exploiting latency bounds for energy efficient load balancing

21
Exploiting latency bounds for energy efficient load balancing Cruz Monrreal, Daniel Jones, Michael May and Mohit Taneja The University of Texas at Austin

Upload: michael-may

Post on 02-Jul-2015

270 views

Category:

Documents


2 download

DESCRIPTION

These slides are taken from a research paper the PSL group wrote while under the direction of Dr. Vijay Garg at the Universtiy of Texas at Austin. The abstract is provided below. In this paper we explore exploitation of latency bounds in order to gain energy efficiency in load balancing applications. We are proposing an energy aware job scheduler that uses vary-on, vary-off features in order to maximize time spent at peak utilization, while maintaining bounded latency. Computing resources will either be at load saturation(highest work per joule) or off. The premise being that servers are most efficient at peak utilization, measured in terms of energy per calculation. We explore the efficiency gains achieved through this approach and compare our results to other methods.

TRANSCRIPT

Page 1: Exploiting latency bounds for energy efficient load balancing

Exploiting latency bounds for energy efficient load balancing

Cruz Monrreal, Daniel Jones, Michael May and Mohit Taneja

The University of Texas at Austin

Page 2: Exploiting latency bounds for energy efficient load balancing

Overview

● Problem● State of server power (lack of power

proportionality)● Inspiration● Current Solutions● Assumptions● Our Solution● Comparisons of results● Limitations of our solutions● Future prospects● Q & A

Page 3: Exploiting latency bounds for energy efficient load balancing

Description of Problem

Current server implementations are power inefficient during low load hours.

Many requests do not need to be serviced as fast as possible thus have an acceptable stall period.

Page 4: Exploiting latency bounds for energy efficient load balancing

System Power Consumption

Source - http://static.usenix.org/events/hotpower08/tech/full_papers/rivoire/rivoire_html/

Page 5: Exploiting latency bounds for energy efficient load balancing

System Power ConsumptionTable 1 - Number of Cores utilized to Power Usage

# of Cores Power (W)

0 0

1 270

2 300

3 320

4 330

Page 6: Exploiting latency bounds for energy efficient load balancing

Inspiration

Page 7: Exploiting latency bounds for energy efficient load balancing

Inspiration (Cont)

Page 8: Exploiting latency bounds for energy efficient load balancing

Assumptions - Model4 core machine

3 servers total

1 job saturates a core

Instant on/off

No background tasks

Page 9: Exploiting latency bounds for energy efficient load balancing

Assumptions - Model (Cont)Load generator simulates sending variable time jobs (service requests) to load balancer.

Load Scheduler distributes jobs to servers.

Server simulates running job by sleeping the given time.

Server sends number of cores running back to load balancer with its own timestamp.

Page 10: Exploiting latency bounds for energy efficient load balancing

Current Solutions

No Power management (Round Robin)

Basic Power management (Round Robin)

Advanced Power management

Page 11: Exploiting latency bounds for energy efficient load balancing

Current Solutions - No Power Management

Load Scheduler uniformly schedules jobs to each server in sequence

Problems:Lots of time spent in idleFew cores used = low efficiency

Page 12: Exploiting latency bounds for energy efficient load balancing

Current Solutions - Round Robin w/o Power Management

Load Scheduler uniformly schedules jobs to each server in sequenceTurns off unused machines

Problems:Few cores used = low efficiency

Page 13: Exploiting latency bounds for energy efficient load balancing

Current Solutions - Round Robin w/ Server Toggling

Load Scheduler uniformly schedules jobs to each server by sending 4 jobs at a time sequentially. Turns off unused servers.

Problems:Does not fully utilize latency

Page 14: Exploiting latency bounds for energy efficient load balancing

Overview of Solution

If any servers are running but not full, the load balancer will send a job to the server with the most jobs running.

If all servers are full on/off than the load balancer will wait the given stall time until sending a job to an off server (thus turning it on).

Page 15: Exploiting latency bounds for energy efficient load balancing

Comparisons

Run at average load = 25%,50%,75%,100%

Vary job time around average load job time.

Example: 25% load timeJob time = 8-12 millisecondsStall time = 500 millisecondsTime between jobs = 2-8 milliseconds

Page 16: Exploiting latency bounds for energy efficient load balancing
Page 17: Exploiting latency bounds for energy efficient load balancing

Conclusion

Page 18: Exploiting latency bounds for energy efficient load balancing

Limitations

With large core counts advantages start to diminish

At 100% no gains

Page 19: Exploiting latency bounds for energy efficient load balancing

Future Prospects

Storage systems (SANs)Latency exploitation for networks

Page 20: Exploiting latency bounds for energy efficient load balancing

Q&A

Page 21: Exploiting latency bounds for energy efficient load balancing

Resources

[1] http://web.eecs.umich.edu/~twenisch/papers/asplos12.pdf

[2] http://static.usenix.org/events/hotpower08/tech/full_papers/rivoire/rivoire_html/