A Close Examina.on of Performance and Power Characteris.cs of 4G LTE Networks

Transcription

1 A Close Examina.on of Performance and Power Characteris.cs of 4G LTE Networks Junxian Huang 1 Feng Qian 1 Alexandre Gerber 2 Z. Morley Mao 1 Subhabrata Sen 2 Oliver Spatscheck 2 1 University of Michigan 2 AT&T Labs - Research June

2 LTE is new, requires explora.on 4G LTE (Long Term Evolu/on) is future trend Ini.ated by 3GPP in Mbps DL, 50Mbps UL, <5ms latency Entered commercial markets in 2009 Lessons from 3G UMTS networks Radio Resource Control (RRC) state machine is important App traffic pa\erns trigger state transi.ons, different states determine UE power usage and user experience State transi.ons incur energy, delay, signaling overhead

3 LTE state machine LTE power model Network performance Energy efficiency Parameter configura.on Mobile applica.on

4 RRC state transi.ons in LTE Continuous Reception Short Ti Tis Long Ttail RRC_CONNECTED Timer expiration RRC_IDLE Data transfer

5 RRC state transi.ons in LTE RRC_IDLE Continuous Reception No radio resource allocated Ti Low power state: 11.36mW average power Tis Short Long Promo.on delay from RRC_IDLE to RRC_CONNECTED: 260ms RRC_CONNECTED Timer expiration Ttail RRC_IDLE Data transfer

6 RRC state transi.ons in LTE Continuous Reception RRC_CONNECTED Short Ti Tis Long Radio resource allocated Power state is a func.on of Ttail data rate: 1060mW is the base power consump.on Up to 3300mW transmicng at full speed RRC_CONNECTED Timer expiration RRC_IDLE Data transfer

7 RRC state transi.ons in LTE Con/nuous Recep/on Continuous Reception Send/receive a packet Promote to RRC_CONNECTED Short Ti Tis Long Ttail Reset Ttail RRC_CONNECTED Timer expiration RRC_IDLE Data transfer

8 RRC state transi.ons in LTE Continuous Reception Short Ti Tis Long RRC_CONNECTED Timer expiration Ttail Ttail expires Ttail stops Demote to RRC_IDLE RRC_IDLE Data transfer

9 Tradeoffs of Ttail secngs Ttail seing Energy Consump/on # of state transi/ons Responsiveness Long High Small Fast Short Low Large Slow

10 RRC state transi.ons in LTE Continuous Reception : Discon/nuous Recep/on Listens to Ti downlink channel periodically for a short Ttail dura.on and sleeps for the rest.me to save energy Tis Short at the cost of responsiveness Long RRC_CONNECTED Timer expiration RRC_IDLE Data transfer

11 Discon.nuous Recep.on (): micro- sleeps for energy saving In LTE 4G, makes UE micro- sleep periodically in the RRC_CONNECTED state Short Long incurs tradeoffs between energy usage and latency Short sleep less and respond faster Long sleep more and respond slower In contrast, in UMTS 3G, UE is always listening to the downlink control channel in the data transmission states

12 in LTE A cycle consists of On Dura.on - UE monitors the downlink control channel (PDCCH) Off Dura.on - skip recep.on of downlink channel T i : Con.nuous recep.on inac.vity.mer When to start Short T is : Short inac.vity.mer When to start Long Data transfer On Duration Ti starts Ti expires Tis starts Tis expires Long cycle Continuous Reception Short cycle Long cycle

13 LTE state machine LTE power model Network performance Energy efficiency Parameter configura.on Mobile applica.on

14 Power trace of RRC state transi.ons 4000 t 2 : Data transfer starts Power (mw) t 1 : Promotion starts 0 t 1 t t t 4 25 Time (second) t 3 : Tail starts t 4 : Tail ends The data points are sampled and in RRC_CONNECTED tail is not obvious due to the low sampling rate

15 LTE power model Measured with a LTE phone and Monsoon power meter, averaged with repeated samples

16 LTE power model Measured with a LTE phone and Monsoon power meter, averaged with repeated samples

17 LTE power model Measured with a LTE phone and Monsoon power meter, averaged with repeated samples

18 LTE power model Measured with a LTE phone and Monsoon power meter, averaged with repeated samples

19 LTE power model Measured with a LTE phone and Monsoon power meter, averaged with repeated samples P(on) P(off) = 620mW, saves 36% energy in RRC_CONNECTED High power levels in both On and Off dura/ons in the cycle of RRC_CONNECTED

