Technically speaking: Will standards mesh?

Oct. 24, 2004
Reliability, scalability and mobility are the trump cards of wireless mesh networks and should result in triumph over other wireless topologies, according to Senior Technical Editor Dan Hebert, PE.
By Dan Heber, PE, Senior Technical EditorParticipants in the wired bus wars are down to a manageable few, and we now move on to the wireless network battleground. Hopefully, early establishment of open standards will come sooner to wireless than wired.

Wireless network standards need to be established for both topology and protocol. There are three basic types of topology for wired and wireless networks. The first is point-to-point and its logical extension daisy-chain. Such topologies connect a single device to another.

The second is the star topology. Star has a central host that communicates to each node. Star topologies are used for most commercial wired Ethernet systems. A wireless star system can replace wiring from a central controller to field devices in an industrial application.

The third and newest is the wireless mesh topology. Mesh topologies form a true peer-to-peer network. Every wireless device in a mesh network is a mini-router that can send, receive and relay information. Mesh topologies offer a host of advantages over the other two  and are likely to emerge the victor in the wireless world.

A combination of point-to-point and star topologies works well for many industrial wired installations and is used by industrial fieldbuses, where a host typically communicates with many devices. Some devices are connected directly to the host, and other devices are connected to the host via a daisy chain network.

The problem with both point-to-point and star for wireless is dependence on a single link, either from device to device or from host to device. Wired links are seldom broken and communicate reliably, but wireless links are inherently more erratic.

A mesh network does not rely on a single link because messages can be automatically routed from one device to another via any of the devices in the network. A mesh network with four nodes would have five different communication paths between any two nodes, and the number of possible communication paths grows quickly as nodes are added.

 Mesh topology fits the wireless world where communication paths are subject to interference, vulnerable to physical obstructions, and limited by distance. If a path is blocked, either permanently or temporarily, an alternate path can be automatically found by using the intelligence embedded in each node. This "self-healing" automatic path search is one of the chief benefits of mesh topologies.

If a path search reveals no viable path, one can be created by simply adding another node. Most mesh networks contain software that automatically recognizes and configures new nodes. This is why mesh topology-based networks are often described as self-configuring. In theory, nodes can be added at will and a mesh network can scale to hundreds or even thousands of nodes.

As one might expect, this all-dancing, all-singing, self-configuring, self-healing, and scalable network must be powered by fiendishly complex software. "The single biggest improvement to mesh networking in recent years is more powerful algorithms to reliably link hundreds and thousands of nodes," says Venkat Bahl, the vice president of marketing for Ember.

"These new algorithms enable mesh networks suitable for large-scale and wide-area applications such as meter reading," continues Bahl.

Node mobility is also easily handled by mesh networks. According to Rick Rotondo, vice president of technical marketing for MeshNetworks, "Mobile mesh networking supports high-speed hand-off and proactive route selection. Mesh networks have been used to communicate with Indy race cars traveling over 230 mph."

Reliability, scalability and mobility are the trump cards of wireless mesh networks and should result in triumph over other wireless topologies. If and when this occurs, the last standardization hurdle to overcome will be protocol.

Most wireless mesh firms, including the aforementioned Ember and MeshNetworks as well as Firetide, have products that support their own proprietary protocol as well as emerging industry standards.

ZigBee, is one such standard. ZigBee Alliance members, including process control heavyweights Honeywell and Invensys, are defining a global specification for wireless applications based on the IEEE 802.15.4 standard. ZigBee takes full advantage of IEEE 802.15.4 and adds logical network, security, and application software.

According to a recent article in Business Week, the U.S. Energy Department has hired Honeywell to test ZigBee at companies such as Alcoa, Dow Chemical, and ExxonMobil to track energy loss from piping systems and to monitor the use of gases in production processes.

With any luck, wireless mesh networks will arrange as a standard topology powered by ZigBee and IEEE 802.15.4. Establishment of a standard will create true competition among vendors, drive prices down, and allow implementation of multi-vendor mesh networks.