Exploring Workload Patterns for Saving Power
|
|
- Matilda Andrews
- 5 years ago
- Views:
Transcription
1 Exploring Workload Patterns for Saving Power Evgenia Smirni College of William & Mary joint work with Andrew Caniff, Lei Lu, Ningfang Mi (William and Mary), Lucy Cherkasova, HP Labs Robert Birke and Lydia Chen, IBM Zurich Lab
2 Overview Mo*va*on (big picture view) Data centers: workload characteriza*on Is burs*ness for real? Burs*ness (focus on the small picture) Index of dispersion: to capture burs*ness Alloca*on algorithm: Fastract Performance evalua*on Open problems 2
3 Motivation Traffic surges are frequent but chanllenging Slashdot effect : a web page linked by a popular blog or media site suddenly experices a huge increase of number of hits Auction sites (e.g., ebay), special offers or marketing campains Burstiness or temporal surges leads to uncontrolled increase of response time or resource overprovisioning Image from 3
4 Motivation Traffic surges are frequent but chanllenging Slashdot effect : a web page linked by a popular blog or media site suddenly experices a huge increase of number of hits Auction sites (e.g., ebay), special offers or marketing campains Burstiness or temporal surges leads to uncontrolled increase of response time or resource overprovisioning 4
5 Motivation: looking at mature Data Centers ALL IBM data centers (thousands of servers) Two year study (June 2009 to May 2011) Global view Empirical distributions of resource utilizations Different time scales Two year time frame Macro view vs micro view Small one: 393 servers Large one: 3672 servers Expected: overprovisioning Seasonality: patterns Resource demand trends: patterns, burstiness 5
6 Motivation: looking at mature Data Centers Utilizations across basic resources Hidden CPU peaks Measurement granularities 6
7 Motivation: looking at mature Data Centers Looking at CPU Hidden CPU peaks Some fitting (elementary) 7
8 Motivation: looking at mature Data Centers Micro view: two specific enterprises 8
9 Motivation: looking at mature Data Centers Macro view: two specific data centers Some shared servers, some not Small data center view Large data center view 9
10 Motivation: looking at mature Data Centers Small data center: enterprise 1 10
11 Motivation: looking at mature Data Centers Small data center: enterprise 2 11
12 Motivation: looking at mature Data Centers Small data center: enterprise 3 12
13 Motivation: looking at mature Data Centers Large data center: enterprise 5 13
14 Motivation: looking at mature Data Centers Large data center: enterprise 6 14
15 Summary Big differences on how each client uses the DC CPU is consistently overprovisioned Disk utilizations are usually very high Dips in across time graphs! Upgrades/where the money is! 15
16 Summary (cont.) Overprovisioning Utilization vs. other measures Arrival/service processes Google data center data Google cloud clusters See: code.google.com/p/googleclusterdata/wiki/ ClusterData2011_1 Arrival process is clearly bursty (high autocorrelation) 16
17 Summary (cont.) Overprovisioning Utilization vs. other measures Arrival/service processes Google data center data Google cloud clusters See: code.google.com/p/googleclusterdata/wiki/ ClusterData2011_1 Arrival process is clearly bursty (high autocorrelation) 17
18 Challenge Back into the micro view... [Related work] Multi-tiered systems Single-tier dynamic provisioning Bottleneck switch Determining how many servers to provision and where is far more complex. Bursty/seasonal workloads Design an analytical model of multi-tier application to capture the required capacity of different tiers Burstiness precludes the use of classic analytical models Need simple models of seasonality 18
19 Outline Mo*va*on & Challenges Index of dispersion: Capture burs*ness Alloca*on algorithm: Fastract Performance Evalua*on Future work & Conclusions 19
20 Index of Dispersion: Capture Burstiness A metric measures variability and temporal structure Variability Temporal structure Squared Coeffient of Varia9on Lag k autocorrela9on I = SCV (1 + 2 ρk ) k = 1 20
21 Index of Dispersion Definition [Cox1966][Gusella1991] Given a time series of random variables {Xn}, where n = 0,...,, the index of dispersion is defined: I = SCV (1 + 2 ρk ) k = 1 2 where, quantifies the variability in the series, = C ov( X, X ) / Var( X ) lag-k autocorrelation is the cross-correlation of the time series variable against a k stepshifted version of itself. SCV = Var( X ) / E ( X ) ρ k i i+ k 21
22 Index of Dispersion: Example (1) Poisson Process I =1 Renewal process (i.i.d.) (e.g., HyperExponential) 2 I = SCV = Var( X ) / E ( X ) Markov Modulated Poisson Process (MMPP(2)) I 2 σ σ ( λ λ ) = 1 + ( σ σ ) ( λ σ λ σ ) I = Classic FIFA World Cup trace 22
23 Index of Dispersion: Example (2) Same mean and SCV Hyper Exponential (iid) versus MAP I = 20 I =
24 Challenge: What to do with? Using batch size K = avg arrival rate C (we use C = 2000) Adding the first 10% of K instead of I value detects the workload change I = I = 1~107 I = I = 3~
