Relation between mean delay and network size

Question

I am studying the effect of network size on the mean delay per received packet of a wireless sensor network. I am using a duty-cycled asynchronous MAC protocol.The topology is all sources directly send packets to only one destination located at the center. In such case, if the number of sources increases, should delay per packet increase or decrease? If delay per packet decreases with increasing number of nodes, what would be the possible reasons behind this? Would anyone please explain?

Thank you.

Answer 1

In a wireless network, the propagation speed is c₀, so a link's propagation delay is distance / c₀.

The processing delay of a node depends on the node, making it off-topic here.

To complete the delay chain, each (store and forward) buffering node causes a serialization delay, depending on the bit rate. For wireless networks, longer distance usually means lower bit rate, so that delay tends to grow with physical diameter.

Since you're using a shared medium there's also a delay for successfully accessing the medium (also factoring in attempts lost to contention or collisions). Depending on the exact access method (pure CSMA/CA, RTS/CTS, ...), that delay statistically increases with the number of potentially transmitting nodes, caused by more parties actually trying to use the single channel. There's no reason why that delay should decrease with increasing node numbers.

What you describe might reflect that nodes have less opportunity to enter low-power mode and subsequently require less time to wake up (needing to wake up potentially adds to the overall delay, depending on the access method). Note that that is host and configuration specific and explicitly off topic here.