ptec: a system for predictive thermal and energy control ... · ptec: a system for predictive...
TRANSCRIPT
![Page 1: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/1.jpg)
PTEC: A System for Predictive Thermal and Energy Control in Data Centers
Jinzhu Chen1 Rui Tan3 Guoliang Xing1 Xiaorui Wang2
1 Michigan State University, USA 2 Ohio State University, USA
3 Advanced Digital Sciences Center, Illinois at Singapore
![Page 2: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/2.jpg)
Conservative Cooling Settings
• Data centers eat massive energy
– An industry data center = a mid-size town
• 60% non-computing energy ratio [Uptime 2012]
– 50% for cooling 24°C in 90% data centers vs. recommended 27°C
– 10% for circulation High fan speeds and simple control
EMC's new data center in Durham, NC Raised-floor cooling [www.42u.com]
CR
AC
CR
AC
rack
![Page 3: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/3.jpg)
Related Work
• New data centers: 10% for non-computing [Google]
– Clean slate redesign
– Retrofitting technologies for existing data centers
• Thermal & energy control: prevent overheat & reduce non-computing energy
– Single-variable (e.g., server workload)
– Multi-variable Ours: AC + server fan (major correlated energy eaters)
• React to detected hotspots
– Low temperature setpoint to resolve
– Less energy-efficient
![Page 4: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/4.jpg)
Predictive Thermal & Energy Control
• Energy-efficiently prevent hot spots
• Predict energy consumption & thermal conditions for each possible AC & fan control action
– Minimize predicted AC and fan energy
– No predicted hot spots at server inlets
Server inlets Server outlets
cool air hot air
![Page 5: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/5.jpg)
Challenges
• Complex cyber-physical dynamics
– Air flow, server workload
– Coupling btw control and thermal condition
• Real-time and scalable
– No polynomial-time algorithms
– Large # of controllable variables
Row 1
Row 2
Raised-floor
cold air
Server
exhaust
Dynamic workload in MSU HPCC over 12 days
![Page 6: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/6.jpg)
Sensing (CPU, fan speed, temperature, airflow)
Real-time Prediction
Approach Overview
Data Center
• Environment sensing – Built-in sensors, external sensor network
• Temperature & energy prediction – Sensor data + energy models + candidate control action
• Predictive controller – Constrained optimization
Server Fans
AC
Predictive Controller
candidate
solution
predicted
temperatures
![Page 7: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/7.jpg)
Outline • Motivation & Approach Overview
• Sensing and Prediction
• Predictive Thermal & Energy Control
• Evaluation
Sensing (CPU, fan speed, temperature, airflow)
Real-time Prediction
Data Center
Server Fans
AC
Predictive Controller
Candidate
solution
predicted
temperatures
![Page 8: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/8.jpg)
Inlet / outlet temperature
Sensing
Air velocity Air temperature
CPU utilization
Fan speed
Wireless sensor / built-in sensor
Airflow velocity
Server network
Sensing
Sensor sampling rate: 30 seconds
![Page 9: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/9.jpg)
Inlet / outlet temperature
Sensing
Air velocity Air temperature
CPU utilization
Fan speed
Wireless sensor / built-in sensor
Airflow velocity
Server network
Sensing
Sensor sampling rate: 30 seconds
![Page 10: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/10.jpg)
Inlet / outlet temperature
Sensing
Air velocity Air temperature
CPU utilization
Fan speed
Wireless sensor / built-in sensor
Airflow velocity
Server network
Sensing
Sensor sampling rate: 30 seconds
![Page 11: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/11.jpg)
Inlet / outlet temperature
Sensing
Air velocity Air temperature
CPU utilization
Fan speed
Wireless sensor / built-in sensor
Airflow velocity
Server network
Sensing
Sensor sampling rate: 30 seconds
![Page 12: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/12.jpg)
• Vector pt: Sensor measurements at moment t
• Prediction with a horizon of k sampling periods
Real-time Temperature Prediction 1
] [11
htttkkt
pppAt
1 A High-Fidelity Temperature Distribution Forecasting System for Data Centers.
RTSS 2012.