25 Detect Burstiness: a close view Algorithm for detecting burstiness Together with arrival rate to know whether a burst starts or burst ends Detec9ng Burst Burst Start Burst Ends 1. current_i calculate I value for current batch 2. current_rate arrival rate for current *me 3. Update total average arrival rate and SCV 4. IF ( current_i old_i > 2*SCV ) AND (old_i / current_i> 2 OR old_i/current_i < 0.5) I = I = 1~107 I = IF (current_rate > total_rate) burst starts 6. ELSE burst I = ends 3~
26 Detect Burstiness: a big picture Capture the beginning and end of most bursty periods -- not really perfect Slower in the detection of a quiet period Taking longer to gather after the conclusion of burst 26
27 Talk Outline Mo*va*on Index of dispersion: Capture burs*ness Alloca*on algorithm: Fastract Performance Evalua*on Future work & Conclusions 27
28 System Overview TCP-W 2-tier implementation 8 potential front servers, 1 database server Focus on front tier allocation Users Web + App server Dispatcher Database Front server pool 28
29 Fastract algorithm (1) Proactive adjustment Reactive adjustment Proac9ve Adjustment (Primary Resource Alloca9on for request burst Management ) 1. Increase front servers when burst starts and front server is the bovleneck 2. Decrease front serves when burst ends Burst starts!! Users Database Dispatcher Front server pool 29
30 Fastract algorithm (2) Proactive adjustment Reactive adjustment Proac9ve Adjustment (Primary Resource Alloca9on for request burst Management ) 1. Increase front servers when burst starts and front server is the bo`leneck 2. Decrease front serves when burst ends Burst ends! Users Database Dispatcher Front server pool 30
31 Fastract algorithm (3) Proactive adjustment Reactive adjustment Reac9ve Adjustment (Secondary Resource Alloca9on for SLA guarantees) 1. Increase front servers, when SLO is violated, front server is the bovleneck and current number of front servers can not sa9sfy SLO 2. Decrease front servers, when SLO is sa*sfied and front server is not the bo`leneck Users SLO Database Dispatcher Front server pool 31
32 Fastract algorithm (4) Proactive adjustment Reactive adjustment Reac9ve Adjustment (Secondary Resource Alloca9on for SLA guarantees) 1. Increase front servers, when SLO is violated, front server is the bo`leneck and current number of front servers can not sa*sfy SLO 2. Decrease front servers, when SLO is sa9sfied and front server is not the bovleneck SLO Users Database Dispatcher Front server pool 32
33 Experiment setup Different transaction mix: Browsing mix: 95% browsing and 5% ordering Shopping mix: 80% browsing and 20% ordering Ordering mix: 50% browsing and 50% ordering Think time of emulated browsers (EB) are modeled by MAPs (injecting burstiness) [ICAC 09] Mean = 7 seconds, SCV = 20, Index of dispersion = 41 (low) and 1806 (high) Using a MAP for the service processes in front and database servers [Middleware 08] Compare results to systems with a static number of server Power usage Raw machine seconds 33
34 Evaluation results (1) Persistent bottleneck: ordering mix Front servers are always the bottleneck Using minimal power while reaching SLO target Transient behavior of ordering mix EB =
35 Evaluation results (2) Persistent Bottleneck: shopping mix Database is the system bottleneck and dominate the end-toend response time Do not increase front server even in the presence of a burst Bottleneck switch: browsing mix Locate bottleneck before provisioning 35
36 Interesting open problems (to a simple problem ) Prediction of burst length More details about power usage measurement Cost of power up/power down Real (more complicated) problems abound Virtualization Multiple resources Sharing? How to share best? Best match? How to give different priorities? SLOs? Multiclass vs single class? Effective modeled-based allocation [HPDC 12, DSN- PDS 12 (to appear) 36
37 Acknowledgements: Work funded by NSF grant NSF, #CCF , Computing and Communication Foundations, Computing Processes and Artifacts, (CPA-ACR-CSA) Effective Resource Allocation under Temporal Dependence Data center workload characterization has been done during my sabbatical leave at IBM-Research Zurich 37
38 38
Injecting Realistic Burstiness to a Traditional Client-Server Benchmark
Injecting Realistic Burstiness to a Traditional Client-Server Benchmark Ningfang Mi, Giuliano Casale, Ludmila Cherkasova, Evgenia Smirni HP Laboratories HPL-29-114 Keyword(s): Client-server benchmarks,
More informationPerformance impacts of autocorrelated flows in multi-tiered systems
Performance Evaluation ( ) www.elsevier.com/locate/peva Performance impacts of autocorrelated flows in multi-tiered systems Ningfang Mi a,, Qi Zhang b, Alma Riska c, Evgenia Smirni a, Erik Riedel c a College
More informationBurstiness in Multi-tier Applications: Symptoms, Causes, and New Models
