Battery consumption comparison between MANET routing protocols

[rockjaz...@gmail.com]

Hello dear all.

I’m working on a comparison between routing protocols (AODV, DSDV, DSR, OLSR) in a MANET.... Pretty much standard stuff (PDR, mean delay, throughput… you name it). While I’ve been getting results comparable with existing literature on those measurements, I found something about battery consumption that doesn’t add up (please see attached picture)

To begin with. I’m working on ns-3.26, I’ve used the basic energy model in my nodes and the TotalEnergy trace function to measure battery consumption with 802.11b.

I get that the battery consumption in DSR is a reflection of the way it operates at different network densities (the three heuristics...) but what I don’t get is why AODV is the one that consumes the most energy since it’s also a reactive routing protocol? It consumes even more resources than both proactive protocols.

I’ve double checked all my work (both calmed and hysterically) to make sure there are no "silly" mistakes… I’m sorry I can’t upload my ns-3 code right now since it’s an ongoing investigation… But I’d greatly appreciate your feedback on this issue to help me understand this behavior.

Caleb V.

[Konstantinos]

Hi,

By default in ns-3, AODV has enabled HELLO messages, so that might be the reason. So it is not 100% re-active. There is some pro-active part (not the same information as in OLSR through)

Try disabling HELLO's and you should get a similar trend as DSR.

EnableHello: Indicates whether a hello messages enable.

• Set with class: BooleanValue
• Underlying type: bool
• Initial value: true
• Flags: construct write read

Regards

K

[rockjaz...@gmail.com]

Thanks so much for your fast response Konstantinos... I'd definitely check on that and come back to show what happened. 

Caleb V.

[MsOne Handly]

Hi,

Did you manage with the code and is it possible for you to share it? I'm new to ns3 and I'm struggling to compare the routing protocols based on energy consumption. Kindly assist.