CoolSpots

1. CoolSpots Yuvraj Agarwal, CSE, UCSD Trevor Pering, Intel Research Rajesh Gupta, CSE, UCSDRoy Want, Intel Research

2. Motivation: Wireless Power Is a Problem!

3. Many devices already have multiple wireless interfaces… PDA’s HP h6300 (GSM/GPRS, BT, 802.11) Mobile Phones - Motorola CN620 (BT, 802.11, GSM) Laptops (Wi-Fi, BT, GSM, …) Opportunity: Devices With Multiple Radios Leverage these already present multiple radios on commodity devices to provide a similar user experience while saving energy and increasing battery lifetimeLeverage these already present multiple radios on commodity devices to provide a similar user experience while saving energy and increasing battery lifetime

4. Properties of Common Radio Standards

5. Low-power Access Within a WiFi Hot-spot

7. Inter/Intra Technology Power Management Add power numbers! Add power numbers!

8. CoolSpots Network Architecture

9. Switching Overview Three main components contribute to the behavior of a multi-radio system: where, what, and when Position: Where you are Need to address the difference in range between Bluetooth and WiFi Benchmarks: What you are doing Application traffic patterns greatly affect underlying policies Policies: When to switch interfaces A non-intrusive way to tell which interface to use

10. Where: Position Bluetooth and WiFi have very different operating ranges! (approx. 10m vs. 100m) Optimal switching point will depend on exact operating conditions, not just range Experiments and (effective) policies will measure and take into account a variety of operating conditions

11. What: Benchmarks

12. When: Policies The switching policy determines how the system will react under different operating conditions

13. Experimental Setup Characterize power for WiFi and BT Multiple Policies Different locations Suite of benchmark applications Stargate research platform 400Mhz processor, 64MB RAM, Linux Allows detailed power measurement Tested using “today’s” wireless: WiFi is NetGear MA701 CF card Bluetooth is a CSR BlueCore3 module Use the geometric mean to combine benchmarks into an aggregate result Moved devices around on a cart to vary channel characteristics Now in order to quantify the benefits of using CoolSpots we wanted to characterize the power for each of the wireless interfaces – namely Bluetooth and WiFi. This had to be done for the multiple policies , at different locations and for each of the benchmark applications. We instrumented the Stargate platform with current sense resistors in order to measure instantaneous power consumption for WiFi and BT ! Mention that for the range experiments we put the device on a cart and moved it around at different locations.  Benchmarks suite  - idle  - 2 - File Transfer traces  - 2 - web traffic traces  - 3 - media streams at different bit rates 

14. Switching Example: MPEG4 streaming This is just an illustration of the switching behaviour of the bandwidth only policy of CoolSpots. We chose an MPEG-4 streaming application which starts off using WiFi and then switches to BT.  Note in this case when the application has buffered enough data towards the end the system switches to Bluetooth to maintain the trickle bandwidth ! WiFi in this case would still consume considerable power ! This would be the case even if using WiFi in PSM mode

15. Results Overview (Intermediate Location) This chart gives an overview of the energy saving benefits of a selction of CoolSpots Policies at an intermediate locationacross all benchmarks ! All policies do better than Wifi in CAM and the CoolSpot polices do much better than WiFi in PSM mode … Mention the latency – is the increase in total benchmark time Bluetooth -- the increase in application latency leads to reduced user experience … (How can we word this ?)These policies as we show next are also affected by operating distance/range with some of them not performing as well at differnet ranges ..(next slide) This chart gives an overview of the energy saving benefits of a selction of CoolSpots Policies at an intermediate locationacross all benchmarks !

16. Impact of Range/Distance In order to evaluate the effects of range we set up three location configurations (1,2 LOS and 3 NLOS) The Bandwidth policies save substantial power, however at longer ranges at least one benchmarks starts to fail … Cap-static does better, but that too starts to fail at Location 3 configuration. Cap-dynamic which dynamically remembers the last seen location B/W does well and does not fail (and thus is relatively robust), while maintaining a low energy bound. As expected bluetooth-only gives lowest energy but benchmarks fail as well as it increases latency as described earlier. In order to evaluate the effects of range we set up three location configurations (1,2 LOS and 3 NLOS) The Bandwidth policies save substantial power, however at longer ranges at least one benchmarks starts to fail … Cap-static does better, but that too starts to fail at Location 3 configuration. Cap-dynamic which dynamically remembers the last seen location B/W does well and does not fail (and thus is relatively robust), while maintaining a low energy bound. As expected bluetooth-only gives lowest energy but benchmarks fail as well as it increases latency as described earlier.

17. Results across various benchmarks This chart illustrates the energy breakdown for the benchmarks for some representative policies. We have chosen an intermediate location where none of the policies fail. Results are normalized to WiFi in Cam mode (100%) Talk about the switching overhead amortizes as seen from transfer-1 and transfer-2. This chart illustrates the energy breakdown for the benchmarks for some representative policies. We have chosen an intermediate location where none of the policies fail. Results are normalized to WiFi in Cam mode (100%)

18. Cap-Dynamic Switching Policy Switch up based on measured channel capacity (ping time > Y) Remember last seen Bluetooth bandwidth (Z=kbps) Switch down based on remembered bandwidth (kbps < Z) Add about both swith-up and switch down having some hysteresis. Add about both swith-up and switch down having some hysteresis.

19. Switching Policies – Analysis “Wifi-Fixed” Policy (WiFi in Power Save Mode) Works best for as-fast-as-you-can data transfer Higher power consumption, especially idle power “Blue-Fixed” Policy Very low idle power consumption Increases total application latency, fails at longer ranges “Bandwidth” Policy Static coded bandwidth thresholds, fails to adapt at longer ranges Switches too soon (bandwidth-0) or switches too late (bandwidth-50) “Capacity-Static” Policy Estimates channel capacity and uses that to switch up Fails at longer ranges due to incorrect switch-down point “Capacity-Dynamic” Policy Dynamic policy, remembers the last seem switch-up bandwidth Performs well across all benchmarks and location configurations! In the light of the results present we can now analyze why certain policies do better that the others. In the light of the results present we can now analyze why certain policies do better that the others.

20. Conclusions A dynamic system can leverage the different underlying radio characteristics to reduce communication energy while still maintaining good performance Advanced policies can adapt well to changing operating conditions Application behavior Radio link quality Evaluation of CoolSpots policies shows around a 50% reduction in energy consumption over the present power management scheme in WiFi (PSM) across a range of situations Presented detailed experiments and evaluation of various “policies” For multiple applications, at different ranges  Implemented the CoolSpots switching framework No change to the application themselves !  Coolspots leverages the good features of these “heterogeneous” radios Presented detailed experiments and evaluation of various “policies” For multiple applications, at different ranges

21. Thank you! Questions?