Bugzilla – Bug 1851
WifiRadioEnergyModel energy consumption values are taken from a 802.15.4 chip
Last modified: 2014-09-07 15:32:26 EDT
The doxygen say: /** * \ingroup energy * \brief A WiFi radio energy model. * [..] * * Default values for power consumption are based on CC2420 radio chip, with * supply voltage as 2.5V and currents as 17.4 mA (TX), 18.8 mA (RX), 20 uA * (sleep) and 426 uA (idle). * */ As a matter of fact, the CC2420 Data Sheet is confirming it: http://www.ti.com/product/cc2420 - page 13 The problem is: the CC2420 isn't a WiFi chipset, it's 802.15.4 ! Totally different power consumptions. We need a better set of values, possibly with reference values for different Tx power. T.
I worked on energy consumption issues in 802.11 networks using ns-3. The problem is that very different consumption values are observed between different manufacturers. I may provide some typical average values calculated among several datasheet specifications. Could this help you?
Hi Sébastien, definitely it would help. Either, a realistic one (i.e., one from a particular chipset widely used) would help. Cheers (In reply to Sébastien Deronne from comment #1) > I worked on energy consumption issues in 802.11 networks using ns-3. The > problem is that very different consumption values are observed between > different manufacturers. > > I may provide some typical average values calculated among several datasheet > specifications. Could this help you?
Since most Wi-Fi cards are nowadays IEEE 802.11n compliant and that many users have Intel NICs, I propose to use values measured on Intel 5300 a/b/g/n wireless network adapter (3.3V): 100 mW (sleep), 820 mW (idle), 940 mW (Rx) and 1280 mW (Tx). I guess those values are measured at full power, which is 15 dBm for this NIC. Note that those values are reported in Daniel Halperin et al., "Demystifying 802.11n Power Consumption". I also compared those values to other cards, and they are very close to the average value. You also mentioned the impact of TX power levels on power consumption. This is not easy to have those values. But some people showed that we can use a linear relation between TX power (in Watt) and power consumption (in Watt), considering an average efficiency for the power amplifier of 10%.
Great, thanks. Can you provide a patch and a manual clarification ? Thanks, T.
Hello, I am going to implement the support for having different power consumption based on the transmission power. Sébastien, you mentioned that "some people showed that we can use a linear relation between TX power (in Watt) and power consumption (in Watt), considering an average efficiency for the power amplifier of 10%". Could you please provide some reference? Thanks, Stefano
Sorry Tommaso, I had forgotten to reply to your previous comment :/ To reply to Stefano, efficiency of power amplifiers are generally between 10% and 30%. You can refer to those two articles: [1] F. Wang, D.F. Kimball, J.D. Popp, A.H. Yang, D.Y. Lie, P.M. Asbeck, and L.E. Larson. An improved power-added efficiency 19-dbm hybrid envelope elimination and restoration power amplifier for 802.11g wlan applications. IEEE Transactions on Microwave Theory and Techniques, 54(12) :4086–4099, dec. 2006. [2] J. Van Driessche, G. Cantone, W. Eberle, B. Come, and S. Donnay. Transmitter cost/efficiency exploration for 5-ghz wlan. In Radio and Wireless Conference, 2003. RAWCON ’03. Proceedings, pages 35–38, aug. 2003. In the thesis from Yves Josse about Wi-Fi energy efficiency in radio-over-fiber systems, available at http://tel.archives-ouvertes.fr/tel-00790028 (sorry it's in french), an efficiency of 10% has been considered. I will try to provide a patch as soon as possible, but I'm quite short of time to meet some other deadlines.
Stevano Avallone updated this bug here: http://mailman.isi.edu/pipermail/ns-developers/2014-April/011941.html with code review here: https://codereview.appspot.com/57210044/
Fixed in changeset 10883:d919e7194e23