![Page 13: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/13.jpg)
• Vector pt: Sensor measurements at moment t
• Prediction with a horizon of k sampling periods
Real-time Temperature Prediction 1
Regression matrix (offline trained)
] [11
htttkkt
pppAt
1 A High-Fidelity Temperature Distribution Forecasting System for Data Centers.
RTSS 2012.
![Page 14: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/14.jpg)
• Vector pt: Sensor measurements at moment t
• Prediction with a horizon of k sampling periods
Real-time Temperature Prediction 1
Regression matrix (offline trained)
] [11
htttkkt
pppAt
1 A High-Fidelity Temperature Distribution Forecasting System for Data Centers.
RTSS 2012.
Historical measurements
![Page 15: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/15.jpg)
• Vector pt: Sensor measurements at moment t
• Prediction with a horizon of k sampling periods
•Increasing k: temperature distribution evolution
Real-time Temperature Prediction 1
Regression matrix (offline trained)
] [11
htttkkt
pppAt
1 A High-Fidelity Temperature Distribution Forecasting System for Data Centers.
RTSS 2012.
Historical measurements
![Page 16: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/16.jpg)
• Vector pt: Sensor measurements at moment t
• Prediction with a horizon of k sampling periods
•Increasing k: temperature distribution evolution
•Less than 0.5 oC error when k < 10 min
•Error increases with k
Real-time Temperature Prediction 1
Regression matrix (offline trained)
] [11
htttkkt
pppAt
1 A High-Fidelity Temperature Distribution Forecasting System for Data Centers.
RTSS 2012.
Historical measurements
![Page 17: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/17.jpg)
• Fan regulates speed by duty cycle of PWM signal
– Part of control solution
• Offline / online learning
PWM duty cycle
(%)
Fan power
(W)
60 73 87 100
12
20
29.4
Server Fan Power Model
Delta Electronics BFB1012EH
![Page 18: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/18.jpg)
• AC power consumption
– Temperature setpoint, blower speed, return hot air temperature
– Offline learning or from spec.
• AC in our testbed
– Binary blower state (B)
– Binary compressor state (S)
– Return hot air temp. TH
])([201
HAC
TSBP
AC Power Model
Pow
er
(kW
)
(oC)
Return hot air temp. TH
B=1 S=1
![Page 19: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/19.jpg)
Outline • Motivation & Approach Overview
• Sensing and Prediction
• Predictive Thermal & Energy Control
• Evaluation
Sensing (CPU, fan speed, temperature, airflow)
Real-time Prediction
Data Center
Server Fans
AC
Predictive Controller
Candidate
solution
predicted
temperatures
![Page 20: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/20.jpg)
• Find fan speeds, AC settings
– Minimize predicted total power in opt. horizon
– Predicted inlet temp. upper-bounded Prevent overheating
– Predicted inlet temp. variation upper-bounded Failure rate increases with variation [El-Sayed 2012]
Problem Formulation
Optimization horizon
Control cycle
Time
Sampling
period
Now
![Page 21: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/21.jpg)
• Find fan speeds, AC settings
– Minimize predicted total power in opt. horizon
– Predicted inlet temp. upper-bounded Prevent overheating
– Predicted inlet temp. variation upper-bounded Failure rate increases with variation [El-Sayed 2012]
• Compute-intensive
– A small data center: 237 controllable variables 229 fans + 4 blower speeds + 4 temp. setpoints
Problem Formulation
Optimization horizon
Control cycle
Time
Sampling
period
Now
![Page 22: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/22.jpg)
Coordinated Control • Autonomous fan control
– speed = f (inlet temperature, CPU utilization)
Temperature prediction
Solution evaluator
Autonomous fan control algorithm
Predicted temperatures
Predicted
CPU utilization
if available
Fan power models AC power models
+
Sensor
data
Inlet temp. < T1? Variation < T2?
Predicted
fan speeds
Candidate
solution
(AC settings)
Predicted total energy Valid solution?
![Page 23: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/23.jpg)
Coordinated Control • Autonomous fan control
– speed = f (inlet temperature, CPU utilization)
Temperature prediction
Solution evaluator
Autonomous fan control algorithm
Predicted temperatures
Predicted
CPU utilization
if available
Fan power models AC power models
+
Sensor
data
Inlet temp. < T1? Variation < T2?
Predicted
fan speeds
Candidate
solution
(AC settings)
Predicted total energy Valid solution?
