Inside Mesh Networks

Part 1: Ad-hoc wireless mesh networks will be the great enabler for future devices.


Mesh networks could revolutionize communications in the future. Independent of the Internet we know today, wireless mesh networks (WMN) allow both ad-hoc and fixed wireless “nodes” to form a communications net that can become a very powerful information sharing hub.

The idea is that all devices, both user-controlled and autonomous, would be open to act as relay points for the transmission of user and sensor data. Originally developed for military use, this approach has since migrated to the commercial world, and even to underground programs for communities that cannot afford to pay, or that have repressive political regimes, for Internet and peer-to-peer connections. Mesh networks also can play a vital role during natural disasters where a typical network of wired and wireless connection may be compromised.

One example of this is Havana, Cuba. Because of both government repression and high cost, the Internet is a luxury there. So residents of that city have created what is called the Havana StreetNet (SNet). To become part of this “underground” wireless network, all one needs is a computer, a high-power Wi-Fi router and some cabling.

Components of a mesh network. Source: Meraka Institute

While such a network can, of course, connect to the Internet, it also can just act as a local network, sharing whatever data is desired without Internet connectivity (see Figure 1). This approach isn’t limited to small areas, either. Such networks can be area wide, including entire cities, if desired. All they need are enough wireless hotspots, and/or traditional Wi-Fi APs.

How it works
In reality, wireless mesh networks are nothing more than a bunch of radio transmitters set to function as wireless routers. For this discussion, Wi-Fi will be the chosen platform, although theoretically any platform, from cellular to Bluetooth, to Zigbee to Ant, can be made into a mesh network.

“However, if the networks are made up of dissimilar devices and platforms this can lead to some very challenging issues,” said Steven Woo, vice president of enterprise solutions technology at Rambus.

Fig. 1: Peer-to-peer mesh network. Source: Engineering Radio

Wi-Fi devices operate in one of the 802.XX modes (a, b, g, etc.), just as any other Wi-Fi network. Figure 2 shows one of these radios.

The difference in mesh nets is node programming. There is special code programmed into them to tell them how to interact with both the wireless infrastructure, and each other. Essentially this code enables a dynamic routing type of node hopping platform, where the best path is automatically determined and used without user intervention. This is different from standard Wi-Fi networks where there is a server and many clients. In this case, the radio nodes act as both server and/or client depending on whether the node is simply a relay point (as in Figure2), or is a computer to which someone is attached (as in Figure 1).

Fig. 2: Radio node. Source: Motorola.

The major advantage of wireless mesh networks is that they are 100% wireless. In traditional wireless networks, almost all access points are connected to the Internet and only wirelessly connect to the immediate set of clients nearby. The connection is almost always fiber or copper and done via Ethernet.

True wireless mesh networks need only one access point, anywhere in the network to have connectivity to the Internet if desired. That node can share its Internet connection with all of the nodes in the mesh network. Recall that in wireless mesh networks, connectivity to the Internet is optional. If it doesn’t exist, then the communication remains within the mesh network, and only among peers. However, if one of the nodes is interconnected to the internet, then any node in the mesh network can have access to the Internet thought that interconnected node. This will become an extremely important metric with the IoE since many of the devices on the IoE will have only wireless connectivity. And not directly to the IoE.

There are a number of advantages to mesh networks. The first is advances in digital technology. This technology stratifies Note in Figure two that there four antennas. This design is used in what is called antenna diversity. Antenna diversity is a methodology that uses advanced technology to sense where the signal is strongest. While this is a very well understood and mature concept, the advent of DSP ratchets it up an order or two in magnitude, with how it can be implemented to give the network the same bump in reliability.

Why it works so well for this application is because signals used in such networks are at least 2 GHz or higher, and the wavelength is relatively short. That way it is easy to connect multiple antennas to a small form-factor radio (a typical home Wi-Fi router with two antennas, for example). For clarity, let’s take a look at the math behind the antenna diversity spacing.

The formula for determining the length of the frequency wave is:


λ= wavelength (meters); f = frequency; and v = speed of light (also noted as c in some instances, and is 299,792,458 meters per second).

