Intel Developer Forum had traditionally closed with sneak peeks into Intels research efforts. Were not talking a specific product area — i.e., researching 90nm technology in order to build a CPU. Rather, Intel unveils research initiatives that are more far-flung. Some are clearly efforts to develop additional business opportunities, while others simply explore possibilities that might or might not lead to additional products.
Pat Gelsinger, Intel VP and Chief Technology Officer delivered the keynote address that kicked off research day. Gelsingers general theme was the leap from what he called “microconvergence” to “macroconvergence”. Microconvergence is simply a restatement of the overall theme of this IDF — melding computing and communications. Macroconvergence is taking these technologies and applying them to real-world applications areas beyond those traditionally thought of as computing applications — in other words, the convergence between disparate industries.
One key example of this is healthcare. One interesting demographic statistic, as noted in Mark Hachmans news piece on the keynote (http://www.extremetech.com/article2/0,3973,899835,00.asp), is that 30% of the population worldwide would be over 60 by the year 2050. However, hitting closer to home, another point made was that many baby boomers would be spending more money and time taking care of aging parents than they would spend putting their kids through college. By the same token, senior citizens want to avoid nursing homes and maintain their independence for as long as possible.
To this end, pervasive, context-aware, location-aware computing and networking might play key roles in this arena. Gelsinger demonstrated a slightly different twist on the “living room of the future”. This version had sensors placed in televisions, telephones and furniture (i.e., the couch). In addition, infrared cameras placed such that stereoscopic images could be generated of people walking in the area were mounted near the “ceiling”.
The idea is not to invade the privacy of the people living there, but to monitor their health and well-being. One example was the ability for a smart application using stochastic pattern recognition to detect when someone changes their gait significantly, perhaps indicating a physical problem just beginning to crop up before anyone is aware of it.
Of course, privacy is an issue, though Intel dealt with it rather glibly — is it better to live with all these sensors monitoring you, or live in a nursing home? Of course, its not a binary choice by any means, and care will need to be taken to assure security and privacy issues. After all, you could postulate a situation where sensors report an emerging cancer in your body, and your company canceling your health insurance before your aware of the situation. In the end, though, such pervasive communications and computing technologies could reduce healthcare costs and improve quality of life if properly handled.
Wireless Mesh Networking
We also attended a highly intriguing session on wireless mesh networking, presented by Steve Conner, a senior network software engineer from Intel.
Conner related a personal experience, in which his own home wireless networking has dead zones within a single room, where his wireless internet connection loses connection. All of us who have used Wi-Fi networks have experienced these types of dropouts. The potential sources of interference are substantial.
One potential solution is the mesh network. Instead of a central access point, which can act as a bottleneck, Conner postulates an environment where all of the networking connections (including those in the clients, such as a PC or handheld computer) can act as routers. One concept shown were tiny “micro-routers” the size of a nightlight that could easily be plugged into household power outlets and left in place. Each node in the mesh network only has to talk to the closest adjacent nodes; routing protocols decide the best route through the network to deliver data.
The advantages of this approach are varied:
- More reliable connections, and hence, reliable behavior in the network.
- Lower power utilization — the clients dont have to radiate large amounts of energy to reach the access point, and vice-versa.
- Potential interference happens on a more local scale, and the mesh can re-route around it, offering greater redundancy.
However, there are still key problems to be addressed. One issue is security. If you have a mesh network, and so does your neighbor, then the potential exists for either snooping into private data or simply hogging the available bandwidth of a neighboring, but alien, node.
Existing 802.11 products can actually be leveraged to create mesh networks at the MAC level, but you cant simply use standard Ethernet protocols. You have to use better routing algorithms. Also, the problem of QoS (quality of service) once again rears its head. After all, if youre blasting a video image from home theater system to your laptop PC, the last thing you want is dropouts in video or audio data.
In the end, wireless mesh networks could solve a number of problems that extend beyond simply eliminating dead zones in your home. For example, one key problem has been getting broadband access from the big backbones to homes that can use it — the “last mile” problem. If one home can connect to the backbone, a wireless mesh could extend access to the entire neighborhood. Another scenario is the corporate environment, where a mesh network could extend the reach of a wired network, without needing additional access points that have to tie into the wired infrastructure.
ExtremeTechs latest IDF Coverage