Implementing mesh networking using 802.15.4

One important operational concept of the IEEE 802.15.4 standard, as promoted by the 'Zigbee' consortium but not widely discussed, is its ability to form mesh networks. In other words, each and every 802.15.4 node can act as a dynamic router potentially providing infinite range from a gateway as long as there are nodes in range of one another (check out Ember's Web site: http://www.ember.com/news/press/press-120103.html). This is not a new concept, although making it work as a self-organizing, self-healing network is a challenge.

Is it possible to implement mesh networking using IEEE 802.15.4, and will this similarly be practicable with IEEE 802.16 technology? That is, potentially offering virtually unlimited range for broadband wireless access to the Internet through co-operative inter-working?

    Requires Free Membership to View

    Login

As we deploy more and more wireless networks, utilizing a variety of different technologies, there is a very strong possibility that these networks will eventually be tied together to operate in a 'mesh' environment -- or at least be able to communicate with each other across different boundaries and standards. Right now, cellular technology makes use of the mesh communications framework by allowing users to roam throughout the country and the world by seamlessly hopping from one cell tower to the next. The client device reaches out and grabs the most powerful and available signal to make a wireless connection.

Emerging companies like Cometa Networks are currently investigating ways to build 802.11 (Wi-Fi) networks that overlap and provide a mesh environment to offer wide-scale wireless access. This essentially is the idea behind metropolitan area networks (MANs). The problem is that in order to be operational and reliable, all of the access points in the network must be up and running to provide that wireless bridge on which to hopscotch signals. If one segment is down, then it becomes the weak link and service is compromised -- sort of like one bulb malfunctioning in a string of Christmas lights.

While Zigbee promises to inject wireless capabilities into a variety of devices and previously non-communicative systems -- such as in-building environmental control technologies -- its very nature makes it a poor choice for pervasive mesh networking. The reason is that Zigbee is not designed for constant communications, but to remain dormant until a pre-set time or when some action awakens it. This maintains battery life, allowing these systems to function for extended periods of time. Also, Zigbee is designed to transmit small packets of data, rather than extensive volumes of information.

The same can be said of RFID systems, which are designed to transmit limited amounts of information, and for the most part are passive until readers spark them into life. We will eventually have more and or active RFID systems installed, but they will most likely be used for targeted retail and transportation applications.

The wireless technology that is best positioned for the mesh movement is 802.11 and the various flavors of this technology. There are already tens of thousands of Wi-Fi hot spots available throughout the U.S. and the world. More advanced and faster variations are already under development and will surface sometime within the next year or so. In addition, companies like Intel are dabbling with embedded wireless systems that can function as wireless access points, which means that every system equipped with such technology can also function as a hand-off bridge to wider-area networking.

Management, control, and billing are problems, although these will be solved. We can already monitor and control RF, so it is only a matter of time before we can apply layered technology and algorithms to manage the flow from Point A to extended Points Z. These mesh networks will communicate and exchange data with the Zigbees and the RFIDs of the world, but 802.11 will be the general, and for now favored 'mesh mosh pit.'

This was first published in December 2003

Back to top