Burstiness in Multi-tier Applications: Symptoms, Causes, and New Models Ningfang Mi 1, Giuliano Casale 1, Ludmila Cherkasova 2, and Evgenia Smirni 1 1 College of William and Mary, Williamsburg, VA 23187,
More informationModel-Driven Geo-Elasticity In Database Clouds
Model-Driven Geo-Elasticity In Database Clouds Tian Guo, Prashant Shenoy College of Information and Computer Sciences University of Massachusetts, Amherst This work is supported by NSF grant 1345300, 1229059
More informationWorkload Prediction For Efficient Performance Isolation And System Reliability
College of William and Mary W&M ScholarWorks Dissertations, Theses, and Masters Projects Theses, Dissertations, & Master Projects Spring 217 Workload Prediction For Efficient Performance Isolation And
More informationPerformance-Guided Load (Un)Balancing Under Autocorrelated Flows
Performance-Guided Load (Un)Balancing Under Autocorrelated Flows Qi Zhang Ningfang Mi Alma Riska 2 Evgenia Smirni Department of Computer Science, College of William and Mary, Williamsburg, VA 2 Seagate
More informationTowards Energy Proportionality for Large-Scale Latency-Critical Workloads
Towards Energy Proportionality for Large-Scale Latency-Critical Workloads David Lo *, Liqun Cheng *, Rama Govindaraju *, Luiz André Barroso *, Christos Kozyrakis Stanford University * Google Inc. 2012
More informationA regression-based analytic model for capacity planning of multi-tier applications
Cluster Comput (2008) 11: 197 211 DOI 10.1007/s10586-008-0052-0 A regression-based analytic model for capacity planning of multi-tier applications Qi Zhang Ludmila Cherkasova Ningfang Mi Evgenia Smirni
More informationPrediction-Based Admission Control for IaaS Clouds with Multiple Service Classes
Prediction-Based Admission Control for IaaS Clouds with Multiple Service Classes Marcus Carvalho, Daniel Menascé, Francisco Brasileiro 2015 IEEE Intl. Conf. Cloud Computing Technology and Science Summarized
More informationApplication Performance Management in the Cloud using Learning, Optimization, and Control
Application Performance Management in the Cloud using Learning, Optimization, and Control Xiaoyun Zhu May 9, 2014 2014 VMware Inc. All rights reserved. Rising adoption of cloud-based services 47% 34% Source:
More informationCloud Computing. What is cloud computing. CS 537 Fall 2017
Cloud Computing CS 537 Fall 2017 What is cloud computing Illusion of infinite computing resources available on demand Scale-up for most apps Elimination of up-front commitment Small initial investment,
More informationContents. Abstract Motivation... 63
Contents Abstract 3 1 Introduction 4 1.1 Computational Resource Management in Cluster Systems.......... 5 1.2 Data Management in Multi-tiered Storage Systems.............. 6 1.3 Data Management in Flash-based
More informationStorage Workload Isolation via Tier Warming: How Models Can Help
Storage Workload Isolation via Tier Warming: How Models Can Help By Ji Xue, Feng Yan, Alma Riska, and Evgenia Smirni Presented By Christian Contreras Proposed solution Storage workload prediction model
More informationVersatile Models of Systems Using MAP Queueing Networks
Versatile Models of Systems Using MAP Queueing Networks Giuliano Casale, Ningfang Mi, and Evgenia Smirni College of William and Mary Department of Computer Science Williamsburg, VA { casale, ningfang,
More informationBuilding Adaptive Performance Models for Dynamic Resource Allocation in Cloud Data Centers
Building Adaptive Performance Models for Dynamic Resource Allocation in Cloud Data Centers Jin Chen University of Toronto Joint work with Gokul Soundararajan and Prof. Cristiana Amza. Today s Cloud Pay
More informationNested QoS: Providing Flexible Performance in Shared IO Environment
Nested QoS: Providing Flexible Performance in Shared IO Environment Hui Wang Peter Varman Rice University Houston, TX 1 Outline Introduction System model Analysis Evaluation Conclusions and future work
More informationMean Value Analysis and Related Techniques
Mean Value Analysis and Related Techniques 34-1 Overview 1. Analysis of Open Queueing Networks 2. Mean-Value Analysis 3. Approximate MVA 4. Balanced Job Bounds 34-2 Analysis of Open Queueing Networks Used
More informationAutomated Control for Elastic Storage
Automated Control for Elastic Storage Summarized by Matthew Jablonski George Mason University mjablons@gmu.edu October 26, 2015 Lim, H. C. and Babu, S. and Chase, J. S. (2010) Automated Control for Elastic
More informationMulti-tenancy version of BigDataBench
Multi-tenancy version of BigDataBench Gang Lu Institute of Computing Technology, Chinese Academy of Sciences BigDataBench Tutorial MICRO 2014 Cambridge, UK INSTITUTE OF COMPUTING TECHNOLOGY 1 Multi-tenancy
More informationAutomated Control for Elastic Storage Harold Lim, Shivnath Babu, Jeff Chase Duke University
D u k e S y s t e m s Automated Control for Elastic Storage Harold Lim, Shivnath Babu, Jeff Chase Duke University Motivation We address challenges for controlling elastic applications, specifically storage.