![Page 24: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/24.jpg)
Coordinated Control • Autonomous fan control
– speed = f (inlet temperature, CPU utilization)
Temperature prediction
Solution evaluator
Autonomous fan control algorithm
Predicted temperatures
Predicted
CPU utilization
if available
Fan power models AC power models
+
Sensor
data
Inlet temp. < T1? Variation < T2?
Predicted
fan speeds
Candidate
solution
(AC settings)
Predicted total energy Valid solution?
![Page 25: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/25.jpg)
Coordinated Control • Autonomous fan control
– speed = f (inlet temperature, CPU utilization)
Temperature prediction
Solution evaluator
Autonomous fan control algorithm
Predicted temperatures
Predicted
CPU utilization
if available
Fan power models AC power models
+
Sensor
data
Inlet temp. < T1? Variation < T2?
Predicted
fan speeds
Candidate
solution
(AC settings)
Predicted total energy Valid solution?
![Page 26: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/26.jpg)
Coordinated Control • Autonomous fan control
– speed = f (inlet temperature, CPU utilization)
Temperature prediction
Solution evaluator
Autonomous fan control algorithm
Predicted temperatures
Predicted
CPU utilization
if available
Fan power models AC power models
+
Sensor
data
Inlet temp. < T1? Variation < T2?
Predicted
fan speeds
Candidate
solution
(AC settings)
Predicted total energy Valid solution?
![Page 27: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/27.jpg)
Coordinated Control • Autonomous fan control
– speed = f (inlet temperature, CPU utilization)
Temperature prediction
Solution evaluator
Autonomous fan control algorithm
Predicted temperatures
Predicted
CPU utilization
if available
Fan power models AC power models
+
Sensor
data
Inlet temp. < T1? Variation < T2?
Predicted
fan speeds
Candidate
solution
(AC settings)
Predicted total energy Valid solution?
Simulated annealing
![Page 28: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/28.jpg)
Coordinated Control • Autonomous fan control
– speed = f (inlet temperature, CPU utilization)
Temperature prediction
Solution evaluator
Autonomous fan control algorithm
Predicted temperatures
Predicted
CPU utilization
if available
Fan power models AC power models
+
Sensor
data
Inlet temp. < T1? Variation < T2?
Predicted
fan speeds
Candidate
solution
(AC settings)
Predicted total energy Valid solution?
Simulated annealing
Fans’ autonomous behavior
considered in optimization
![Page 29: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/29.jpg)
• Ensure upper-bounded CPU temperature
– Measurement-based approach
Autonomous Fan Control M
in P
WM
duty
cycle
(%
)
60
73
87
100
CPU=0.9
30 35 40 45 Inlet temperature (oC)
CPU temperature upper bound = 50 oC
CPU=0.5 CPU=0.1
![Page 30: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/30.jpg)
Outline
• Motivation & Approach Overview
• Sensing and Prediction
• Predictive Thermal & Energy Control
• Evaluation
![Page 31: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/31.jpg)
Single-Rack Experiments
• Setup – 15 servers, 32 wireless sensors, portable AC with wireless power relay
– Controllable CPU utilization
• System implementation – Predictive controller: MATLAB on a desktop
– Fan control: BASH on servers
![Page 32: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/32.jpg)
• Max Cooling: fixed low AC setpoint, full server fan speed
Compare with Max Cooling
5 5.5 6 6.5 7
Inle
t (o
C)
Fan (
W)
AC
(W
) C
PU
util.
Time (hour)
Inlet temp. upper bound
![Page 33: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/33.jpg)
• Max Cooling: fixed low AC setpoint, full server fan speed
Compare with Max Cooling
5 5.5 6 6.5 7
Inle
t (o
C)
Fan (
W)
AC
(W
) C
PU
util.
Time (hour)
Inlet temp. upper bound
Our approach
![Page 34: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/34.jpg)
• Max Cooling: fixed low AC setpoint, full server fan speed
Compare with Max Cooling
5 5.5 6 6.5 7
Inle
t (o
C)
Fan (
W)
AC
(W
) C
PU
util.
Time (hour)
Inlet temp. upper bound
Our approach Max cooling
![Page 35: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/35.jpg)
• Max Cooling: fixed low AC setpoint, full server fan speed
Compare with Max Cooling
5 5.5 6 6.5 7
Inle
t (o
C)
Fan (
W)
AC
(W
) C
PU
util.