So at 2.4 GHz, λ= (2.99792458 X 10^8 m/s)/(2.4 X 10^9 1/s) 125 mm, or just under five inches. That means that multiple antennas can be hooked to a voting receiver at five or more inches apart and they can be viewed as individual, isolated receivers. Because the frequency is so high, it is easy to use full wavelength spacing. The same technique can be applied for λ/2 (half-wave) and λ/4 (quarter wave) spacing with some mechanical (R/L/C) manipulation of the antenna electrical length, which is common at lower frequencies such as the CB band where separation becomes a challenge.

All radios are subject to the multipath phenomenon and antenna diversity is a very elegant way to improve the quality of the link. In a nutshell, it has to do with how the multiple occurrences of the same signal arrives at the receiver at slightly different times. A well-known example of this phenomenon is where one is driving along and listening to the radio and coming to a stop at a light, for example, and the signal either gets noisy or fades. By simply moving a few inches, the signal is restored. If the vehicle’s radio has multiple antennas connected to a voting circuit, the fade would not happen. A common commercial solution can be seen on big rig trucks with dual CB antennas, one on each side of the vehicle. In this case the receiver has a choice of picking the stronger of the signals from different antennas on each side of the cab.

On the technical side, we are only talking microns here, but there are times when a reflected signal, such as one that bounces off of a concrete building, may actually be stronger that the direct line of sight (LOS) due to LOS attenuation characteristics, even if the separation is a few microseconds, or less. The above discussion focused on propagation because it is the most important parameter of wireless transmission technology and is the basis upon which all other wireless technologies depend. Digital technology and diversity will likely be the major enabler for mesh networks, especially when it comes to uptime and throughput.

Another advantage is scalability. Mesh networks can have an unlimited footprint. And such a network has few economic challenges. All it needs, theoretically, is one connection to the infrastructure if desired and enough clients to cover the area. One example of this is in Abu Dhabi, where a mesh network covers more than 3,000 square miles to service more than 1 million electricity and water meters. Another example is in Google’s Mountain View network, with more than 19,000 daily users. Such high-end mesh networks boast performance of greater than 10 Mbps throughput at each mesh router, with less than 1 ms of latency per hop, and the capacity to move more than 1 TB of data daily—and more, eventually.

There are other advantages, as well. Just being wireless is probably the top one. But the ability for mesh technology transmit data different devices simultaneously is a big one, as well. This is particularly significant in high traffic areas and times. If there is a failure in a mesh device, it has negligible effect on the network. The data is simply rerouted to another of the many devices on the network. And expanding the network, due to its scalable nature, can be accomplished without any disruption to the network.

Finally, mesh networks can offer extended range and coverage, have better non-LOS because nodes easily can be added to go around obstacles and there are virtually unlimited alternate propagation paths.

All of this makes mesh nets a very promising technology on the wireless scene. There are a lot of good things about mesh nets, but there also are a number of challenges. The foremost is that wireless and is subject to the same fading, intermodulation, signal propagation characteristics as any other wireless technology. And combining disparaging technologies isn’t always easy.

Mesh nets are just now coming on the scene. The industry has at its disposal an arsenal of digital, radio, and software technology that will enable all sorts of new platforms for the IoE and 5G wireless. Mesh networks are an interesting and elegant solution to one of the wireless’ world biggest challenges, namely how to make IoE things dynamic, intelligent and autonomous.

Part two will focus on how mesh networks can leverage tools and techniques used to combat cyber-attacks in the enterprise and financial institutions, government agencies, and other wireless infrastructures.

Commonly Used Acronyms

AP – Access Point
DSP – Digital Signal Processing
IoE – Internet of Everything
IP – Internet Protocol
LOS – Line of Sight
MANETS – Mobile Ad Hoc Network
PSK – Pre-shared Keys
R/C/L – Resistor/Capacitor/Inductor
TCP – Transmission Control Protocol
UDP – User Datagram Protocol
WMN – Wireless Mesh Network