SlideShare une entreprise Scribd logo
1  sur  21
Télécharger pour lire hors ligne
Exploiting latency bounds for
energy efficient load balancing
       Cruz Monrreal, Daniel Jones,
       Michael May and Mohit Taneja
    The University of Texas at Austin
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
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.
System Power Consumption




    Source - http://static.usenix.org/events/hotpower08/tech/full_papers/rivoire/rivoire_html/
System Power Consumption
  Table 1 - Number of Cores utilized to Power Usage



             # of Cores   Power (W)

             0            0

             1            270

             2            300

             3            320

             4            330
Inspiration
Inspiration (Cont)
Assumptions - Model
4 core machine

3 servers total

1 job saturates a core

Instant on/off

No background tasks
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.
Current Solutions

No Power management (Round Robin)

Basic Power management (Round Robin)

Advanced Power management
Current Solutions - No Power
Management
                  Load Scheduler
                  uniformly schedules
                  jobs to each server in
                  sequence

                  Problems:
                  Lots of time spent in idle
                  Few cores used = low
                  efficiency
Current Solutions - Round Robin w/o
Power Management
                  Load Scheduler
                  uniformly schedules
                  jobs to each server in
                  sequence
                  Turns off unused
                  machines

                  Problems:
                  Few cores used = low
                  efficiency
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
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).
Comparisons
Run at average load = 25%,50%,75%,100%

Vary job time around average load job time.

Example: 25% load time
Job time = 8-12 milliseconds
Stall time = 500 milliseconds
Time between jobs = 2-8 milliseconds
Conclusion
Limitations
With large core counts advantages start to
diminish




At 100% no gains
Future Prospects
Storage systems (SANs)
Latency exploitation for networks
Q&A
Resources
[1]   http://web.eecs.umich.edu/~twenisch/papers/asplos12.pdf


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

Contenu connexe

Tendances

Load Balancing In Cloud Computing newppt
Load Balancing In Cloud Computing newpptLoad Balancing In Cloud Computing newppt
Load Balancing In Cloud Computing newpptUtshab Saha
 
Load Balancing in Cloud
Load Balancing in CloudLoad Balancing in Cloud
Load Balancing in CloudMphasis
 
Php ieee project & abstract
Php ieee project & abstractPhp ieee project & abstract
Php ieee project & abstractaswin tbbc
 
Load Balancing from the Cloud - Layer 7 Aware Solution
Load Balancing from the Cloud - Layer 7 Aware SolutionLoad Balancing from the Cloud - Layer 7 Aware Solution
Load Balancing from the Cloud - Layer 7 Aware SolutionImperva Incapsula
 
PowerTrade SurgeGuard
PowerTrade SurgeGuardPowerTrade SurgeGuard
PowerTrade SurgeGuardShi Junxiao
 
load balancing in public cloud ppt
load balancing in public cloud pptload balancing in public cloud ppt
load balancing in public cloud pptKrishna Kumar
 
Load balancing
Load balancingLoad balancing
Load balancingSoujanya V
 
Взгляд на облака с точки зрения HPC
Взгляд на облака с точки зрения HPCВзгляд на облака с точки зрения HPC
Взгляд на облака с точки зрения HPCOlga Lavrentieva
 
Variations in Performance and Scalability when Migrating n-Tier Applications ...
Variations in Performance and Scalability when Migrating n-Tier Applications ...Variations in Performance and Scalability when Migrating n-Tier Applications ...
Variations in Performance and Scalability when Migrating n-Tier Applications ...deepalk
 
An Introduction To Space Based Architecture
An Introduction To Space Based ArchitectureAn Introduction To Space Based Architecture
An Introduction To Space Based ArchitectureAmin Abbaspour
 
Webinar Slides: MySQL Multi-Site Multi-Master Done Right
Webinar Slides: MySQL Multi-Site Multi-Master Done RightWebinar Slides: MySQL Multi-Site Multi-Master Done Right
Webinar Slides: MySQL Multi-Site Multi-Master Done RightContinuent
 
Velocity 2018 preetha appan final
Velocity 2018   preetha appan finalVelocity 2018   preetha appan final
Velocity 2018 preetha appan finalpreethaappan
 
Avoiding the ring of death
Avoiding the ring of deathAvoiding the ring of death
Avoiding the ring of deathAishvarya Verma
 
Probabilistic consolidation of virtual machines in self organizing cloud data...
Probabilistic consolidation of virtual machines in self organizing cloud data...Probabilistic consolidation of virtual machines in self organizing cloud data...
Probabilistic consolidation of virtual machines in self organizing cloud data...WMLab,NCU
 
Shreelakshmi(power).pptx
Shreelakshmi(power).pptxShreelakshmi(power).pptx
Shreelakshmi(power).pptxsurbhi agarwal
 
Fast dataarchitecture
Fast dataarchitectureFast dataarchitecture
Fast dataarchitectureKnoldus Inc.
 
Sql disaster recovery
Sql disaster recoverySql disaster recovery
Sql disaster recoverySqlperfomance
 
LOAD BALANCING ALGORITHM TO IMPROVE RESPONSE TIME ON CLOUD COMPUTING
LOAD BALANCING ALGORITHM TO IMPROVE RESPONSE TIME ON CLOUD COMPUTINGLOAD BALANCING ALGORITHM TO IMPROVE RESPONSE TIME ON CLOUD COMPUTING
LOAD BALANCING ALGORITHM TO IMPROVE RESPONSE TIME ON CLOUD COMPUTINGijccsa
 
The Dell EMC PowerMax 8000 outperformed another vendor's array on an OLTP-lik...
The Dell EMC PowerMax 8000 outperformed another vendor's array on an OLTP-lik...The Dell EMC PowerMax 8000 outperformed another vendor's array on an OLTP-lik...
The Dell EMC PowerMax 8000 outperformed another vendor's array on an OLTP-lik...Principled Technologies
 

Tendances (20)

Load Balancing In Cloud Computing newppt
Load Balancing In Cloud Computing newpptLoad Balancing In Cloud Computing newppt
Load Balancing In Cloud Computing newppt
 
Load Balancing in Cloud
Load Balancing in CloudLoad Balancing in Cloud
Load Balancing in Cloud
 
Php ieee project & abstract
Php ieee project & abstractPhp ieee project & abstract
Php ieee project & abstract
 
Load Balancing from the Cloud - Layer 7 Aware Solution
Load Balancing from the Cloud - Layer 7 Aware SolutionLoad Balancing from the Cloud - Layer 7 Aware Solution
Load Balancing from the Cloud - Layer 7 Aware Solution
 
PowerTrade SurgeGuard
PowerTrade SurgeGuardPowerTrade SurgeGuard
PowerTrade SurgeGuard
 
load balancing in public cloud ppt
load balancing in public cloud pptload balancing in public cloud ppt
load balancing in public cloud ppt
 
Load balancing
Load balancingLoad balancing
Load balancing
 
Взгляд на облака с точки зрения HPC
Взгляд на облака с точки зрения HPCВзгляд на облака с точки зрения HPC
Взгляд на облака с точки зрения HPC
 
Variations in Performance and Scalability when Migrating n-Tier Applications ...
Variations in Performance and Scalability when Migrating n-Tier Applications ...Variations in Performance and Scalability when Migrating n-Tier Applications ...
Variations in Performance and Scalability when Migrating n-Tier Applications ...
 
An Introduction To Space Based Architecture
An Introduction To Space Based ArchitectureAn Introduction To Space Based Architecture
An Introduction To Space Based Architecture
 
Webinar Slides: MySQL Multi-Site Multi-Master Done Right
Webinar Slides: MySQL Multi-Site Multi-Master Done RightWebinar Slides: MySQL Multi-Site Multi-Master Done Right
Webinar Slides: MySQL Multi-Site Multi-Master Done Right
 
Velocity 2018 preetha appan final
Velocity 2018   preetha appan finalVelocity 2018   preetha appan final
Velocity 2018 preetha appan final
 
Avoiding the ring of death
Avoiding the ring of deathAvoiding the ring of death
Avoiding the ring of death
 
Probabilistic consolidation of virtual machines in self organizing cloud data...
Probabilistic consolidation of virtual machines in self organizing cloud data...Probabilistic consolidation of virtual machines in self organizing cloud data...
Probabilistic consolidation of virtual machines in self organizing cloud data...
 
Shreelakshmi(power).pptx
Shreelakshmi(power).pptxShreelakshmi(power).pptx
Shreelakshmi(power).pptx
 
Fast dataarchitecture
Fast dataarchitectureFast dataarchitecture
Fast dataarchitecture
 
Map reduce
Map reduceMap reduce
Map reduce
 
Sql disaster recovery
Sql disaster recoverySql disaster recovery
Sql disaster recovery
 
LOAD BALANCING ALGORITHM TO IMPROVE RESPONSE TIME ON CLOUD COMPUTING
LOAD BALANCING ALGORITHM TO IMPROVE RESPONSE TIME ON CLOUD COMPUTINGLOAD BALANCING ALGORITHM TO IMPROVE RESPONSE TIME ON CLOUD COMPUTING
LOAD BALANCING ALGORITHM TO IMPROVE RESPONSE TIME ON CLOUD COMPUTING
 
The Dell EMC PowerMax 8000 outperformed another vendor's array on an OLTP-lik...
The Dell EMC PowerMax 8000 outperformed another vendor's array on an OLTP-lik...The Dell EMC PowerMax 8000 outperformed another vendor's array on an OLTP-lik...
The Dell EMC PowerMax 8000 outperformed another vendor's array on an OLTP-lik...
 

Similaire à Exploiting latency bounds for energy efficient load balancing

load-balancing-method-for-embedded-rt-system-20120711-0940
load-balancing-method-for-embedded-rt-system-20120711-0940load-balancing-method-for-embedded-rt-system-20120711-0940
load-balancing-method-for-embedded-rt-system-20120711-0940Samsung Electronics
 
How Workload Prioritization Reduces Your Datacenter Footprint
How Workload Prioritization Reduces Your Datacenter FootprintHow Workload Prioritization Reduces Your Datacenter Footprint
How Workload Prioritization Reduces Your Datacenter FootprintScyllaDB
 
A Study on Task Scheduling in Could Data Centers for Energy Efficacy
A Study on Task Scheduling in Could Data Centers for Energy Efficacy A Study on Task Scheduling in Could Data Centers for Energy Efficacy
A Study on Task Scheduling in Could Data Centers for Energy Efficacy Ehsan Sharifi
 
Adding Value in the Cloud with Performance Test
Adding Value in the Cloud with Performance TestAdding Value in the Cloud with Performance Test
Adding Value in the Cloud with Performance TestRodolfo Kohn
 
Cloudsim & Green Cloud
Cloudsim & Green CloudCloudsim & Green Cloud
Cloudsim & Green CloudNeda Maleki
 
Cloudsim & greencloud
Cloudsim & greencloud Cloudsim & greencloud
Cloudsim & greencloud nedamaleki87
 
Cassandra Summit 2014: Cassandra Compute Cloud: An elastic Cassandra Infrastr...
Cassandra Summit 2014: Cassandra Compute Cloud: An elastic Cassandra Infrastr...Cassandra Summit 2014: Cassandra Compute Cloud: An elastic Cassandra Infrastr...
Cassandra Summit 2014: Cassandra Compute Cloud: An elastic Cassandra Infrastr...DataStax Academy
 
Iaetsd appliances of harmonizing model in cloud
Iaetsd appliances of harmonizing model in cloudIaetsd appliances of harmonizing model in cloud
Iaetsd appliances of harmonizing model in cloudIaetsd Iaetsd
 
Load Balancing in Parallel and Distributed Database
Load Balancing in Parallel and Distributed DatabaseLoad Balancing in Parallel and Distributed Database
Load Balancing in Parallel and Distributed DatabaseMd. Shamsur Rahim
 
Scalability using Node.js
Scalability using Node.jsScalability using Node.js
Scalability using Node.jsratankadam
 
Autonomic Decentralised Elasticity Management of Cloud Applications
Autonomic Decentralised Elasticity Management of Cloud ApplicationsAutonomic Decentralised Elasticity Management of Cloud Applications
Autonomic Decentralised Elasticity Management of Cloud ApplicationsSrikumar Venugopal
 
Towards True Elasticity of Spark-(Michael Le and Min Li, IBM)
Towards True Elasticity of Spark-(Michael Le and Min Li, IBM)Towards True Elasticity of Spark-(Michael Le and Min Li, IBM)
Towards True Elasticity of Spark-(Michael Le and Min Li, IBM)Spark Summit
 
IEEE 2014 JAVA CLOUD COMPUTING PROJECTS Optimal power allocation and load dis...
IEEE 2014 JAVA CLOUD COMPUTING PROJECTS Optimal power allocation and load dis...IEEE 2014 JAVA CLOUD COMPUTING PROJECTS Optimal power allocation and load dis...
IEEE 2014 JAVA CLOUD COMPUTING PROJECTS Optimal power allocation and load dis...IEEEGLOBALSOFTSTUDENTPROJECTS
 
2014 IEEE JAVA CLOUD COMPUTING PROJECT Optimal power allocation and load dist...
2014 IEEE JAVA CLOUD COMPUTING PROJECT Optimal power allocation and load dist...2014 IEEE JAVA CLOUD COMPUTING PROJECT Optimal power allocation and load dist...
2014 IEEE JAVA CLOUD COMPUTING PROJECT Optimal power allocation and load dist...IEEEFINALYEARSTUDENTPROJECT
 
Scylla Summit 2022: Scylla 5.0 New Features, Part 1
Scylla Summit 2022: Scylla 5.0 New Features, Part 1Scylla Summit 2022: Scylla 5.0 New Features, Part 1
Scylla Summit 2022: Scylla 5.0 New Features, Part 1ScyllaDB
 
Planning For Catastrophe with IBM WAS and IBM BPM
Planning For Catastrophe with IBM WAS and IBM BPMPlanning For Catastrophe with IBM WAS and IBM BPM
Planning For Catastrophe with IBM WAS and IBM BPMWASdev Community
 
cmandtracing-1560359.pdf
cmandtracing-1560359.pdfcmandtracing-1560359.pdf
cmandtracing-1560359.pdfkriole13
 
How Busy Is Too Busy? Automating Your System for Maximum Throughput
How Busy Is Too Busy? Automating Your System for Maximum Throughput How Busy Is Too Busy? Automating Your System for Maximum Throughput
How Busy Is Too Busy? Automating Your System for Maximum Throughput Compuware
 
Probabilistic consolidation of virtual machines in self organizing cloud data...
Probabilistic consolidation of virtual machines in self organizing cloud data...Probabilistic consolidation of virtual machines in self organizing cloud data...
Probabilistic consolidation of virtual machines in self organizing cloud data...Papitha Velumani
 
Power aware load balancing in cloud
Power aware load balancing in cloud Power aware load balancing in cloud
Power aware load balancing in cloud manjula manju
 

Similaire à Exploiting latency bounds for energy efficient load balancing (20)

load-balancing-method-for-embedded-rt-system-20120711-0940
load-balancing-method-for-embedded-rt-system-20120711-0940load-balancing-method-for-embedded-rt-system-20120711-0940
load-balancing-method-for-embedded-rt-system-20120711-0940
 
How Workload Prioritization Reduces Your Datacenter Footprint
How Workload Prioritization Reduces Your Datacenter FootprintHow Workload Prioritization Reduces Your Datacenter Footprint
How Workload Prioritization Reduces Your Datacenter Footprint
 
A Study on Task Scheduling in Could Data Centers for Energy Efficacy
A Study on Task Scheduling in Could Data Centers for Energy Efficacy A Study on Task Scheduling in Could Data Centers for Energy Efficacy
A Study on Task Scheduling in Could Data Centers for Energy Efficacy
 
Adding Value in the Cloud with Performance Test
Adding Value in the Cloud with Performance TestAdding Value in the Cloud with Performance Test
Adding Value in the Cloud with Performance Test
 
Cloudsim & Green Cloud
Cloudsim & Green CloudCloudsim & Green Cloud
Cloudsim & Green Cloud
 
Cloudsim & greencloud
Cloudsim & greencloud Cloudsim & greencloud
Cloudsim & greencloud
 
Cassandra Summit 2014: Cassandra Compute Cloud: An elastic Cassandra Infrastr...
Cassandra Summit 2014: Cassandra Compute Cloud: An elastic Cassandra Infrastr...Cassandra Summit 2014: Cassandra Compute Cloud: An elastic Cassandra Infrastr...
Cassandra Summit 2014: Cassandra Compute Cloud: An elastic Cassandra Infrastr...
 
Iaetsd appliances of harmonizing model in cloud
Iaetsd appliances of harmonizing model in cloudIaetsd appliances of harmonizing model in cloud
Iaetsd appliances of harmonizing model in cloud
 
Load Balancing in Parallel and Distributed Database
Load Balancing in Parallel and Distributed DatabaseLoad Balancing in Parallel and Distributed Database
Load Balancing in Parallel and Distributed Database
 
Scalability using Node.js
Scalability using Node.jsScalability using Node.js
Scalability using Node.js
 
Autonomic Decentralised Elasticity Management of Cloud Applications
Autonomic Decentralised Elasticity Management of Cloud ApplicationsAutonomic Decentralised Elasticity Management of Cloud Applications
Autonomic Decentralised Elasticity Management of Cloud Applications
 
