Posts Tagged ‘WSN Requirements’

Good Wireless Sensor Network?

Many people believe that if a good wireless sensor node is built with a good radio, a good wireless sensor network can be built. Although a good sensor node/radio is an important part of the whole wireless sensor network, this is not whole story. The characteristics of a “good” sensor network include: scalability, reliability, responsiveness, mobility, and power efficiency.

A “sensor node” is just a very small part of the wireless sensor network. There are more important design challenges that go into making a network “good.” To build a good wireless sensor network, all of the above factors must work in harmony. The challenge is that this harmony can be difficult to achieve. The complex inter-relationships between these factors is a balance; if these factors are not managed well, the network can suffer from overhead that negates its applicability in the real world.

A. Scalability

Scalability refers to the ability of the network to grow, in terms of the number of nodes, without excessive overhead. This is an important real-world requirement where networks must support more than the small handful of nodes typical in a pilot implementation.

This is due to the network overhead that comes with the increased size of the network. In ad hoc networks, the network is formed without any predetermined topology or shape. Therefore, any node wishing to communicate with other nodes should generate more packets than its data packets – i.e. “control packets” or “network overhead.” As the size of the network grows, more control packets will be needed to find and keep the routing paths. Moreover, as the network size increases, there is higher chance that communication links get broken in communication paths, which will end up with creating more control packets. In a small network, the amount of control packets is almost negligible. But when the network size starts increasing, the overhead increases rapidly. Since the available overall bandwidth is limited, the increase of overhead results in the decrease of usable bandwidth for data transmission. As the network size grows further, there will be very small amount of bandwidth left for application data transmission.

B. Reliability

Reliability is the ability of the network to ensure reliable data transmission in a state of continuous change of network structure. Typically there is an inverse relationship between scalability and reliability in ad hoc wireless networks; as the number of nodes in the network increases, the more difficult it becomes to ensure reliability.

This scalability characteristics of ad hoc networks described above imposes an interesting question on the reliability of the network. Since an ad hoc network is designed to automatically adapt itself to a changing environment or interference, it will issue more control packets when it faces dynamics. More dynamics in the environment will increase the number of control packets and, at some point, the network cannot sustain the amount of overhead caused by the dynamics, which will result in less reliability of data transmission. This breaking point will show up earlier in a large-sized network. So, network scalability and reliability are closely coupled and typically they act against each other.

C. Responsiveness (and Latency)

Responsiveness is the ability of the network to quickly adapt itself to changes in topology. To achieve high responsiveness, an ad hoc network should issue and exchange more control packets, which will naturally result in less scalability and less reliability. In general, the latency of packet delivery in dynamic environment decreases in the network with high responsiveness.

D. Mobility

Mobility refers to the ability of the network to handle mobile nodes and changeable data paths. Generally, a wireless sensor network that includes a number of mobile nodes should have high responsiveness to deal with the mobility. So, it is not easy to design a large scale and highly mobile wireless sensor network.

E. Power Efficiency

Power efficiency – the ability of the network to operate at extremely low power levels – also plays an important role in this complex equation. A typical method for designing a low-power wireless sensor network is to reduce the duty cycle of each node. The drawback is that as the wireless sensor node stays longer in sleep mode to save power, there is less chance that the node can communicate with its neighbors. This will decrease the network responsiveness and may also lower reliability due to the lack of the exchange of control packets and delays in packet delivery. In addition, a more complicated synchronization technique will be necessary to keep more nodes in low duty cycle, which may also affect scalability.

F. Managing the Design Tradeoffs

The complex issue of managing these tradeoffs comes down to how the communication overhead can be minimized while maintaining the network reliability and responsiveness. As explained above, there are many conflicting factors involved in the design of wireless sensor networks, and there are always tradeoffs. When choosing a wireless sensor network for an application, careful consideration of the balance of these factors within the context of the needs of the application is critical.

See Millennial Net‘s MeshScape technology as a good example.

High Performance Wireless Sensor Network

Wireless mesh networks enable numerous embedded applications to be independent of the issues of wiring costs and physical constraints. At the same time, the unique nature of the wireless mesh networks requires many fundamental challenges to be addressed. Among them, robustness and the scalability are the most important issues that need to be addressed. In typical wireless mesh network systems, these two factors generally go against each other due to the self-adjusting and non-hierarchical nature of the networks. For example, as the number of nodes increases, the robustness of network becomes harder to guarantee. On the other hand, to make the network more robust, smaller sized networks are preferred. When the issue of network responsiveness is added on top of these two issues, the equation becomes even more complicated.

Using the wireless mesh networking technology, hundreds or thousands of sensors and actuators can be placed without any wiring constraints. Due to the ad hoc nature of wireless mesh networks, the sensor nodes form a network automatically with minimal human intervention. The network is maintained autonomously, healing itself if any damage occurs to the network. The wireless mesh network is reliable and robust because the network “learns” based on its own changes or problems in the topology and adapts itself to the situation very quickly, even at the individual packet-transmission level. In essence, a wireless mesh network does not assume any predefined topology or placement. The network is designed to adapt itself to them at the initialization, and keep adapting itself continuously.

Much of the technology development in the wireless sensor networking field comes out of the traditional static networking industry (i.e. DSDV, DSR, AODV) that is based on a relatively static environment and stable links. Most of the existing traditional wireless sensor networking technologies such as Zigbee, are based on these assumptions. As a result, these technologies are challenged to address all of the above-named requirements to the level required for production purposes. By contrast, the approach of Persistent Dynamic Routing (PDR) used by MeshScape comes out of the sensor research field that presumes an unstable, rapidly changing environment.

Throughout the next several posts, I will discuss about several important factors that determine the performance of the wireless sensor network. In most cases, these factors have a conflicting impact on each other so that trade-off decisions are unavoidable in network design. Among those factors, scalability, reliability, responsiveness, mobility, and power efficiency will be discussed more in detail. In addition, the advantage of Persistent Dynamic Routing will be discussed and finally, the concept of increasing capacity to improve scalability will be introduced along with several techniques to achieve this.


Get every new post delivered to your Inbox.

%d bloggers like this: