| Hi-index | 0.00 |
We study the scaling capacity of wireless sensor networks in case of multiple gateways are presented. In such a heterogeneous network, many gateways are needed to gather data from the sensor nodes in network. Different from the previously studied single-gateway model, our proposed gateway access model allows multiple gateway nodes being seamlessly connected with high speed wire-line or wide-band radios. Once the data from the sensor node is sent to any gateway node, it can be collaboratively sent to the appropriate sink node among multiple gateways, which could effectively avoid throughput reduction due to multi-hop routing among the sensor node. This mechanism is called ``Destination-hub", which could ensure the bounded time delay and network capacity. However under such a cooperative scheme, two research issues need to be tackled. One is the number of nodes in network and the second is the scaling capacity for each node when multiple gateway nodes are presented. Our proposed model is a fundamental research work for scaling capacity, which could also provide an instructive proposal for future gateway deployment. In this paper, we characterize the scaling capacity between gateways and sensor nodes under different gateway access model. If gateway nodes are randomly deployed, the number is $\Omega(\log n)$, and there are at least $4 \log l$ gateway nodes are needed in each squarelet region so as to keep the gateway connectivity, where $r$ is the transmitting range of sensor nodes, and $l$ is the side length of the deployed square region. And we can conclude that, the gateway nodes can be tightly bounded by $\Theta(\log n)$. For random placement, the scaling capacity for such a scheme is $\Theta(n)$. And we have also proposed an optimized scheme for reducing the number of gateway nodes. Considering the interference Voronoi cell, the reduced gateway nodes can be $2 \lnl$, and still bounded by $\Theta(\log {n})$.