Virtualization : IBM Insights In Sizing Servers For .
Virtualization : IBM Insights inSizing Servers for VirtualizationBob ZuberWW Program Manager for Virtualization
Learning ObjectivesTitle: Virtualization - IBM Insights in Sizing Servers for VirtualizationAbstract: Come hear how IBM can help you build a better virtualized infrastructure.Learn how to use the new IBM Virtualization Sizing Guide, to help you understand themethod and tools that IBM has used to assist their customers in many consolidationengagements. In addition, learn how IBM System x and Bladecenter servers can bedeployed to provide the most efficient and reliable hardware platform for your VMwarevirtualization needs. Also see how IBM Director and Virtualization Manager provide acomplete systems management solution, via a single pane of glass interface, for yourvirtual and physical servers
AgendaSizingThe elements of sizingIBM’s insights into sizingIBM PortfolioPortfolio PositioningCompetitive AdvantageSystems Management – IBM DirectorWhat is itHow it helps virtualizationSummary
Long-term Focus On Virtualization Across Our SystemsSystem z9 System p5 pSeries zSeries System x BladeCenter iSeries IBM SystemStorage OpenPower While virtualization sounds complex, it’s really a simple idea. IBM Systemscan provide virtualization capabilities that are unique in the marketplace. IBM mainframe virtualization – 40 year history of world-class innovationIBM X-Architecture designed for virtualization, shared cross platformCoolBlue - Power and Cooling designs that lead the industryVirtualization Management software that simplifies your environmentVirtualization features do not require “rip and replace” upgradesVMware’s number one OEM vendor(1)Note (1) : Based on bundled license revenue
Why is Sizing ImportantThere are so many sizing considerations, either the source or targetConsider just the Source serverAverage and Peak CPU utilizationAverage and Peakmemory utilizationPage Size and pageextentsDisk I/O throughputNetwork I/O throughputAnd now the Target serverCPU UtilizationMemory UtilizationPage size page extentsDisk I/O throughput
Rules for VirtualizationHot spot avoidance – Reduce risk of multiple workloads saturatinggiven physical serverIf the virtual servers are spread over multiple small servers (e.g., four 2-ways)rather than hosted by a large server (e.g, a single 8-way x460) then surges inworkload demand may saturate a given physical server while other physicalservers are underutilized. A larger shared server can avoid artificial hot spotswhich impact workload QoS (throughput and response time).Headroom for hard to predict workloadsIt is often the case in planning a new multi-system configuration that the resourceneeds of some systems/workloads are not known with confidence. In such cases,there is a big advantage in using a scalable server such as an x460 to hostmultiple virtual servers so that surprises in individual system/workload demand canbe handled without acquiring additional server hardware or making physicalchanges.Peak load ResponsivenessWith a large server rather than N smaller servers, there is a greater probability thata processor will be free at any point in time to handle spikes in any workloadsdemand.Source: J. RymarczykIBM Fellow, Chief Virtualization Technologist
Gather The DataVMware CapacityPlannerIBM CDATBoth gather inventoryand performanceinformation for serversand clients.Data is needed toprovide the best sizingfor your customersworkloads.Using Virtualization forServer Consolidation isonly as good as theinitial sizinginformation.
