One of the technologies that many observers have pointed to as causing the most change over the next few years is wireless MESH networking. But critics have pointed out the limited success of W-MESH so far. That is due to a number of factors summed up by the old "chicken and egg" scenario. For W-MESH to establish itself it needs mass market acceptance. In fact, mesh networks rely on significant numbers of users for the self-forming, self-dealing networks to have any practical benefit. But W-MESH systems that have been built so far have either been proprietary and expensive or built on top of a hodge-podge of WiFi based equipment that has been difficult for the average user to buy or use. There are several grass roots W-MESH efforts that have tried to make a go of it so far. Notable among these is that of Locustworld.com. Locustworld has a lot of the right ingredients for widespread adoption: 1] Ability to port the same W-MESH software to a number of different WiFi equipment and configurations. 2] Relatively low cost - $300-$500 per W-MESH router. 3] ability to monitor and upgrade the software
via the network. But the stuff still isn't something the average user can buy off the shelf and load up with software and then use. It still takes hacking together the router box, loading the software, configuring the system . . . rather archaic steps to get to a useful appliance for the non-hacker. And to gain widespread acceptance W-MESH needs to be "idiot proof".
Another problem has been that there is no official W-MESH standard. And so far there has not been a fertile field for a de facto standard to emerge.
But this all may be ready to change. A common requirement for W-MESH to establish itself is for low cost, standards based equipment that can host the W-MESH routing software to become widespread. The major problem with using current WiFi equipment as a base for W-MESH, ignoring issues of range and channel limitations, has been that the routing processor has been external to the wireless radio. Systems have either had to be fashioned using a proprietary router architecture such as an ORiNOCO RG-1000 or by using WiFi cards in a host computer, such as an Epia single-board, fanless computer or a Sokris (or similar) board. Those platforms can run standard Linux kernels to host the W-MESH software. But each implementation takes some hacking and the result is often less than optimal: power consumption is too high for solar battery power, too large for some installations and not able to operate at temperature extremes found in outside environments. Systems can be hacked that work but all that hacking is not what leads to mass market acceptance.
COMMON WIFI+ ARCHITECTURE IS EMERGING
Recently the trend has emerged for WiFi and WiMAX chip designs to incorporate common on-chip elements that can enable W-MESH without the need for external host computers or circuitry except, perhaps, some external SRAM or Flash memory. This architecture is comprised of an ARM processor for which small Linux kernel OS is available, some on-chip SRAM and.or Flash memory, external memory I/O, Ethernet, PCI and USB controllers, and, of course, the 802.11 radio. Whether this is a one or two chip design is mostly irrelevant as long as the device form factor, range and power consumption are competitive. Another common element to emerge is the inclusion of on-chip security encryption (DES plus other algorithm). This makes it possible to keep data encrypted throughout the W-MESH until it reaches it's final destination. Because it is encrypted within the chip itself, it becomes almost impossible to hack.
The latest WiFi chip designs from TI, BroadCom, Intersil, Atheros and others fit this common W-MESH enabling architecture. Designs for WiMAX chips from Fujitsu also fit this architecture. I don't know enough details about Intel's designs. There is a good chance that they won't use an ARM processor but also a good chance that they will use an (Intel) embedded processor and ensure that a Linus kernel is available to run on it.
The result of these new chips will soon be a proliferation of wireless devices in various form factors that can be programmed with common W-MESH software (firmware). The big difference is that this will be self-contained in the standard WiFi devices. Of course to do multi-point routers or systems that use independent sector antennas, either multiple cards or devices will have to be used. In some cases that goes back to the need for a host computer. And certainly there will be the need for host routers or PCs for network management, security and control. But the equation is shifting for the mass market of end users: When the user can be downloaded the W-MESH router software via the network and the system easily configured, or better yet, when the user can simply buy W-MESH enabled cards and boxes off the shelf is about to be a real possibility.
W-MESH combined with broadened spectrum at 5 GHz and multi-mode designs able to use spectrum at 2.4 GHz and 5 GHz may become a "next big thing" while almost nobody is looking. |