20 LTE consumes more instant power than 3G/WiFi in the high- power tail Average power for WiFi tail 120 mw Average power for 3G tail 800 mw Average power for LTE tail 1080 mw

21 Power model for data transfer A linear model is used to quan.fy instant power level: Downlink throughput t d Mbps Uplink throughput t u Mbps < 6% error rate in evalua/ons with real applica/ons

22 Energy per bit comparison LTE s high throughput compensates for the promo.on energy and tail energy Transfer Size LTE μ J / bit WiFi μ J / bit 3G μ J / bit 10KB MB Total energy per bit for downlink bulk data transfer

23 Energy per bit comparison LTE s high throughput compensates for the promo.on energy and tail energy Transfer LTE WiFi 3G Small data transfer, LTE wastes energy Size μ J / bit μ J / bit μ J / bit Large data transfer, LTE is energy efficient 10KB MB Total energy per bit for downlink bulk data transfer

24 LTE state machine LTE power model Network performance Energy efficiency Parameter configura.on Mobile applica.on

25 Network characteris.cs 4GTest on Android h>p://mobiperf.com/4g.html Measures network performance with the help of 46 M- Lab nodes across the world 3,300 users and 14,000 runs in 2 months 10/15/2011 ~ 12/15/2011 Latitude WiFi 25 WiMAX LTE Longitude 4GTest user coverage in the U.S.

26 Downlink throughput LTE median is 13Mbps, up to 30Mbps The LTE network is rela.vely unloaded WiFi, WiMAX < 5Mbps median 30 Y1: Network throughput (Mbps) WiFi LTE WiMAX ehrpd EVDO_A 1

27 Uplink throughput LTE median is 5.6Mbps, up to 20Mbps WiFi, WiMAX < 2Mbps median 30 Y1: Network throughput (Mbps) WiFi LTE WiMAX ehrpd EVDO_A 1

28 RTT LTE median 70ms WiFi similar to LTE WiMAX higher 30 Y1: Network throughput (Mbps) WiFi LTE WiMAX ehrpd EVDO_A 1

29 LTE state machine LTE power model Network performance Energy efficiency Parameter configura.on Mobile applica.on

30 User trace based analysis UMICH data set Collected from 20 volunteer smartphone users for five months, totaling 118GB Contains packet traces including full payload Trace- driven modeling methodology Network model simulator Simulates network states, such as RRC state transi.ons Power model simulator Calculates power usage based on the network states

31 Comparing total energy of all user traces via simula.on in LTE/3G/WiFi Total energy usage LTE/WiFi 23 3G/WiFi Energy ratio: LTE/WiFi Energy ratio: 3G/WiFi Energy ratio All User ID

32 Energy consump.on break down Tail energy dominates LTE energy consump.on, similar to 3G % of total energy LTE WiFi 3G % Idle energy % Tail energy % Promotion energy % Data transfer energy 0 All User ID The total energy for different networks and users is normalized to be 100%

33 Energy consump.on break down Tail energy dominates LTE energy consump.on, similar to 3G % Idle energy % Tail energy % Promotion energy % Data transfer energy 100 The tail problem is the key factor for 80 LTE s high energy consump.on, 60 similar to 3G networks 40 % of total energy 20 LTE WiFi 3G 0 All User ID The total energy for different networks and users is normalized to be 100%

34 LTE state machine LTE power model Network performance Energy efficiency Parameter configura.on Mobile applica.on

35 Impact of configuring LTE tail.mer (T tail ) S is defined to be the number of promo.ons T tail has significant impact on radio energy E, channel scheduling delay D, and signaling overhead S (relative change) E D S 0 5 T D T tail (second) T D is the default secng for T tail in the measured network

36 LTE state machine LTE power model Network performance Energy efficiency Parameter configura.on Mobile applica.on

37 App case study Studied 5 web- based apps LTE has comparable page loading.me as WiFi, with 3G lagging behind CPU usage for LTE/WiFi is between 80% ~ 90% during page loading Network does not appear to be the bo\leneck Total energy consump.on: LTE > 3G >> WiFi