Real-World Workload EnvironmentsNon-Steady State(Unpredictable)Steady State(Predictable)Non-Steady vs Steady State workloads1.Peak utilization should be examined for all Non-Steady workloads. Email server peak periods may beentirely different from the CRM server, causing peaks and valleys to either negate each other, or doublingthe amount of resources needed.2.Average utilization can be examined for Steady state workloads, once they have achieved steady state.Measurements can be performed prior to Steady State, but this would not reflect the workloads trueresource needs.Source : Intel Technology Journal, Vol 10 Issue 3, published August 10, 2006 - “Redefining ServerPerformance Characterization for Virtualization Benchmarking”
What is VMware Capacity Planner?Agent-less discovery,inventory,and performance collectionData Sources: WMI, Registry &Perfmon API calls on Windowsand Remote SSH Sessionsusing UNIX and Linux utilitiesHourly performancemetrics for days/weeksCDAT compatibleWeb-based reportingIndustry benchmarking /comparisonConsolidation andvirtualizationscenario modelingIncorporated into BCEand VirtualizationAssessment reportsSelect IBMServer ModelCreate Server GroupsModel for VMware
What is CDAT? – Consolidation, Discovery & AnalysisDiscovers servers on the networkEstablishes communication with the servers on the network:SNMP for all serversexcept xSeriesStandard API calls forxSeriesTakes a snap shot ofthe data that iscollected and storedon the server by theOS and theperformance countersExports the data intoExcel spread sheetwhich can beexported to anyspread sheet of yourchoiceServergroupedintocategoriesby CDATfor datacollectionpurposesThe contents ofthe SQL servergroup – selecteda group, server,serverinformation
Customer Consolidation StudyMethodology for the Virtualization Sizing GuideUsed customer survey data to define key workload classifications targeted forserver consolidationAnalyzed recent CDAT(1) studies for 30 customers Organized their legacy servers by application type into the key workloadclassifications Entered more than 3,200 servers in the customer study database Defined typical average and peak utilization statistics for each key workloadclassificationInput typical workloads from each workload classification into IBM’s VISIAN(2) tool VISIAN consolidated the workloads into virtual machines onto a target serveraccording to the Headroom Rules. VISIAN defined the limiting factor for adding additional workloads (CPU,Memory, Virtual CPU)Notes :1. IBM’s CDAT is a stand-alone discovery tool that automatically gathers a significant amount of server inventory, configuration and performance data, generatingan enterprise-wide view of server population and usage. IBM has performed over 2,000 customer server consolidation studies involving over 200,000 servers.2. IBM’s VISIAN is a tool that facilitates the calculation of the optimal number of machines that can be consolidated onto an IBM System x server running VMwareESX Server 2.5. VISIAN is an internal-use-only tool, and an patent is pending.
Virtualization Sizing Customer consolidation studies highlight a greater variation between average CPU utilization and peak CPU utilization on legacy 2-Psource workloads than on 4-P and 8-P source workloads.Variations in the CPU utilization will determine the number of virtual machines possible on a server configuration.ServerPeakHeadroomand AverageUtilizationSize for excess capacity to ensure that workloads have headroom for peak periods of operation when they require greater serverresources than their average or steady state period of operation.To calculate the correct amount of headroom, consider the variation between average and peak server resource requirements foreach consolidated workload.The Headroom Rules account for this variation.Scheduler ContentionPeak Utilization versus Average UtilizationTo correctly size the number of virtual machines per platform, consider the average and peak utilization of the processor, memory,and disk.The chart on the next page shows the recommended number of VMs based on both the peak and average utilizations. The firstlettered indicator aligns with the number of recommended VMs based on the peak utilization values. The second lettered indicatoraligns with the number of recommended VMs based on the average utilization values.The letter represents the constraining factor which is the resource on the server that limits the recommended number of VMs for thatserver configuration.Memory versus CPU constraintsPhysical CPU utilization is not always the primary constraint for consolidated workloads.Workloads that are consolidated based on peak utilizations tend to be CPU-constrained and those based on average utilizationstend to be memory-constrained or vCPU-constrained.VMware Scheduler ContentionThe amount of work that can be scheduled has a non-linear relationship to the number of CPUs and the number of vCPUS definedfor a given virtual machine.VMs with more than 1 vCPU have lower scheduler latency on SMP servers with more CPU cores compared to servers with lessCPU cores.