More informationHigh Performance Resource Allocation and Request Redirection Algorithms for Web Clusters
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS 1 High Performance Resource Allocation and Request Redirection Algorithms for Web Clusters Supranamaya Ranjan, Member, IEEE, and Edward Knightly, Senior
More informationBigDataBench-MT: Multi-tenancy version of BigDataBench
BigDataBench-MT: Multi-tenancy version of BigDataBench Gang Lu Beijing Academy of Frontier Science and Technology BigDataBench Tutorial, ASPLOS 2016 Atlanta, GA, USA n Software perspective Multi-tenancy
More informationLies, Damn Lies and Performance Metrics. PRESENTATION TITLE GOES HERE Barry Cooks Virtual Instruments
Lies, Damn Lies and Performance Metrics PRESENTATION TITLE GOES HERE Barry Cooks Virtual Instruments Goal for This Talk Take away a sense of how to make the move from: Improving your mean time to innocence
More informationCh. 7: Benchmarks and Performance Tests
Ch. 7: Benchmarks and Performance Tests Kenneth Mitchell School of Computing & Engineering, University of Missouri-Kansas City, Kansas City, MO 64110 Kenneth Mitchell, CS & EE dept., SCE, UMKC p. 1/3 Introduction
More informationTripS: Automated Multi-tiered Data Placement in a Geo-distributed Cloud Environment
TripS: Automated Multi-tiered Data Placement in a Geo-distributed Cloud Environment Kwangsung Oh, Abhishek Chandra, and Jon Weissman Department of Computer Science and Engineering University of Minnesota
More informationEnterprise System Installation And. Performance Evaluation
Enterprise System Installation And Performance Evaluation A Thesis Presented by Jiahui Chen to The Department of Electrical and Computer Engineering in partial fulfillment of the requirements for the degree
More informationGame Traffic Analysis: An MMORPG Perspective
Appeared in ACM NOSSDAV 2005 (15th International Workshop on Network and Operating System Support for Digital Audio and Video) Game Traffic Analysis: An MMORPG Perspective (MMORPG: Massive Multiplayer
More informationWeek 7: Traffic Models and QoS
Week 7: Traffic Models and QoS Acknowledgement: Some slides are adapted from Computer Networking: A Top Down Approach Featuring the Internet, 2 nd edition, J.F Kurose and K.W. Ross All Rights Reserved,
More informationAn Experimental Study of Rapidly Alternating Bottleneck in n-tier Applications
An Experimental Study of Rapidly Alternating Bottleneck in n-tier Applications Qingyang Wang, Yasuhiko Kanemasa, Jack Li, Deepal Jayasinghe, Toshihiro Shimizu, Masazumi Matsubara, Motoyuki Kawaba, Calton
More informationConsolidating Complementary VMs with Spatial/Temporalawareness
Consolidating Complementary VMs with Spatial/Temporalawareness in Cloud Datacenters Liuhua Chen and Haiying Shen Dept. of Electrical and Computer Engineering Clemson University, SC, USA 1 Outline Introduction
More informationDaniel A. Menascé, Ph. D. Dept. of Computer Science George Mason University
Daniel A. Menascé, Ph. D. Dept. of Computer Science George Mason University menasce@cs.gmu.edu www.cs.gmu.edu/faculty/menasce.html D. Menascé. All Rights Reserved. 1 Benchmark System Under Test (SUT) SPEC
More informationA New Metric for Analyzing Storage System Performance Under Varied Workloads
A New Metric for Analyzing Storage System Performance Under Varied Workloads Touch Rate Steven Hetzler IBM Fellow Manager, Cloud Data Architecture Flash Memory Summit 2015 Steven Hetzler. IBM 1 Overview
More informationQuantifying Load Imbalance on Virtualized Enterprise Servers
Quantifying Load Imbalance on Virtualized Enterprise Servers Emmanuel Arzuaga and David Kaeli Department of Electrical and Computer Engineering Northeastern University Boston MA 1 Traditional Data Centers
More informationDecoupling Datacenter Studies from Access to Large-Scale Applications: A Modeling Approach for Storage Workloads
Decoupling Datacenter Studies from Access to Large-Scale Applications: A Modeling Approach for Storage Workloads Christina Delimitrou 1, Sriram Sankar 2, Kushagra Vaid 2, Christos Kozyrakis 1 1 Stanford
More informationA Large-Scale Study of Soft- Errors on GPUs in the Field
A Large-Scale Study of Soft- Errors on GPUs in the Field Bin Nie*, Devesh Tiwari +, Saurabh Gupta +, Evgenia Smirni*, and James H. Rogers + *College of William and Mary + Oak Ridge National Laboratory
More informationPower Provisioning for Diverse Datacenter Workloads
Power Provisioning for Diverse Datacenter Workloads Christopher Stewart The Ohio State University Jing Li The Ohio State University Datacenters: A New Research Area Datacenters are the de facto architecture
More informationResource Allocation Strategies for Multiple Job Classes
Resource Allocation Strategies for Multiple Job Classes by Ye Hu A thesis presented to the University of Waterloo in fulfillment of the thesis requirement for the degree of Master of Mathematics in Computer
More informationDiagnostics in Testing and Performance Engineering
Diagnostics in Testing and Performance Engineering This document talks about importance of diagnostics in application testing and performance engineering space. Here are some of the diagnostics best practices
More informationEfficient Evaluation and Management of Temperature and Reliability for Multiprocessor Systems
Efficient Evaluation and Management of Temperature and Reliability for Multiprocessor Systems Ayse K. Coskun Electrical and Computer Engineering Department Boston University http://people.bu.edu/acoskun
More informationI/O Characterization of Commercial Workloads
I/O Characterization of Commercial Workloads Kimberly Keeton, Alistair Veitch, Doug Obal, and John Wilkes Storage Systems Program Hewlett-Packard Laboratories www.hpl.hp.com/research/itc/csl/ssp kkeeton@hpl.hp.com
More informationG Robert Grimm New York University
G22.3250-001 Receiver Livelock Robert Grimm New York University Altogether Now: The Three Questions What is the problem? What is new or different? What are the contributions and limitations? Motivation
More informationMean Value Analysis and Related Techniques
Mean Value Analysis and Related Techniques Raj Jain Washington University in Saint Louis Saint Louis, MO 63130 Jain@cse.wustl.edu Audio/Video recordings of this lecture are available at: 34-1 Overview
More informationPerformance and Scalability: Tuning, Testing, and Monitoring
Performance and Scalability: Tuning, Testing, and Monitoring Andrew Sakowicz, asakowicz@esri.com Steve McCarthy, Steven.McCarthy@Williams.com Frank Pizzi, fpizzi@esri.com Agenda Process, Tools, Value Performance
More informationReplicate It! Scalable Content Delivery: Why? Scalable Content Delivery: How? Scalable Content Delivery: How? Scalable Content Delivery: What?
Accelerating Internet Streaming Media Delivery using Azer Bestavros and Shudong Jin Boston University http://www.cs.bu.edu/groups/wing Scalable Content Delivery: Why? Need to manage resource usage as demand
More informationThe next step in Software-Defined Storage with Virtual SAN
The next step in Software-Defined Storage with Virtual SAN Osama I. Al-Dosary VMware vforum, 2014 2014 VMware Inc. All rights reserved. Agenda Virtual SAN s Place in the SDDC Overview Features and Benefits
More informationTypical scenario in shared infrastructures
Got control? AutoControl: Automated Control of MultipleVirtualized Resources Pradeep Padala, Karen Hou, Xiaoyun Zhu*, Mustfa Uysal, Zhikui Wang, Sharad Singhal, Arif Merchant, Kang G. Shin University of
More informationPublic Cloud Leverage For IT/Business Alignment Business Goals Agility to speed time to market, adapt to market demands Elasticity to meet demand whil
LHC2386BU True Costs Savings Modeling and Costing A Migration to VMware Cloud on AWS Chris Grossmeier chrisg@cloudphysics.com John Blumenthal john@cloudphysics.com #VMworld Public Cloud Leverage For IT/Business
More informationData Centers and Cloud Computing. Slides courtesy of Tim Wood
Data Centers and Cloud Computing Slides courtesy of Tim Wood 1 Data Centers Large server and storage farms 1000s of servers Many TBs or PBs of data Used by Enterprises for server applications Internet
More informationAmit. Amit - Active Middleware. Technology Overview. IBM Research Lab in Haifa Active Technologies October 2002
Amit Amit - Active Middleware Technology Overview IBM Research Lab in Haifa Active Technologies October 2002 OUTLINE: The Active Technologies Amit Active Middleware Technology Related Active Management
More informationUsing Alluxio to Improve the Performance and Consistency of HDFS Clusters
ARTICLE Using Alluxio to Improve the Performance and Consistency of HDFS Clusters Calvin Jia Software Engineer at Alluxio Learn how Alluxio is used in clusters with co-located compute and storage to improve
More informationData Centers and Cloud Computing. Data Centers
Data Centers and Cloud Computing Slides courtesy of Tim Wood 1 Data Centers Large server and storage farms 1000s of servers Many TBs or PBs of data Used by Enterprises for server applications Internet
More informationA Regression-Based Analytic Model for Dynamic Resource Provisioning of Multi-Tier Applications
A Regression-Based Analytic Model for Dynamic Resource Provisioning of Multi-Tier Applications Qi Zhang College of William and Mary Williamsburg, VA 23187, USA Ludmila Cherkasova Hewlett-Packard Labs Palo
More informationLeveraging Software-Defined Storage to Meet Today and Tomorrow s Infrastructure Demands
Leveraging Software-Defined Storage to Meet Today and Tomorrow s Infrastructure Demands Unleash Your Data Center s Hidden Power September 16, 2014 Molly Rector CMO, EVP Product Management & WW Marketing
More informationYouChoose: A Performance Interface Enabling Convenient and Efficient QoS Support for Consolidated Storage Systems
YouChoose: A Performance Interface Enabling Convenient and Efficient QoS Support for Consolidated Storage Systems Xuechen Zhang Yuhai Xu Song Jiang The ECE Department Wayne State University Detroit, MI
More informationVirtualization. Dr. Yingwu Zhu
Virtualization Dr. Yingwu Zhu Virtualization Definition Framework or methodology of dividing the resources of a computer into multiple execution environments. Types Platform Virtualization: Simulate a
More informationDesign Considerations for Using Flash Memory for Caching
Design Considerations for Using Flash Memory for Caching Edi Shmueli, IBM XIV Storage Systems edi@il.ibm.com Santa Clara, CA August 2010 1 Solid-State Storage In a few decades solid-state storage will
More informationTraffic Characteristics of Bulk Data Transfer using TCP/IP over Gigabit Ethernet
Traffic Characteristics of Bulk Data Transfer using TCP/IP over Gigabit Ethernet Aamir Shaikh and Kenneth J. Christensen Department of Computer Science and Engineering University of South Florida Tampa,
More informationMulti-tier architecture performance analysis. Papers covered
Multi-tier architecture performance analysis Papers covered Emmanuel Cecchet, Julie Marguerie, Willy Zwaenepoel: Performance and Scalability of EJB Applications. OOPSLA 02 Yan Liu, Alan Fekete, Ian Gorton:
More informationHeterogeneous Resources Management In Modern Data Centers with Dynamic Workloads Ningfang Mi
Heterogeneous Resources Management In Modern Data Centers with Dynamic Workloads Ningfang Mi Electrical and Computer Engineering Dept. Northeastern University ningfang@ece.neu.edu 1 Research Focus To investigate
More informationDynamic control and Resource management for Mission Critical Multi-tier Applications in Cloud Data Center
Institute Institute of of Advanced Advanced Engineering Engineering and and Science Science International Journal of Electrical and Computer Engineering (IJECE) Vol. 6, No. 3, June 206, pp. 023 030 ISSN:
More informationFast and Accurate Load Balancing for Geo-Distributed Storage Systems
Fast and Accurate Load Balancing for Geo-Distributed Storage Systems Kirill L. Bogdanov 1 Waleed Reda 1,2 Gerald Q. Maguire Jr. 1 Dejan Kostic 1 Marco Canini 3 1 KTH Royal Institute of Technology 2 Université
More informationLecture 5: Performance Analysis I
CS 6323 : Modeling and Inference Lecture 5: Performance Analysis I Prof. Gregory Provan Department of Computer Science University College Cork Slides: Based on M. Yin (Performability Analysis) Overview
More informationCPSC 426/526. Cloud Computing. Ennan Zhai. Computer Science Department Yale University
CPSC 426/526 Cloud Computing Ennan Zhai Computer Science Department Yale University Recall: Lec-7 In the lec-7, I talked about: - P2P vs Enterprise control - Firewall - NATs - Software defined network
More informationWhen Average is Not Average: Large Response Time Fluctuations in n-tier Applications. Qingyang Wang, Yasuhiko Kanemasa, Calton Pu, Motoyuki Kawaba
When Average is Not Average: Large Response Time Fluctuations in n-tier Applications Qingyang Wang, Yasuhiko Kanemasa, Calton Pu, Motoyuki Kawaba Background & Motivation Analysis of the Large Response
More informationvsan Mixed Workloads First Published On: Last Updated On:
First Published On: 03-05-2018 Last Updated On: 03-05-2018 1 1. Mixed Workloads on HCI 1.1.Solution Overview Table of Contents 2 1. Mixed Workloads on HCI 3 1.1 Solution Overview Eliminate the Complexity
More informationDDSS: Dynamic Dedicated Servers Scheduling for Multi Priority Level Classes in Cloud Computing
DDSS: Dynamic Dedicated Servers Scheduling for Multi Priority Level Classes in Cloud Computing Husnu Saner Narman Md. Shohrab Hossain Mohammed Atiquzzaman School of Computer Science University of Oklahoma,
More informationCopyright 2018, Oracle and/or its affiliates. All rights reserved.
Beyond SQL Tuning: Insider's Guide to Maximizing SQL Performance Monday, Oct 22 10:30 a.m. - 11:15 a.m. Marriott Marquis (Golden Gate Level) - Golden Gate A Ashish Agrawal Group Product Manager Oracle
More informationCisco SAN Analytics and SAN Telemetry Streaming
Cisco SAN Analytics and SAN Telemetry Streaming A deeper look at enterprise storage infrastructure The enterprise storage industry is going through a historic transformation. On one end, deep adoption
More informationOASIS: Self-tuning Storage for Applications
OASIS: Self-tuning Storage for Applications Kostas Magoutis, Prasenjit Sarkar, Gauri Shah 14 th NASA Goddard- 23 rd IEEE Mass Storage Systems Technologies, College Park, MD, May 17, 2006 Outline Motivation
More informationGetting it Right: Testing Storage Arrays The Way They ll be Used
Getting it Right: Testing Storage Arrays The Way They ll be Used Peter Murray Virtual Instruments Flash Memory Summit 2017 Santa Clara, CA 1 The Journey: How Did we Get Here? Storage testing was black
More informationDemystifying the Cloud With a Look at Hybrid Hosting and OpenStack
Demystifying the Cloud With a Look at Hybrid Hosting and OpenStack Robert Collazo Systems Engineer Rackspace Hosting The Rackspace Vision Agenda Truly a New Era of Computing 70 s 80 s Mainframe Era 90
More informationThe Elasticity and Plasticity in Semi-Containerized Colocating Cloud Workload: a view from Alibaba Trace
The Elasticity and Plasticity in Semi-Containerized Colocating Cloud Workload: a view from Alibaba Trace Qixiao Liu* and Zhibin Yu Shenzhen Institute of Advanced Technology Chinese Academy of Science @SoCC
More informationArcGIS in the Cloud. Andrew Sakowicz & Alec Walker
ArcGIS in the Cloud Andrew Sakowicz & Alec Walker Key Takeaways How to Identify Organizational Strategy & Priorities Esri s Cloud Offerings A Broad Spectrum Successfully Executing Your Strategy The Cloud
More informationPerformance Extrapolation across Servers
Performance Extrapolation across Servers Subhasri Duttagupta www.cmgindia.org 1 Outline Why do performance extrapolation across servers? What are the techniques for extrapolation? SPEC-Rates of servers
More informationFrom Internet Data Centers to Data Centers in the Cloud
From Internet Data Centers to Data Centers in the Cloud This case study is a short extract from a keynote address given to the Doctoral Symposium at Middleware 2009 by Lucy Cherkasova of HP Research Labs
More informationEliminate the Complexity of Multiple Infrastructure Silos
SOLUTION OVERVIEW Eliminate the Complexity of Multiple Infrastructure Silos A common approach to building out compute and storage infrastructure for varying workloads has been dedicated resources based
More informationKubernetes Integration with Virtuozzo Storage
Kubernetes Integration with Virtuozzo Storage A Technical OCTOBER, 2017 2017 Virtuozzo. All rights reserved. 1 Application Container Storage Application containers appear to be the perfect tool for supporting
More informationVolley: Automated Data Placement for Geo-Distributed Cloud Services
Volley: Automated Data Placement for Geo-Distributed Cloud Services Authors: Sharad Agarwal, John Dunagen, Navendu Jain, Stefan Saroiu, Alec Wolman, Harbinder Bogan 7th USENIX Symposium on Networked Systems
More informationHéctor Fernández and G. Pierre Vrije Universiteit Amsterdam
Héctor Fernández and G. Pierre Vrije Universiteit Amsterdam Cloud Computing Day, November 20th 2012 contrail is co-funded by the EC 7th Framework Programme under Grant Agreement nr. 257438 1 Typical Cloud
More informationRELIABILITY IN CLOUD COMPUTING SYSTEMS: SESSION 1
RELIABILITY IN CLOUD COMPUTING SYSTEMS: SESSION 1 Dr. Bahman Javadi School of Computing, Engineering and Mathematics Western Sydney University, Australia 1 TUTORIAL AGENDA Session 1: Reliability in Cloud
More informationData Centers and Cloud Computing
Data Centers and Cloud Computing CS677 Guest Lecture Tim Wood 1 Data Centers Large server and storage farms 1000s of servers Many TBs or PBs of data Used by Enterprises for server applications Internet
More informationAnálise e Modelagem de Desempenho de Sistemas de Computação: Component Level Performance Models of Computer Systems
Análise e Modelagem de Desempenho de Sistemas de Computação: Component Level Performance Models of Computer Systems Virgilio ili A. F. Almeida 1 o Semestre de 2009 Introdução: Semana 5 Computer Science
More informationIdentifying Workloads for the Cloud
Identifying Workloads for the Cloud 1 This brief is based on a webinar in RightScale s I m in the Cloud Now What? series. Browse our entire library for webinars on cloud computing management. Meet our
More informationLecture 15: Datacenter TCP"
Lecture 15: Datacenter TCP" CSE 222A: Computer Communication Networks Alex C. Snoeren Thanks: Mohammad Alizadeh Lecture 15 Overview" Datacenter workload discussion DC-TCP Overview 2 Datacenter Review"
More informationMeasurement of packet networks, e.g. the internet
Measurement of packet networks, e.g. the internet John Schormans (EE) Ben Parker (SMS) (next speaker in this joint talk) and Steven Gilmour (SMS Head of the Statistics Research Group and Director for the
More informationR-Capriccio: A Capacity Planning and Anomaly Detection Tool for Enterprise Services with Live Workloads
R-Capriccio: A Capacity Planning and Anomaly Detection Tool for Enterprise Services with Live Workloads Qi Zhang, Lucy Cherkasova, Guy Matthews, Wayne Greene, Evgenia Smirni Enterprise Systems and Software
More informationBottlenecks and Their Performance Implications in E-commerce Systems
Bottlenecks and Their Performance Implications in E-commerce Systems Qi Zhang, Alma Riska 2, Erik Riedel 2, and Evgenia Smirni College of William and Mary, Williamsburg, VA 2387 {qizhang,esmirni}@cs.wm.edu
More informationHybrid Auto-scaling of Multi-tier Web Applications: A Case of Using Amazon Public Cloud
Hybrid Auto-scaling of Multi-tier Web Applications: A Case of Using Amazon Public Cloud Abid Nisar, Waheed Iqbal, Fawaz S. Bokhari, and Faisal Bukhari Punjab University College of Information and Technology,Lahore
More informationVMware vcloud Architecture Toolkit Cloud Bursting
VMware vcloud Architecture Toolkit VMware vcloud Architecture Toolkit Version 3.0 September 2012 Version 2.0.1 This product is protected by U.S. and international copyright and intellectual property laws.
More informationDistributed Autonomous Virtual Resource Management in Datacenters Using Finite- Markov Decision Process
Distributed Autonomous Virtual Resource Management in Datacenters Using Finite- Markov Decision Process Liuhua Chen, Haiying Shen and Karan Sapra Department of Electrical and Computer Engineering Clemson
More informationDATABASES AND THE CLOUD. Gustavo Alonso Systems Group / ECC Dept. of Computer Science ETH Zürich, Switzerland
DATABASES AND THE CLOUD Gustavo Alonso Systems Group / ECC Dept. of Computer Science ETH Zürich, Switzerland AVALOQ Conference Zürich June 2011 Systems Group www.systems.ethz.ch Enterprise Computing Center
More informationAPPLICATION NOTE. XCellAir s Wi-Fi Radio Resource Optimization Solution. Features, Test Results & Methodology
APPLICATION NOTE XCellAir s Wi-Fi Radio Resource Optimization Solution Features, Test Results & Methodology Introduction Multi Service Operators (MSOs) and Internet service providers have been aggressively
More informationBest Practices for Validating the Performance of Data Center Infrastructure. Henry He Ixia
Best Practices for Validating the Performance of Data Center Infrastructure Henry He Ixia Game Changers Big data - the world is getting hungrier and hungrier for data 2.5B pieces of content 500+ TB ingested
More informationresearch How Manual Tasks Sabotage the Potential of Natural Search Marketers
research How Manual Tasks Sabotage the Potential of Natural Search Marketers Executive Summary Due to the technical nature of the SEO industry and its immaturity relative to other marketing disciplines,
More informationTrace-Based Evaluation of Job Runtime and Queue Wait Time Predictions in Grids
Trace-Based Evaluation of Job Runtime and Queue Wait Time Predictions in Grids Ozan Sonmez, Nezih Yigitbasi, Alexandru Iosup, Dick Epema Parallel and Distributed Systems Group (PDS) Department of Software
More information3.3. Traffic models and teletraffic dimensioning
3.3. Traffic models and teletraffic dimensioning Sándor Molnár, author Béla Frajka: reviewer 3.3.1. Introduction The basic teletraffic principles, equations and an overview of the nature of network traffic
More informationTable of Contents HOL-SDC-1317
Table of Contents Lab Overview - Components... 2 Business Critical Applications - About this Lab... 3 Infrastructure Components - VMware vcenter... 5 Infrastructure Components - VMware ESXi hosts... 6
More informationvsan 6.6 Performance Improvements First Published On: Last Updated On:
vsan 6.6 Performance Improvements First Published On: 07-24-2017 Last Updated On: 07-28-2017 1 Table of Contents 1. Overview 1.1.Executive Summary 1.2.Introduction 2. vsan Testing Configuration and Conditions
More informationPredictive Elastic Database Systems. Rebecca Taft HPTS 2017
Predictive Elastic Database Systems Rebecca Taft becca@cockroachlabs.com HPTS 2017 1 Modern OLTP Applications Large Scale Cloud-Based Performance is Critical 2 Challenges to transaction performance: skew
More informationStochastic Processing Networks: What, Why and How? Ruth J. Williams University of California, San Diego
Stochastic Processing Networks: What, Why and How? Ruth J. Williams University of California, San Diego http://www.math.ucsd.edu/~williams 1 OUTLINE! What is a Stochastic Processing Network?! Applications!
More information