1. How Green is IP-Telephony? Salman Abdul Baset*, Joshua Reich*, Jan Janak**, Pavel Kasparek**, Vishal Misra*, Dan Rubenstein*, Henning Schulzrinne* Department of Computer Science, Columbia University* Tekelec Corporation**

2. 2 Traditional Telephony Place call (Signaling) Directory lookup Circuit reservation Talk (Connectivity) Transfer voice data (analog, digital) Variations on these themes Multi-party conferencing Voicemail

3. 3 IP-Based Communication Systems Place call (Signaling) Directory lookup Packet switched routes Talk (Connectivity) Direct packet routing Media relaying PSTN/mobile gateways Telephony And More Video IM Status / buddy list

4. Trend away from traditional telephony infrastructure Vonage, Packet8, Verizon FiOS Skype Mobile 4 Trends & Implications To single infrastructure (IP) for all data/voice/etc. More efficient (one system to maintain, improve) Much less expensive (for now) More fragile (one system to fail) More complex So what does this mean energy wise?

5. 5 Our Questions 1)Where is energy consumed? 2)How do different design choices effect energy consumption? 3)How we can make IP-telephony more energy efficient?

6. 6 Outline IP-Telephony and power consumption Answering our questions: 1)Where is energy consumed? 2)How do different design choices effect energy consumption? 3)How we can make IP-telephony more energy efficient? Conclusion & Future Work

7. 7 Outline IP-Telephony and power consumption Answering our questions: 1)Where is energy consumed? 2)How do different design choices effect energy consumption? 3)How we can make IP-telephony more energy efficient? Conclusion & Future Work

8. 8 IP Communication Flavors Client-Server (C/S)Peer-to-Peer (P2P)

9. 9 IP Communication Flavors Traditional Telephony Replacement Communication Addendum

10. 10 How Does C/S IP-Telephony Work? SIP registrar / proxy server REGISTER (ip addr) REGISTER (ip addr) User agent (1) signaling (2) media (voice, video, IM) SIP registrar / proxy / presence server Utopian Internet No NATs or firewalls IP-PSTN gateway

11. 11 And In The Real-World… SIP registrar / proxy / presence / server User agent media server NAT / firewall

12. 12 Media Servers Bypass Firewalls SIP registrar / proxy / presence / server User agents (1) signaling (2) media (voice, video, IM) (UDP or TCP) media server NAT / firewall IP-PSTN gateway

13. 13 How Does P2P IP-Telephony Work? P2P / PSTN gateway NAT / firewall network address of node B? (3) signaling (4) media network address of node E? (2) signaling (3) media (TCP) node C node B media relay (or relay) node A node D node E (1) (2) node = user agent nodes form an overlay share responsibilities for message routing, signaling, media relaying super nodes, ordinary nodes (1) (2) (1)

14. 14 Sources of Energy Consumption End-point –Handsets –VoIP conversion boxes –PCs Core –Signaling / directory –Media relaying –PSTN / mobile gateways Network

15. 15 Assessing Energy Consumption Data (from C/S VoIP provider) –100 K users (mostly business) –15 calls per second (CPS) –~5K calls in system –NAT keep-alive traffic –All calls relayed Modeling –C/S –P2P Power Meters –Wattsup –Killawatt Hardware Measurements –SIP Server –Relay Server –Desktop clients –Laptop clients –Hardware SIP phones –Software phones –Skype peers

16. 16 Outline IP-Telephony and power consumption Answering our questions: 1)Where is energy consumed? 2)How do different design choices effect energy consumption? 3)How we can make IP-telephony more energy efficient? Conclusion & Future Work

17. 17 Where is Energy Consumed? PSTN replacement VoIP servers consume less than 0.04% of total! – >10K users, voice traffic –a server can handle signaling workload for 500k users –a server can handle media workload for 50k users –even after a redundancy factor of 2, and conservative PUE of 2!

18. 18 Where is Energy Consumed? Non-PSTN replacement More complicated If softphone draws little additional power –Still likely that end-point biggest component –But may not dominate consumption If users leave PCs on just as phones –Possibly even worse than PSTN! User / hardware study needed.

19. 19 How Do Design Choices Effect Power Consumption? C/S Inefficiencies –Power utilization efficiency – (PUE) Ratio of data center power draw to IT power draw e.g., cooling, network equipment, etc. –Idle power consumption (can be addressed in larger systems by techniques such as Somniloquy or Sleep Proxy –Percentage of user population that requires relaying major determinant of core energy consumption. SIP registrar / proxy / presence / server User agents INVITE media (voice, video, IM) (UDP or TCP) media server NAT / firewall network address of node E? (2) signaling (3) media (TCP) node C node B media relay (or relay) node A node D node E (1)

20. 20 How Do Design Choices Effect Power Consumption? P2P Consumption –Avoids these overheads by using machines that are already on In theory general user population In practice appears to be heavily subsidized by university machines What happens when machines are idling less… –Incurs small additional energy use for signaling and relaying But how small? SIP registrar / proxy / presence / server User agents INVITE media (voice, video, IM) (UDP or TCP) media server NAT / firewall network address of node E? (2) signaling (3) media (TCP) node C node B media relay (or relay) node A node D node E (1)

21. 21 Comparing C/S and P2P Compare under same load –Active calls –Call duration –Percentage of PSTN calls Generic C/S and P2P –Both use standard VoIP (e.g., not Skype) Isolate only services that differ between P2P, CS –Directory service –Call signaling –Media session –Presence

22. 22 Modeling P2P and C/S C/S model –C/S power consumption = #servers * Watts/server * redundancy factor * PUE P2P model –S super nodes active –p s super node consumption p s = 52mW P2P energy efficient when: S * p s < C/S power consumption One active super node per relayed call. Media server fully loaded. 100% calls relayed P2P may consume more than C/S!

23. 23 Caveats Peers –External meters do not provide sufficient resolution to determine p s w/ confidence –Will be in different states when relay starts Medium load unlikely to incur much extra overhead Low or high loads, p s could conceivably be large Consequently, prior distribution effects efficiency Servers –Energy usage not linear w/ load –Lower utilization hurts energy efficiency

24. 24 Making IP-Telephony Greener Make phones energy efficient –LCD, processor, WOL for phones? NATs & Firewalls –Get rid of NATs or rearchitect them –Use TCP to avoid NAT keep-alive –Make firewalls VoIP-friendly. Set up SIP user agents on gateways PC wakeup on receiving calls

25. 25 Outline IP-Telephony and power consumption Answering our questions: 1)Where is energy consumed? 2)How do different design choices effect energy consumption? 3)How we can make IP-telephony more energy efficient? Conclusion & Future Work

26. 26 Conclusions VoIP endpoints dominate total energy consumption in PSTN replacement systems P2P not necessarily more energy efficient than C/S. NATs and firewalls create the need for media relaying, one of the biggest components of core energy consumption.

27. 27 Future Work Obtain data on PSTN power consumption Work on accurately measuring p s Measure path length / routing differences between of direct and media-relayed calls. Study user behavior viz-a-vis softphone use –How much extra time are machines left on –Power draw during those periods Develop WOL capable hard-phones