Server Headroom CalculationData collected from 3,000 servers viaCDAT consolidation studiesWindows and Linux non-virtualized environmentsAverage seven different application workloadsLarger servers with more resources are lessvolatile2005 & 2006 studies adds another 12K serversto our dataVISIAN consolidated the workloads into virtualmachines onto a target server.VISIAN defined the limiting factor for addingadditional workloads (CPU, Memory, VirtualCPU)VirtualizedLegacy 2-PWorkloadsSource WorkloadsServerAvg CPUUtilizationPeak %5%35%90%55%8-P68%3%25%90%65%Virtualization increases server utilization, but proper configuration mustallow for application usage spikes2-P Add 37% headroom to 7% average to achieve 44% avg and 90% peak virtual capacity4-P Add 55% headroom to 5% average to achieve 60% avg and 90% peak virtual capacity8-P Add 65% to 3% average to achieve 68% avg and 90% peak virtual capacityRule of thumb for virtualized System x Servers:2P 44% 4P 60% 8P 68%NOTE : The information in the following table represents the conclusions of IBM from testing of systems in a controlledenvironment. Actual environmental costs and performance characteristics will vary depending on individual customerconfigurations and conditions. IBM makes no representation or warranty that an individual user will achieve resultsequivalent to the levels stated in this document.
Virtualization Sizing InsightsServer HeadroomPeak and Average UtilizationScheduler Contention
x86 Server Utilization ObservationsThe CDAT information wasbroken into two types ofworkloads, 2-P and 4-P.The sizing was based onboth the average and peakfor CPU, memory, disk, andnetwork I/O.The disk and network I/Owas not included in theprint document, it was nota constraint that was seenin the sizing table.This information was used tosize the target servers.Each cell of the sizingguide is represented bythese 6 workloads, andadded in a round robinfashion to achieve theneeded workloads.Consolidation Parameters for Source WorkloadsLegacy rminal ServerEmailLegacy rminal ServerEmailTypicalProcessorXeon 2.0GHzXeon 1.8GHzXeon 1.8GHzXeon 1.8GHzPIII 1.4GHzXeon 2.0GHzTypicalProcessorXeon MP 2.5GHzXeon MP 2.5GHzXeon MP 2.7GHzXeon MP 2.5GHzXeon MP 2.7GHzXeon MP 2.8GHzAvg CPUUtilizationPeak CPUUtilization8%5%8%9%9%6%Avg CPUUtilization48%47%52%60%70%50%Peak CPUUtilization6%4%4%5%6%4%35%24%34%37%45%34%Table C - Consolidation Parameters for Source WorkloadsNOTE : The information in the following table represents the conclusions of IBM from testing of systems in a controlledenvironment. Actual environmental costs and performance characteristics will vary depending on individual customerconfigurations and conditions. IBM makes no representation or warranty that an individual user will achieve resultsequivalent to the levels stated in this ,5361,536
Virtualization Sizing GuidanceLegacy 2-P Workloads virtualized to a VM defined as a Single processor (1vCPU)Single Core - # of VMs1x346/2DP 3.8GHz 8 GBx346/2DP 3.8GHz 16 GBx366/4MP 3.66GHz 16 GBx366/4MP 3.66GHz 32 GBx366/4MP 3.66GHz 40 GBx460/8MP 3.33GHz 32 GBx460/8MP 3.33GHz 64 GBDual Core - # of VMs234567CC89 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70MMMCCCMVCC12x3650/2DP 3.0GHz 8 GBx3650/2DP 3.0GHz 16 GBx3850/4MP 3.0GHz 16 GBx3850/4MP 3.0GHz 32 GBx3850/4MP 3.0Ghz 48 GBx3950/8MP 3.0GHz 32 GBx3950/8MP 3.0GHz 64 GBNumber of Recommended VMs34567CC8MLegendVMs running at Peak UtilizationVMs running at Avg UtilizationC CPU ConstraintM