Energy Aware Consolidation for Cloud Computing Srikanaiah, Kansal, Zhao Usenix HotPower 2008.

Slides:



Advertisements
Similar presentations
Ramya (UCSB), Parthasarathy et al (HP Labs). Overview Power delivery, consumption and cooling problems in a data center are being tackled currently by.
Advertisements

Sabyasachi Ghosh Mark Redekopp Murali Annavaram Ming-Hsieh Department of EE USC KnightShift: Enhancing Energy Efficiency by.
Paging: Design Issues. Readings r Silbershatz et al: ,
MINERVA: an automated resource provisioning tool for large-scale storage systems G. Alvarez, E. Borowsky, S. Go, T. Romer, R. Becker-Szendy, R. Golding,
Register Usage Keep as many values in registers as possible Register assignment Register allocation Popular techniques – Local vs. global – Graph coloring.
An Array-Based Algorithm for Simultaneous Multidimensional Aggregates By Yihong Zhao, Prasad M. Desphande and Jeffrey F. Naughton Presented by Kia Hall.
Hadi Goudarzi and Massoud Pedram
Resource Management §A resource can be a logical, such as a shared file, or physical, such as a CPU (a node of the distributed system). One of the functions.
Power Management in Cloud Computing using Green Algorithm -Kushal Mehta COP 6087 University of Central Florida.
ElasticTree: Saving Energy in Data Center Networks Brandon Heller, Srini Seetharaman, Priya Mahadevan, Yiannis Yiakoumis, Puneed Sharma, Sujata Banerjee,
Energy-efficient Virtual Machine Provision Algorithms for Cloud System Ching-Chi Lin Institute of Information Science, Academia Sinica Department of Computer.
Power-aware Resource Allocation for Cpu- and Memory Intense Internet Services Vlasia Anagnostopoulou Susmit Biswas, Heba Saadeldeen,
Green Cloud Computing Hadi Salimi Distributed Systems Lab, School of Computer Engineering, Iran University of Science and Technology,
A Genetic Algorithm for Workload Scheduling In Cloud Based e-Learning Octavian Morariu Cristina Morariu Theodor Borangiu University Politehnica Bucharest.
Efficient Autoscaling in the Cloud using Predictive Models for Workload Forecasting Roy, N., A. Dubey, and A. Gokhale 4th IEEE International Conference.
Power-Aware Routing in Mobile Ad Hoc Networks. Introduction 5 power aware metrics for shortest-cost routing will be presented Compared to the traditional.
A Heuristic Bidding Strategy for Multiple Heterogeneous Auctions Patricia Anthony & Nicholas R. Jennings Dept. of Electronics and Computer Science University.
Three heuristics for transmission scheduling in sensor networks with multiple mobile sinks Damla Turgut and Lotzi Bölöni University of Central Florida.
OS Fall ’ 02 Performance Evaluation Operating Systems Fall 2002.
Energy Efficient Web Server Cluster Andrew Krioukov, Sara Alspaugh, Laura Keys, David Culler, Randy Katz.
On the Task Assignment Problem : Two New Efficient Heuristic Algorithms.
Tree-Building. Methods in Tree Building Phylogenetic trees can be constructed by: clustering method optimality method.
Distributed Process Management1 Learning Objectives Distributed Scheduling Algorithms Coordinator Elections Orphan Processes.
Distributed Systems Meet Economics: Pricing In The Cloud Authors: Hongyi Wang, Qingfeng Jing, Rishan Chen, Bingsheng He, Zhengping He, Lidong Zhou Presenter:
Energy Aware Network Operations Authors: Priya Mahadevan, Puneet Sharma, Sujata Banerjee, Parthasarathy Ranganathan HP Labs IEEE Global Internet Symposium.
Scheduling a Large DataCenter Cliff Stein Columbia University Google Research June, 2009 Monika Henzinger, Ana Radovanovic Google Research.
ElasticTree: Saving Energy in Data Center Networks 許倫愷 2013/5/28.
Energy Efficiency in Cloud Data Centers: Energy Efficient VM Placement for Cloud Data Centers Doctoral Student : Chaima Ghribi Advisor : Djamal Zeghlache.
Dynamic and Decentralized Approaches for Optimal Allocation of Multiple Resources in Virtualized Data Centers Wei Chen, Samuel Hargrove, Heh Miao, Liang.
Dynamic Resource Allocation Using Virtual Machines for Cloud Computing Environment.
07/21/2005 Senmetrics1 Xin Liu Computer Science Department University of California, Davis Joint work with P. Mohapatra On the Deployment of Wireless Sensor.
A Unified Modeling Framework for Distributed Resource Allocation of General Fork and Join Processing Networks in ACM SIGMETRICS
Cloud Computing Energy efficient cloud computing Keke Chen.
Storage Management in Virtualized Cloud Environments Sankaran Sivathanu, Ling Liu, Mei Yiduo and Xing Pu Student Workshop on Frontiers of Cloud Computing,
Cloud Platforms - I Amazon Infrastructure as a Service Cloud Economics - Infrastructure.
RECON: A TOOL TO RECOMMEND DYNAMIC SERVER CONSOLIDATION IN MULTI-CLUSTER DATACENTERS Anindya Neogi IEEE Network Operations and Management Symposium, 2008.
Critical Power Slope Understanding the Runtime Effects of Frequency Scaling Akihiko Miyoshi, Charles Lefurgy, Eric Van Hensbergen Ram Rajamony Raj Rajkumar.
Discussion Week 10 TA: Kyle Dewey. Overview TA Evaluations Project #3 PE 5.1 PE 5.3 PE 11.8 (a,c,d) PE 10.1.
Silberschatz, Galvin and Gagne  Operating System Concepts Chapter 6: CPU Scheduling Basic Concepts Scheduling Criteria Scheduling Algorithms.
CPU Scheduling CSCI 444/544 Operating Systems Fall 2008.
Joint Power Optimization Through VM Placement and Flow Scheduling in Data Centers DAWEI LI, JIE WU (TEMPLE UNIVERISTY) ZHIYONG LIU, AND FA ZHANG (CHINESE.
1 ERCOT LRS Precision Analysis PWG Presentation February 27, 2007.
Managing Server Energy and Operational Costs Chen, Das, Qin, Sivasubramaniam, Wang, Gautam (Penn State) Sigmetrics 2005.
Register Usage Keep as many values in registers as possible Keep as many values in registers as possible Register assignment Register assignment Register.
Presented by: Dardan Xhymshiti Spring 2016:. Authors: Publication:  ICDM 2015 Type:  Research Paper 2 Michael ShekelyamGregor JosseMatthias Schubert.
Thermal Management in Datacenters Ayan Banerjee. Thermal Management using task placement Tasks: Requires a certain number of servers (cores) for a specified.
1 PerfCenter and AutoPerf: Tools and Techniques for Modeling and Measurement of the Performance of Distributed Applications Varsha Apte Faculty Member,
Determining Optimal Processor Speeds for Periodic Real-Time Tasks with Different Power Characteristics H. Aydın, R. Melhem, D. Mossé, P.M. Alvarez University.
Architecture for Resource Allocation Services Supporting Interactive Remote Desktop Sessions in Utility Grids Vanish Talwar, HP Labs Bikash Agarwalla,
BIN SORTING Problem Pack the following items in bins of size Firstly, find the lower bound by summing the numbers to be packed.
A Hierarchical Edge Cloud Architecture for Mobile Computing IEEE INFOCOM 2016 Liang Tong, Yong Li and Wei Gao University of Tennessee – Knoxville 1.
Energy Aware Network Operations
Optimizing Distributed Actor Systems for Dynamic Interactive Services
CPU Scheduling CSSE 332 Operating Systems
Reinforcement Learning Based Virtual Cluster Management
Adel Nadjaran Toosi and Rajkumar Buyya
CS 425 / ECE 428 Distributed Systems Fall 2016 Nov 10, 2016
Analyzing Security and Energy Tradeoffs in Autonomic Capacity Management Wei Wu.
ElasticTree Michael Fruchtman.
CS 425 / ECE 428 Distributed Systems Fall 2017 Nov 16, 2017
Adaptive Cloud Computing Based Services for Mobile Users
Bank-aware Dynamic Cache Partitioning for Multicore Architectures
Xia Zhao*, Zhiying Wang+, Lieven Eeckhout*
Zhen Xiao, Qi Chen, and Haipeng Luo May 2013
ElasticTree: Saving Energy in Data Center Networks
Clustering.
CSCI1600: Embedded and Real Time Software
Rethinking Cost and Performance of Database Systems
Flexible Assembly Systems
IIS Progress Report 2016/01/18.
Presentation transcript:

