LAS VEGAS --One of challenges facing attendees at a big trade show like Networld+Interop is finding the handful of nascent, disruptive technologies that are poised for explosive growth once IT departments get hip to their advantages. These technologies are often characterized by a healthy dose of start-up participation coupled with a lack of involvement (or at least shipping products) from traditional heavyweights. One such technology on exhibit at this year's event is wireless mesh technology.

In these days of Wi-Fi, the phrase "wireless" generally conjures up images of untethered client connectivity. However, the principle behind wireless mesh is really nothing of the sort.

In the same way that a satellite network through which a typical phone call might be routed is transparent to the phone (and the caller), mesh networks are transparent to Ethernet-connected devices. In situations where a mesh moves your existing network's edge to new geographic boundaries, one "side" of the wireless mesh can have a connection to your wired infrastructure while the other "side" of the mesh has the Ethernet-cabled devices that need access to that infrastructure.

Picture a physical Ethernet switch that's been deconstructed in a way that the wires on the backside of each RJ-45 port are disconnected from the switch's backplane and then reconnected to a radio (one radio per port) instead. Then the ports (and their radios) can be located wherever you like, and the radio network between all the radios (the mesh) is what takes the place of the original switch's backplane. As long as each radio can make contact with at least one other radio, the ports associated with those radios form the basis of an Ethernet network to which any Ethernet device can be connected.

Given this architecture, wireless mesh is primarily a new enabler for Ethernet deployments in two scenarios.

One scenario is where wireless technology is used to push the edge of an existing Ethernet to physical locations that, for whatever reasons, didn't make sense to pull wire to before. For example, picture publicly accessible traffic cameras on every light post along a highway that leads into a metropolitan area. With a mesh network, each light post gets a radio and an Ethernet port to which an Ethernet-based camera can be connected. The radios along the highway talk to each other and form the basis of an Ethernet backbone, the end of which is connected to a backhaul over which the traffic images are transmitted to image servers and ultimately the Internet. The mesh obviates the need to dig a trench along the highway.

In the second scenario, a particular problem requires rapid deployment of an isolated Ethernet where wiring is impractical. It would be like setting up an Ethernet in the middle of open field where no connection to an existing network is available or required for the particular application. Examples might be the scene of an emergency or a battlefield.

As the term "mesh" implies, radios aren't just daisy-chained together to form the network (although they can be if you want to). As long as they're located within a certain proximity of each other, they form a mesh where every radio will stay in communication with the other radios that are within range. This helps to overcome congestion problems that might be encountered should one "side" of the mesh become congested with traffic. The mesh is smart enough to seek an alternative path around the congestion.

This sort of routing intelligence, and other advanced mesh features such as encryption and auto-configuration, is where the various mesh offerings not only differ from one another, but where they also form the basis of proprietary offerings. In other words, due to lack of mesh standards, mesh hardware from one vendor is unlikely to work with hardware from another vendor. Today, though some heavyweights like Intel have expressed an interest in wireless mesh networking, the offerings come mostly from relative unknowns such as PacketHop, Tropos Networks, MeshNetworks, BelAir Networks, Strix Systems, and Firetide.

Firetide is here at Networld+Interop showing off its indoor and outdoor mesh hardware, respectively known as the HotPoint 1000S and HotPoint 1000R. According to Ike Nassi, Firetide's chief technology officer, the company has customers who originally looked at wired deployments that would have cost upwards of $100,000 and taken weeks to roll-out. Using HotPoints, said Nassi, they trimmed costs by about 70 percent while deploying the network in a matter of days.

Firetide's HotPoints are self-configuring. Based on the ad-hoc networking mode of 802.11b-based Wi-Fi, they can be programmed with SSID's and encryption keys to prevent rogue HotPoints from entering the network. But, for most deployments, said Nassi, "our customers just turn them on and leave them alone. There's nothing else they need to do." In terms of encryption, all traffic within the mesh can be encrypted to prevent packets from being sniffed from the air. But the encryption is transparent to the Ethernet devices. Once the traffic is on the wire, it's plain-old Ethernet. If security is an issue, VPNs are highly recommended.

Firetide claims that, at 200 milliwatts of power, the $2,000 outdoor 1000R has a range of up to two miles. The only requirement to get a HotPoint working is electricity and the source of that can be anything, including a solar cell. Theoretically, one could set up some pretty big mesh networks covering a large geographic area.

On the downside, one challenge, given its reliance on 802.11b technology, could be performance. Not just performance of the mesh, but also performance of any 802.11b access points that are connected to it so that wireless clients can take advantage of the "extended edge." For 802.11b, which operates across three channels, there's only so much spectrum available for wireless communications. Technically speaking, a WLAN could end up competing with a mesh for spectrum which could negatively impact performance.

Also, outside of competing for air with WLANs, Firetide hasn't done any real benchmarking to figure out when and where a mesh might run out of gas. For example, what sorts of applications generate what sorts of traffic? And what is the optimal number of HotPoints? According to Nassi, the company is working on these questions to help customers with their deployments.

In terms of the future, the best thing about mesh networking hardware is that it's radio agnostic. Nothing prohibits mesh hardware providers from using other 802.11 technologies, or even WiMax when it becomes available. Theoretically, the mesh could be based on WiMax and access points connected to the mesh could be 802.11a, b or g, and there would be no competition for airspace.

Long term, if mesh gets any traction, I expect that the key vendors will get swept up by the bigger networking players like Cisco and 3Com.

You can write to me at david.berlind@cnet.com. If you're looking for my commentaries on other IT topics, check the archives.