Energy Management in Virtualized Environments Gaurav Dhiman, Giacomo Marchetti, Raid Ayoub, Tajana Simunic Rosing (CSE-UCSD) Inside Xen Hypervisor Online.

Slides:

Advertisements

Similar presentations

Energy-efficient Task Scheduling in Heterogeneous Environment 2013/10/25.

Advertisements

Thread Criticality Predictors for Dynamic Performance, Power, and Resource Management in Chip Multiprocessors Abhishek Bhattacharjee Margaret Martonosi.

A Framework for Dynamic Energy Efficiency and Temperature Management (DEETM) Michael Huang, Jose Renau, Seung-Moon Yoo, Josep Torrellas University of Illinois.

Differentiated I/O services in virtualized environments

Difference Engine: Harnessing Memory Redundancy in Virtual Machines by Diwaker Gupta et al. presented by Jonathan Berkhahn.

Energy-Efficient System Virtualization for Mobile and Embedded Systems Final Review 2014/01/21.

1 MemScale: Active Low-Power Modes for Main Memory Qingyuan Deng, David Meisner*, Luiz Ramos, Thomas F. Wenisch*, and Ricardo Bianchini Rutgers University.

Power Management in Cloud Computing using Green Algorithm -Kushal Mehta COP 6087 University of Central Florida.

International Symposium on Low Power Electronics and Design Dynamic Workload Characterization for Power Efficient Scheduling on CMP Systems 1 Gaurav Dhiman,

PVOCL: Power-Aware Dynamic Placement and Migration in Virtualized GPU Environments Palden Lama, Xiaobo Zhou, University of Colorado at Colorado Springs.

VSphere vs. Hyper-V Metron Performance Showdown. Objectives Architecture Available metrics Challenges in virtual environments Test environment and methods.

Project Overview 2014/05/05 1. Current Project “Research on Embedded Hypervisor Scheduler Techniques” ◦ Design an energy-efficient scheduling mechanism.

© 2008 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice In search of a virtual yardstick:

Shimin Chen Big Data Reading Group Presented and modified by Randall Parabicoli.

Towards High-Availability for IP Telephony using Virtual Machines Devdutt Patnaik, Ashish Bijlani and Vishal K Singh.

Disco: Running Commodity Operating Systems on Scalable Multiprocessors Bugnion et al. Presented by: Ahmed Wafa.

Performance and Energy Bounds for Multimedia Applications on Dual-processor Power-aware SoC Platforms Weng-Fai WONG 黄荣辉 Dept. of Computer Science National.

Energy Efficient Web Server Cluster Andrew Krioukov, Sara Alspaugh, Laura Keys, David Culler, Randy Katz.

DPM Dynamic power management. DPM Tree DPM Timeout Adaptive Device dependent Predictive L-shape Exponential average Predictive wakeup Adaptive Disk shutdown.

Processor Frequency Setting for Energy Minimization of Streaming Multimedia Application by A. Acquaviva, L. Benini, and B. Riccò, in Proc. 9th Internation.

By- Jaideep Moses, Ravi Iyer , Ramesh Illikkal and

Synergy.cs.vt.edu Power and Performance Characterization of Computational Kernels on the GPU Yang Jiao, Heshan Lin, Pavan Balaji (ANL), Wu-chun Feng.

1 EE 587 SoC Design & Test Partha Pande School of EECS Washington State University

CS 423 – Operating Systems Design Lecture 22 – Power Management Klara Nahrstedt and Raoul Rivas Spring 2013 CS Spring 2013.

University of Karlsruhe, System Architecture Group Balancing Power Consumption in Multiprocessor Systems Andreas Merkel Frank Bellosa System Architecture.

Software-Hardware Cooperative Power Management Technique for Main Memory So, today I’m going to be talking about a software-hardware cooperative power.

Low Power Techniques in Processor Design

Introduction and Overview Questions answered in this lecture: What is an operating system? How have operating systems evolved? Why study operating systems?

Cloud Computing Energy efficient cloud computing Keke Chen.

Virtual Machine Scheduling for Parallel Soft Real-Time Applications

USTH Presentation Power-aware Scheduler for Virtualization TRAN Giang Son Prof. Daniel HAGIMONT Oct 19th, 2011.

Power and Performance Modeling in a Virtualized Server System M. Pedram and I. Hwang Department of Electrical Engineering Univ. of Southern California.

Thread Criticality Predictors for Dynamic Performance, Power, and Resource Management in Chip Multiprocessors Abhishek Bhattacharjee and Margaret Martonosi.