38 App case study Studied 5 web- based apps LTE has In comparable LTE network, page applica.ons loading.me as WiFi, with 3G lagging behind should more aggressively burst CPU usage traffic for to LTE/WiFi make more is between efficient 80% ~ 90% during page loading use of the bandwidth given the Network does not appear to be the bo\leneck high energy overhead Total energy consump.on: LTE > 3G >> WiFi

39 Summary LTE has significantly higher speed, compared to 3G and WiFi LTE is much less power efficient than WiFi due to its tail energy for small data transfers Derived a power model of a commercial LTE network, with less than 6% error rate UE processing is the bo\leneck for web- based applica.ons in LTE networks Mobile app design should be LTE friendly

40 Thank you! Q & A Contact: Junxian Huang (hjx@umich.edu)

41 Backup slides

42 Power trace of in RRC_CONNECTED Power (mw) Time (ms)

43 RRC state transi.ons in LTE Send/ receive a packet Con/nuous Recep/on Continuous Reception Short Ti Tis Long Ttail Reset Ti RRC_CONNECTED Timer expiration RRC_IDLE Data transfer

44 RRC state transi.ons in LTE Continuous Reception Short Short Ti Tis Long Ttail Ti stops, Tis starts RRC_CONNECTED Timer expiration RRC_IDLE Data transfer

45 RRC state transi.ons in LTE Con/nuous Recep/on Continuous Reception Short Ti Tis Long Ttail Reset Ti, stops Tis RRC_CONNECTED Timer expiration RRC_IDLE Data transfer

46 RRC state transi.ons in LTE Continuous Reception Short Ti Tis expires Long Long Ttail Tis stops RRC_CONNECTED Timer expiration RRC_IDLE Data transfer

47 RRC state transi.ons in LTE Con/nuous Recep/on Continuous Reception Short Ti Tis Long Ttail Reset Ti RRC_CONNECTED Timer expiration RRC_IDLE Data transfer

48 Impact of inac.vity.mer (T i ): Con.nuous recep.on to short Differently, S is defined as the sum of the con.nuous recep.on.me and on dura.ons in RRC_CONNECTED T i has negligible impact on E, however, S is significantly affected (relative change) E D S 0 T D T i (ms)

49 Interes.ng ques.ons about LTE To users: what is the end performance? Network performance, such as RTT and throughput, how it compares with WiFi, 3G and WiMAX, etc. Energy efficiency affec.ng ba\ery life, is LTE more power efficient than 3G or WiFi? To ISPs: what is the impact of configuring LTE- related parameters on UE power saving, and delay/signaling overhead? To OS/applica.on developers: what is the performance bo\leneck of applica.ons in LTE network, CPU or network speed?

50 Energy per bit comparison For large data transfer with maximum rate, LTE s energy efficiency is comparable with WiFi, due to LTE s high downlink throughput µj / bit LTE DOWN LTE UP WiFi DOWN WiFi UP 3G DOWN 3G UP Bulk data size (kb)

51 One way delay and impact of packet size (not quite related) LTE uplink one way delay (OWD) is larger than that of downlink RTT in LTE is more sensi.ve to packet size than WiFi, mainly due to uplink OWD Delay (ms) UP OWD DOWN OWD RTT Packet size without TCP/IP headers (byte)

52 JavaScript execu.on speed: a representa.ve view of smartphone processing capability From 2009 to 2011, smartphones have significantly improved JavaScript execu.on speed Laptop/Chrome Mac OS X iphone 4S/Safari ios iphone 4/Safari ios HTC/Default Android Samsung/IE Windows Phone 7.5 G1/Default Android 1.6 G1/Default Android 1.6 iphone/safari ios 3.0 Samsung/IE Windows Mobile s 2.26s 3.93s 4.42s 9.46s Tested in Nov, 2009 Tested in Nov, s 95.66s s s Time to finish JavaScript benchmark (sec)

53 Power model for data transfer A linear model is used to quan.fy instant power level: Uplink/downlink throughput t u /t d (Mbps) < 6% error rate for predic/ng energy usage of 5 real applica/ons