where should
play

Where should Background Research contributions infrastructure be - PDF document

Sep 03, 2023 •151 likes •297 views

10/24/2014 Outline Introduction (VM) Virtual Machine Research goals (PM) Physical Machine Challenges Research questions Background Research contributions PHD Dissertation Defense Supporting Infrastructure Research

  1. 10/24/2014 Outline � Introduction (VM) Virtual Machine � Research goals (PM) Physical Machine � Challenges � Research questions � Background � Research contributions PHD Dissertation Defense � Supporting Infrastructure � Research Contributions � Performance Modeling for Component Composition Wes J. Lloyd � VM Placement to Reduce Resource Contention October 27, 2014 � Workload Cost Prediction Methodology � Conclusions Colorado State University, Fort Collins, Colorado USA October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 2 Outline Research Goals � Introduction � Research goals � Support application migration: � Challenges VM component composition, dynamic scaling, infrastructure alternatives � Research questions � Background � Maximize: application throughput � Research contributions Requests per second � Supporting Infrastructure � Minimize: hosting costs, server occupancy � Research Contributions Number of VMs, CPU cores, memory, disk space, hosting costs � Performance Modeling for Component Composition � VM Placement to Reduce Resource Contention � Minimize response time � Workload Cost Prediction Methodology Average service execution time (sec/min) � Conclusions October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 3 October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 4 Outline Research Challenges – WHERE � Introduction � Research goals � Challenges � Research questions Where should � Background � Research contributions infrastructure be � Supporting Infrastructure � Research Contributions provisioned? � Performance Modeling for Component Composition � VM Placement to Reduce Resource Contention � Workload Cost Prediction Methodology � Conclusions October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 5 October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 6 1

  2. 10/24/2014 Research Challenges – WHERE Research Challenges - WHAT Service Isolation Component Composition Size Quantity Vertical Scaling Horizontal Scaling Provisioning Server Consolidation VM types Qualitative Variation Resource descriptions VM m1.large What infrastructure VM VM VM VM c3.xlarge Multi-tenancy Overprovisioning Virtualization Virtualization m2.2xlarge VM VM VM VM Overhead Hypervisors m1.small should be provisioned? c1.xlarge VM VM VM VM m3.medium Resource Contention m2.4xlarge VM VM VM VM c1.medium c1.medium c1.medium c1.medium m1.xlarge Performance c3.large October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 7 October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 8 Research Challenges - WHAT Research Challenges - WHEN Size Quantity Vertical Scaling Horizontal Scaling When should VM types Qualitative Resource descriptions VM m1.large VM VM VM VM infrastructure be c3.xlarge Virtualization Virtualization m2.2xlarge VM VM VM VM Overhead Hypervisors m1.small provisioned? c1.xlarge VM VM VM VM m3.medium m2.4xlarge VM VM VM VM c1.medium c1.medium c1.medium c1.medium m1.xlarge c3.large Performance October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 10 October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 9 Outline Research Challenges - WHEN � Introduction � Research goals Hot Spot Detection VM Launch Latency � Challenges � Research questions Future Load Prediction Pre-provisioning � Background � Research contributions � Supporting Infrastructure � Research Contributions � Performance Modeling for Component Composition � VM Placement to Reduce Resource Contention � Workload Cost Prediction Methodology � Conclusions October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 11 11 October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 12 2

  3. 10/24/2014 Research Questions (1/3) Research Questions (2/3) DRQ-2: Performance modeling DRQ-4: VM placement implications What are the most important resource When dynamically scaling cloud infrastructure utilization variables and modeling techniques to address demand spikes how does VM for predicting service oriented application (SOA) placement impact SOA performance? performance? DRQ-5: Noisy neighbors DRQ-3: Component composition How can noisy neighbors , multi-tenant VMs How does resource utilization and SOA that cause resource contention be detected? performance vary relative to component What performance implications result when composition across VMs? ignoring them? October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 13 October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 14 Outline Research Questions (3/3) � Introduction DRQ-6: Infrastructure prediction � Research goals � Challenges How effectively can we predict required � Research questions infrastructure for SOA workload hosting by � Background harnessing resource utilization models and � Research contributions Linux time accounting principles? � Supporting Infrastructure � Research Contributions � Performance Modeling for Component Composition � VM Placement to Reduce Resource Contention � Workload Cost Prediction Methodology � Conclusions October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 15 October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 16 Virtual Machine (VM) Placement Virtual Machine (VM) Placement as “Bin Packing Problem” as “Bin Packing Problem” � Components items � virtual machines (VMs) bins � Components items � virtual machines (VMs) bins n k � Virtual machines (VMs) items � physical machines (PMs) bins 4 15 � Virtual machines (VMs) items � physical machines (PMs) bins 5 52 � Dimensions � Dimensions 6 203 Bell’s Number � # CPU cores, CPU clock speed, architecture � # CPU cores, CPU clock speed, architecture 7 877 � RAM, hard disk size, # cores � RAM, hard disk size, # cores 8 4,140 � Disk read/write throughput � Disk read/write throughput 9 21,147 � Network read/write throughput � Network read/write throughput � PM capacities vary dynamically NP-Hard � PM capacities vary dynamically NP-Hard n . . . � VM resource utilization varies � VM resource utilization varies � Component requirements vary � Component requirements vary October 27, 2014 Wes J. Lloyd PHD Dissertation Defense 17 October 27, 2014 Wes J. Lloyd PHD Dissertation Defense 18 3

  4. 10/24/2014 Outline Why Gaps Exist � Introduction � Research goals � Public clouds � Challenges � Research questions � Research is time/cost prohibitive � Background � Hardware abstraction: Users are not in control � Research contributions � Rapidly changing system implementations � Supporting Infrastructure � Private clouds: systems still evolving � Research Contributions � Performance models (large problem space) � Performance Modeling for Component Composition � VM Placement to Reduce Resource Contention � Virtualization misunderstood or overlooked � Workload Cost Prediction Methodology � Conclusions October 27, 2014 Wes J. Lloyd PHD Dissertation Defense Approaches & Gaps 19 October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 20 Outline Primary Research Contributions � Introduction � In the context of SOA migration to IaaS Clouds � Research goals � Challenges � Resource utilization modeling to predict � Research questions component composition performance � Background � Research contributions � VM placement improvement to reduce contention � Supporting Infrastructure � Private IaaS: LeastBusy VM placement � Research Contributions � Public/Private IaaS: Noisy-Neighbor Detection, Avoid heterogeneous VM type implementations � Performance Modeling for Component Composition � VM Placement to Reduce Resource Contention � Workload cost prediction methodology for � Workload Cost Prediction Methodology infrastructure alternatives to reduce hosting costs � Conclusions October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 21 October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 22 Scientific Modeling Workloads VM-Scaler � CSIP: USDA platform for model services � Service oriented application surrogates � RUSLE2 – Soil erosion model � WEPS – Wind Erosion Prediction System � SWAT-DEG: Stream channel degradation prediction Monte carlo workloads � Comprehensive Flow Analysis tools Load estimator, Load duration curve, Flow duration Curve, Baseflow, Flood analysis, Drought analysis future October 27, 2014 Wes J. Lloyd PhD Dissertation Defense Research Questions & Methodology 23 October 27, 2014 Wes J. Lloyd PhD Dissertation Defense 24 24 4

Recommend

More recommend