© 2006 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice Profiling and Modeling Resource Usage.

SAN FRANCISCO, CA, USA Adaptive Energy-efficient Resource Sharing for Multi-threaded Workloads in Virtualized Systems Can HankendiAyse K. Coskun Boston.

(1) Scheduling for Multithreaded Chip Multiprocessors (Multithreaded CMPs)

Energy Savings with DVFS Reduction in CPU power Extra system power.

Our work on virtualization Chen Haogang, Wang Xiaolin {hchen, Institute of Network and Information Systems School of Electrical Engineering.

Dynamic Resource Monitoring and Allocation in a virtualized environment.

An Energy-Efficient Hypervisor Scheduler for Asymmetric Multi- core 1 Ching-Chi Lin Institute of Information Science, Academia Sinica Department of Computer.

1 University of Maryland Linger-Longer: Fine-Grain Cycle Stealing in Networks of Workstations Kyung Dong Ryu © Copyright 2000, Kyung Dong Ryu, All Rights.

IM&T Vacation Program Benjamin Meyer Virtualisation and Hyper-Threading in Scientific Computing.

An Analysis of Efficient Multi-Core Global Power Management Policies Authors: Canturk Isci†, Alper Buyuktosunoglu†, Chen-Yong Cher†, Pradip Bose† and Margaret.

Power-Aware Scheduling of Virtual Machines in DVFS-enabled Clusters

A dynamic optimization model for power and performance management of virtualized clusters Vinicius Petrucci, Orlando Loques Univ. Federal Fluminense Niteroi,

Dynamic and Secure Application Consolidation with Nested Virtualization and Library OS in Cloud Kouta Sannomiya and Kenichi Kourai (Kyushu Institute of.

Dana Butnariu Princeton University EDGE Lab June – September 2011 OPTIMAL SLEEPING IN DATACENTERS Joint work with Professor Mung Chiang, Ioannis Kamitsos,

© 2012 IBM Corporation Platform Computing 1 IBM Platform Cluster Manager Data Center Operating System April 2013.

Towards Dynamic Green-Sizing for Database Servers Mustafa Korkmaz, Alexey Karyakin, Martin Karsten, Kenneth Salem University of Waterloo.

Performance Comparison Xen vs. KVM vs. Native –Benchmarks: SPEC CPU2006, SPEC JBB 2005, SPEC WEB, TPC –Case studies Design instrumentations for figure.

VGreen: A System for Energy Efficient Manager in Virtualized Environments G. Dhiman, G Marchetti, T Rosing ISLPED 2009.

Dynamic Voltage Frequency Scaling for Multi-tasking Systems Using Online Learning Gaurav DhimanTajana Simunic Rosing Department of Computer Science and.

Embedded System Lab. 정범종 A_DRM: Architecture-aware Distributed Resource Management of Virtualized Clusters H. Wang et al. VEE, 2015.

Full and Para Virtualization

Adaptive Sleep Scheduling for Energy-efficient Movement-predicted Wireless Communication David K. Y. Yau Purdue University Department of Computer Science.

Profiling, Prediction, and Capping of Power in Consolidated Environments Bhuvan Urgaonkar Computer Systems Laboratory The Penn State University Talk at.

Research on Embedded Hypervisor Scheduler Techniques 2014/10/02 1.

Spin-down Disk Model Not Spinning Spinning & Ready Spinning & Access Spinning & Seek Spinning up Spinning down Inactivity Timeout threshold* Request Trigger:

E-MOS: Efficient Energy Management Policies in Operating Systems

Prof. Tajana Šimunić Rosing UCSD. Power/Thermal Management in Virtualized Environments CPU0 CPU1 n/w CPU2 CPUn HDD Hypervisor Guest 1 Guest 2 Guest n.

Determining Optimal Processor Speeds for Periodic Real-Time Tasks with Different Power Characteristics H. Aydın, R. Melhem, D. Mossé, P.M. Alvarez University.

Dynamic Power Management Using Online Learning Gaurav Dhiman, Tajana Simunic Rosing (CSE-UCSD) Existing DPM policies do not adapt optimally with changing.

Application-Aware Traffic Scheduling for Workload Offloading in Mobile Clouds Liang Tong, Wei Gao University of Tennessee – Knoxville IEEE INFOCOM

XEN – The Art of Virtualisation. So what is Virtualisation? ● Makes use of spare capacity ● Run multiple instances of OSes simultaneously ● Multitasking.

Jacob R. Lorch Microsoft Research

Resource Aware Scheduler – Initial Results

Comparison of the Three CPU Schedulers in Xen

Research on Embedded Hypervisor Scheduler Techniques

A workload-aware energy model for VM migration

Progress Report 2015/01/28.

Presentation transcript:

Energy Management in Virtualized Environments Gaurav Dhiman, Giacomo Marchetti, Raid Ayoub, Tajana Simunic Rosing (CSE-UCSD) Inside Xen Hypervisor Online Learning Algorithm Virtualization DPMDPM Performs dynamic evaluation of a set of DPM and DVFS policies Performs dynamic evaluation of a set of DPM and DVFS policies at run time and selects the best suited for the current workload at run time and selects the best suited for the current workload Guarantees convergence and performance close to that of the best Guarantees convergence and performance close to that of the best available policy in the set available policy in the set OS implementation and Results Summary Summary Hypervisor VM scheduler implementation Hypervisor VM scheduler implementation  Power Management: DPM/DVFS  Workload characterization aware  Adaptive Behavior Motivations and Goals Lower datacenter energy consumption Lower datacenter energy consumption Handle non-stationary workloads Handle non-stationary workloads Service - VM - Customization Service - VM - Customization Energy Oriented Scheduler Energy Oriented Scheduler -Implements a scheduler capable of adapting to workload (guest) characteristics -Migration: Guest balancing and clustering -Co-locate guests to free up resources -Online Learning Algorithm Supported by NSF-GreenLight project, CNS, Sun Microsystems, UC Micro, Cisco, GSRC/DARPA Supported by NSF-GreenLight project, CNS, Sun Microsystems, UC Micro, Cisco, GSRC/DARPA CPU0CPU0CPU1CPU1N/WN/WCPU2CPU2CPUnCPUnHDDHDD Hypervisor Hypervisor Guest n I/OCPUs Hardware I/O Intensive? CPU Intensive? Guest 1 Guest 2 AppsApps OSOS AppsApps OSOS AppsApps OSOS Credit Scheduler Workload Characterization Online Learning Algorithm VM Scheduling Virtual Machine Power Oriented Scheduling Virtual Machine Power Oriented Scheduling Workload migration across physical machine Workload migration across physical machine Minimize impact on performance Minimize impact on performance Workload characterization Workload characterization - I/O Intensiveness: Maintain metrics - I/O Intensiveness: Maintain metrics for I/O accesses per guest for I/O accesses per guest - CPU Intensiveness: Use CPU - CPU Intensiveness: Use CPU performance counters performance counters CPU intensive (µ ->1) vs Memory intensive (µ -> 0) CPU intensive (µ ->1) vs Memory intensive (µ -> 0) µ = measure of CPU intensiveness µ = measure of CPU intensiveness Leakage impact (ρ) Leakage impact (ρ) DVFSDVFS For qsort Higher energy savings Lower Perf Delay Identifies both CPU-intensive and memory intensive phases correctly Avg. μ time % 75% CPU intensive mem intensive Energy Saving/Performance Delay Results for CPU Experimental Setup Workloads: qsort, djpeg, blowfish, dgzip Workloads: qsort, djpeg, blowfish, dgzip CPU Xscale CPU Xscale Controller Working Set Device :Dormant Experts Expert selection :Operational Expert Manages Power Expert 1 Expert 2 Expert N Expert 3 DPM & DVFS Experimental Setup AMD quad core CPU AMD quad core CPU SPEC benchmarks SPEC benchmarksBenchmarkFreq%delay %Energy savingsPM-i PM-1PM-2PM-3 mcf bzip art sixtrack Device Trace Name t RI σ t RI HDD HDDHP-1Trace HP-2 Trace HP-3 Trace t RI : Average Request Inter-arrival Time (in sec) t RI : Average Request Inter-arrival Time (in sec) ExpertCharacteristics Fixed Timeout Timeout = 7*T be Adaptive Timeout (Douglis, USENIX’95) Initial timeout = 7*T be ; Adjustment = +1T be /-1T be Exponential Predictive (Hwang, ICCAD’97) I n+l = a i n + (1 – a).I n with a = 0.5 TISMDP (Simunic, TCAD’01) Optimized for delay constraint of 3.5% on HP-1 trace PolicyDescriptionPM-1 switch CPU to ACPI state C1 (remove clock supply) and move to lowest voltage setting PM-2 switch CPU to ACPI state C6 (remove power) PM-3 switch CPU to ACPI state C6 and switch the memory to self- refresh mode Recent CPUs might perform better with a “run to sleep” policy due to: Improved CPU efficiency Idle power management support Idle power management support Power/Performance Results for HDD HP-1 trace Comparison with fixed timeout experts