In Cambridge we have been trialling an FWA system made by Cambridge Broadband Ltd. The key designers in this company formed one of our spin-offs several years ago, they had previously developed our Broadband Radio ATM Project. The system provides aggregate bandwidth of 60 Mbps bidirectionally within each base station quadrant. Non line of site coverage works at up to 5 Kms. from a base station. The current design is aimed at the European 3.4-3.6 GHz bands.
Our interest in this technology is three-fold, First, it lets us leapfrog the development of fibre broadband to the home by providing us with fast connectivity now. Secondly, being ATM based, it allows for QoS through multiple levels of priority over the air. Finally, wireless technology will inevitably form part of a complete broadband strategy to reach destinations which are difficult or too expensive to fibre.