Towards True Elasticity of Spark-(Michael Le and Min Li, IBM)
Towards True Elasticity of Spark-(Michael Le and Min Li, IBM)Towards True Elasticity of Spark-(Michael Le and Min Li, IBM)
Towards True Elasticity of Spark-(Michael Le and Min Li, IBM)
 
IEEE 2014 JAVA CLOUD COMPUTING PROJECTS Optimal power allocation and load dis...
IEEE 2014 JAVA CLOUD COMPUTING PROJECTS Optimal power allocation and load dis...IEEE 2014 JAVA CLOUD COMPUTING PROJECTS Optimal power allocation and load dis...
IEEE 2014 JAVA CLOUD COMPUTING PROJECTS Optimal power allocation and load dis...
 
2014 IEEE JAVA CLOUD COMPUTING PROJECT Optimal power allocation and load dist...
2014 IEEE JAVA CLOUD COMPUTING PROJECT Optimal power allocation and load dist...2014 IEEE JAVA CLOUD COMPUTING PROJECT Optimal power allocation and load dist...
2014 IEEE JAVA CLOUD COMPUTING PROJECT Optimal power allocation and load dist...
 
Scylla Summit 2022: Scylla 5.0 New Features, Part 1
Scylla Summit 2022: Scylla 5.0 New Features, Part 1Scylla Summit 2022: Scylla 5.0 New Features, Part 1
Scylla Summit 2022: Scylla 5.0 New Features, Part 1
 
Planning For Catastrophe with IBM WAS and IBM BPM
Planning For Catastrophe with IBM WAS and IBM BPMPlanning For Catastrophe with IBM WAS and IBM BPM
Planning For Catastrophe with IBM WAS and IBM BPM
 
cmandtracing-1560359.pdf
cmandtracing-1560359.pdfcmandtracing-1560359.pdf
cmandtracing-1560359.pdf
 
How Busy Is Too Busy? Automating Your System for Maximum Throughput
How Busy Is Too Busy? Automating Your System for Maximum Throughput How Busy Is Too Busy? Automating Your System for Maximum Throughput
How Busy Is Too Busy? Automating Your System for Maximum Throughput
 
Probabilistic consolidation of virtual machines in self organizing cloud data...
Probabilistic consolidation of virtual machines in self organizing cloud data...Probabilistic consolidation of virtual machines in self organizing cloud data...
Probabilistic consolidation of virtual machines in self organizing cloud data...
 
Power aware load balancing in cloud
Power aware load balancing in cloud Power aware load balancing in cloud
Power aware load balancing in cloud
 

Exploiting latency bounds for energy efficient load balancing

  • 1. Exploiting latency bounds for energy efficient load balancing Cruz Monrreal, Daniel Jones, Michael May and Mohit Taneja The University of Texas at Austin
  • 2. 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
  • 3. 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.
  • 4. System Power Consumption Source - http://static.usenix.org/events/hotpower08/tech/full_papers/rivoire/rivoire_html/
  • 5. System Power Consumption Table 1 - Number of Cores utilized to Power Usage # of Cores Power (W) 0 0 1 270 2 300 3 320 4 330
  • 8. Assumptions - Model 4 core machine 3 servers total 1 job saturates a core Instant on/off No background tasks
  • 9. 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.
  • 10. Current Solutions No Power management (Round Robin) Basic Power management (Round Robin) Advanced Power management
  • 11. Current Solutions - No Power Management Load Scheduler uniformly schedules jobs to each server in sequence Problems: Lots of time spent in idle Few cores used = low efficiency
  • 12. Current Solutions - Round Robin w/o Power Management Load Scheduler uniformly schedules jobs to each server in sequence Turns off unused machines Problems: Few cores used = low efficiency
  • 13. 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
  • 14. 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).
  • 15. Comparisons Run at average load = 25%,50%,75%,100% Vary job time around average load job time. Example: 25% load time Job time = 8-12 milliseconds Stall time = 500 milliseconds Time between jobs = 2-8 milliseconds
  • 16.
  • 18. Limitations With large core counts advantages start to diminish At 100% no gains
  • 19. Future Prospects Storage systems (SANs) Latency exploitation for networks
  • 20. Q&A
  • 21. Resources [1] http://web.eecs.umich.edu/~twenisch/papers/asplos12.pdf [2] http://static.usenix.org/events/hotpower08/tech/full_papers/rivoire/rivoire_html/