Time (hour)
Inlet temp. upper bound
Our approach Max cooling
![Page 36: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/36.jpg)
• Max Cooling: fixed low AC setpoint, full server fan speed
Compare with Max Cooling
5 5.5 6 6.5 7
Inle
t (o
C)
Fan (
W)
AC
(W
) C
PU
util.
Time (hour)
Inlet temp. upper bound
Our approach Max cooling
![Page 37: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/37.jpg)
• Max Cooling: fixed low AC setpoint, full server fan speed
Compare with Max Cooling
5 5.5 6 6.5 7
Inle
t (o
C)
Fan (
W)
AC
(W
) C
PU
util.
Time (hour)
Inlet temp. upper bound
Our approach Max cooling
![Page 38: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/38.jpg)
Compare with Reactive
Time
Inlet temp. TU
TL
Turn on AC
Turn off AC
![Page 39: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/39.jpg)
Compare with Reactive C
PU
util.
Inle
t (o
C)
Reactive (TL=30 oC, TU=33 oC)
Upper bound
Time (minutes)
Ours
![Page 40: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/40.jpg)
Compare with Reactive C
PU
util.
Inle
t (o
C)
Reactive (TL=30 oC, TU=33 oC)
Upper bound
Time (minutes)
Ours
30 0 60 90 Time (minutes)
Inle
t (o
C) Upper bound
Reactive (TL=30 oC, TU=32 oC) Ours
![Page 41: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/41.jpg)
Compare with Reactive (cont’d)
Reactive Ours
TL (oC)
TU (oC)
Avg power (Watt)
Avg power (Watt)
27 30 916
638
28 30 807
28 31 806
29 31 817
29 32 746
30 32 669
30 33 714
31 33 640
Inlet temp. upper bound = 33oC, server idle
![Page 42: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/42.jpg)
Compare with Reactive (cont’d)
Reactive Ours
TL (oC)
TU (oC)
Avg power (Watt)
Avg power (Watt)
27 30 916
638
28 30 807
28 31 806
29 31 817
29 32 746
30 32 669
30 33 714
31 33 640
Inlet temp. upper bound = 33oC, server idle
>
![Page 43: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/43.jpg)
Compare with Reactive (cont’d)
Reactive Ours
TL (oC)
TU (oC)
Avg power (Watt)
Avg power (Watt)
27 30 916
638
28 30 807
28 31 806
29 31 817
29 32 746
30 32 669
30 33 714
31 33 640
Inlet temp. upper bound = 33oC, server idle
Inlet temp. > 33oC
>
![Page 44: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/44.jpg)
• MSU HPCC
– 5 racks, 229 servers, 4 in-row ACs
• CPU utilization data trace
– 1 sample / minute
– 12 days
In-row ACs
In-row ACs
Trace-Driven CFD Simulations
Row 1
Row 2
Raised-floor
cold air
Server
exhaust
Computational fluid dynamics (CFD) model
![Page 45: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/45.jpg)
• Results for 3 racks
Dynamic Workload C
PU
util.
Inle
t (o
C)
Time (hour) 0 24 12 6 18
Inlet temp. upper bound
![Page 46: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/46.jpg)
• Results for 3 racks
•Compare with TAPO [Huang et al. 2011] •Need fine parameter tuning •Poorly adapt to dynamic workload
Dynamic Workload C
PU
util.
Inle
t (o
C)
Time (hour) 0 24 12 6 18
Inlet temp. upper bound
![Page 47: PTEC: A System for Predictive Thermal and Energy Control ... · PTEC: A System for Predictive Thermal and Energy Control in Data Centers Jinzhu 1Chen 1Rui 2Tan3 Guoliang Xing Xiaorui](https://reader033.vdocument.in/reader033/viewer/2022041910/5e672176c7a96811710d14e9/html5/thumbnails/47.jpg)
Conclusion
• Predictive thermal and energy control – Minimize AC and fan energy – Upper-bound inlet temperature & variation
• Coordinated Control – Autonomous fan control to ensure CPU temp. – Reduce complexity
• Testbed experiments & CFD simulations – Outperform reactive approach – Adapt to dynamic workload