Introduction to the challenges of low-power RF mesh networks

LPW mesh networking basically uses a peer-to-peer (point-to-point-to-point) topology to offer significant advantages over conventional networks, which are typically built around server-centric (point-to-multipoint) topologies.

In effect, an LPW mesh turns every node in the network into an intelligent participant in sending, receiving and routing network traffic. The mesh approach eliminates the high costs associated with installing and maintaining central servers as well as overcoming the single-point-of-failure and potential choke-point performance reliability risks.

In a mesh, network maintainability also is simplified because the self-healing network can be extended and/or adapted around physical obstacles simply by adding more nodes into the mesh.

Click here for Figure 1
Figure 1: Adding nodes to a self-healing network extends reach.

Initially deployed for a variety of industrial automation, control and sensor monitoring applications, LPW mesh networking proved to be a superior approach for many situations where environmental factors, RF interference and/or physical obstacles can make it impractical to use conventional server-centric networks.

LPW meshes have also proven useful in applications in automotive, telecommunications, retail and healthcare and are now poised to empower a wide range of home control and consumer applications, such as residential lighting control, energy management, security systems and whole-home integration applications.

While embedding low-cost, self-configuring LPW mesh networking capabilities into a host of ordinary devices, such as lighting systems, thermostats, and home appliances offers great potential for enhancing home networking, designers of these devices must have a solid set of interoperable and reliable implementation solutions in order to succeed.

Following the evolutionary pattern of most new wireless technology developments, the initial focus has been on the development of the underlying silicon functionality. For LPW RF mesh networks, this has involved the development of RF radios that comply with the IEEE 802.15.4 standard and the associated microcontrollers for managing the node.

In addition to 802.15.4, a foundational set of standards have been developed by the ZigBee Alliance industry consortium, which was formed in 2003 with the objective of fostering interoperability for emerging Low Power RF Mesh applications. With over 300 companies participating in the ZigBee Alliance, the organization's scope encompasses the full ecosystem from silicon companies, software companies, integrators and customers.

While the implementation hardware choices have become pretty straightforward, software has now moved to the forefront as the most important factor for success in implementing LPW RF mesh networks.

In addition, the fact that LPW mesh networking is being deployed into traditionally non-networked devices, such as lighting and appliances, means that the product designers have limited resources and/or knowledge regarding networking. This makes the need for complete software solutions even more critical, since designing software for mesh networks requires very specialized skills and expertise.

In order to assure both immediate ZigBee compliance and the flexibility to extend and adapt their product offerings for special and/or future requirements, developers need to carefully consider the role that their software choices will play in their overall product architecture.

Besides providing the basic functionality and standards-compliance, in the world of LPW meshes, software also plays a vital role in delivering the "three R's" of Robustness (resisting failure), Resilience (recovering from hard/soft failures), and Reliability (continuous operation).

In industrial applications where RF environments tend to get very noisy (and increasingly in other environments as well), the ability of the software to provide 3R quality is the major factor for success or failure.

While functional quality is the basic requirement, reliability which is another facet of quality, is absolutely essential for the deployment and commissioning of a wireless network product. The simple fact of life in any wireless networking environment is that there will always be transmission interruptions and hardware failures. In mesh networks, where each node must act as a full participant (sending, receiving and routing data traffic), the software challenge involves much more than just implementing a protocol stack that is simply controlled by the hardware.

Wireless network products, if not designed for reliability, have a higher probability of malfunctioning. Such a malfunction caused by a single device has the potential to bring down a large network partly or entirely.

Furthermore, a failure of the device/network will not only cause downtime issues but can have other damaging consequences beyond just the network. For example, failure of a wireless sensor device monitoring a critical pressure parameter of an industrial boiler can result in a disaster.

Total software reliability cannot ever be treated as an optional objective. It is a mandatory requirement, which has to be built into the design right from the very beginning and carried all the way through the implementation process.

For LPW RF mesh networking, software represents the essential foundation upon which the product and network's overall functional quality needs to be built.

Some of the key challenges/problems faced by a device in an LPW RF mesh network are outlined in the following table:

Click here for Table1
Table 1: Challenges faced by mesh network devices.

When making the underlying hardware selections, it is also important to consider the requirements of the overall software architecture. For example, the selected processor's horsepower and memory capacity must be adequate for handling the reliability functions built into the software protocol stack.

These "hidden" requirements, which are absolutely critical for successful commissioning of mesh networks, are sometimes overlooked or ignored by designers and device manufacturers, in fear that they may result in increased part-cost of the microcontroller or associated hardware circuits. However, it's critical to remember that reliability is not an option.

In order to assure maximum reliability and the flexibility to optimize hardware choices, the software approach needs to provide a complete operating environment that is tailored for LPW mesh networking implementations.

Just as the operating systems used in computer platforms are intended to both control the hardware environment and insulate it from externally induced failures, the "operating software" on a low power mesh node must perform a similar role.

If a flawed or limited software design exposes the node to adverse external conditions, such as RF interference, and also does not provide mechanisms for graceful recovery, then the node and ultimately the entire network are compromised.

The only rational approach lies in treating the software as the most important element of the LPW mesh networking node design and making software choices that combine compliance with ZigBee standards, the flexibility to support a variety of underlying hardware choices and a comprehensive operating environment designed for robustness, resilience and reliability.

About the author
Ram Satagopan is CTO and vice president, Technology at Airbee Wireless. He specializes in the concept-to-commercialization of products of advanced wireless technologies. Ram received a Bachelor of Science degree in Electronics Engineering from Annamalai University, India and a Master of Science degree in Management of Technology from National University of Singapore. He can be reached at rsatagopan@airbeemail.com.