Energy Aware Consolidation for Cloud Computing Srikanaiah, Kansal, Zhao Usenix HotPower 2008

Power & Consolidation Issues  High power consumption even at low load (at 10% CPU util, 50% of peak power consumed)  Consolidation is not just bin-packing –Packing too much might increase “energy used per unit service provided” –Consolidation can lead to performance degradation –There exists an “optimal” performance vs energy operating point

Experiments for Understanding Consolidation Experimental Set up

Understanding Consolidation: Experimental Results Started with app using 10% CPU, 10% disk Added workloads with varying CPU- disk utils Numbers show CPU-disk utilization mix Point made: consolidation results in performance degradation

Energy Consumption per transaction U-shaped curve: At low utilizations idle power is not “amortized effectively” At high utilizations, energy consumption increases, but throughput degrades hence per tran energy consumption increases Observation: There is an optimal combination of CPU and disk utils for this setup (70% CPU, 50% disk)

Consolidation Problem  Why not straightforward multi-dimensional bin-packing –Performance degradation: resource utilizations are additive, but performance measures are not modeled at all in bin-packing. Minimizing number of bins not equal to minimizing energy (implied to mean energy per transaction) –Power variation: even if minimum number of severs is used, the allocation of workloads will result in varying power usage

Consolidation Algorithm  Algorithm has to do following –Workload “arrives” to host cluster –Arriving workload has known CPU-disk utilization –Arriving workload has to be “assigned” to a host

Consolidation Algorithm 1.Optimal point for each host should be known (details not specified!!)  Assumes that optimal point is a host characteristic - not dependent on application 2.For a particular allocation, the “Euclidean distance” from the optimal point is calculated 3.Pick allocation which maximizes sum of such Euclidiean distances of each server  In authors’ words “This heuristic is based on the intuition that we can use both dimensions of a bin to the fullest (where “full” is defined as the optimal utilization point) after the current allocation is done, if we are left with maximum empty space in each dimension after the allocation.”

Consolidation Heuristic

Evaluation

…Evaluation