Memory ConstraintV vCPU ConstraintM9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70MMMCCCBest Estimate at this timeMMCCMMLegacy 4-P Workloads virtualized to a VM defined as a Dual processor (2vCPU)Single Core - # of VMs12x346/2DP 3.8GHz 8 GB Cx346/2DP 3.8GHz 16 GB Cx366/4MP 3.66GHz 16 GBx366/4MP 3.66GHz 32 GBx366/4MP 3.66GHz 40 GBx460/8MP 3.33GHz 32 GBx460/8MP 3.33GHz 64 GBDual Core - # of VMsx3650/2DP 3.0GHz 8 GBx3650/2DP 3.0GHz 16 GBx3850/4MP 3.0GHz 16 GBx3850/4MP 3.0GHz 32 GBx3850/4MP 3.0Ghz 48 GBx3950/8MP 3.0GHz 32 GBx3950/8MP 3.0GHz 64 GB1345678Number of Recommended VMs9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70MVCCCMVVCC234567MV89 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70CCCMMCCCMVNOTEinformationSizingin theGuidancefollowing table represents the conclusions of IBM from testing of systems in a controlledTable :AThe- Virtualizationenvironment. Actual environmental costs and performance characteristics will vary depending on individual customerconfigurations and conditions. IBM makes no representation or warranty that an individual user will achieve resultsequivalent to the levels stated in this document.
Selection CriteriaLegacy 2-P Workloads virtualized to a VM defined as a Single processor (1vCPU)Dual Core - # of VMs12x3650/2DP 3.0GHz 8 GBx3650/2DP 3.0GHz 16 GBx3850/4MP 3.0GHz 16 GBx3850/4MP 3.0GHz 32 GBx3850/4MP 3.0Ghz 48 GBx3950/8MP 3.0GHz 32 GBx3950/8MP 3.0GHz 64 GB3CC456789 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50MMMCCCBest Estimate at this timeMMCCM2-P4-P8-PZoneZone 11 (VMs/Memory)(VMs/Memory)3 / 8GB9 / 16GB16 / 24GBZoneZone 226 / 8GB18 / 24GB32 / 40GBZoneZone 339 / 16GB27 / 32GB48 / 52GBZone 1 (Conservative)Zone 3 (Aggressive)Aggressive application (ie. Database)Lightly aggressive application (ie. Web)Unpredictable workloadPredictable WorkloadNo tolerance for performance degradationHigh tolerance for performance degradationZone 2 (Moderate)Moderately aggressive application (ie. Collaboration)Predictable workloadSome tolerance for performance degradationNOTE : The information in the following table represents the conclusions of IBM from testing of systems in a controlledenvironment. Actual environmental costs and performance characteristics will vary depending on individual customerconfigurations and conditions. IBM makes no representation or warranty that an individual user will achieve resultsequivalent to the levels stated in this document.
Virtualization Sizing InsightsServer HeadroomPeak and Average UtilizationScheduler Contention
Guidance from VMware The Benefits of CPU Dense ESX Server Hosts:“The chance that the scheduler can find room for a particularworkload without much reshuffling of virtual machines will always bebetter when the scheduler has more CPUs across which it can searchfor idle time. For this reason, it will generally be better to purchase twofour-way ESX Server licenses than to purchase four two-waymachines. Similarly, two eight-way servers will provide morescheduling flexibility than four four-way servers.”Tips and Techniques for Implementing InfrastructureServices on ESX Server
VMware SchedulerCPU 1CPU 2Single Core 2-PFirst Clock Cycle1 vcpu, 100%2 vcpu, 200%2 vcpu, 150%Third Clock Cycle1 vcpu, 50%1 vcpu, 100%Fourth Clock Cycle4 vcpu, 400%LatencySecond Clock Cycle
VMware SchedulerDual Core 2-P&Single Core 4-PCPU 1CPU 2CPU 3CPU 41 vcpu, 100%2 vcpu, 200%2 vcpu, 150%1 vcpu, 50%1 vcpu, 100%Second Clock Cycle4 vcpu, 400%Third Clock CycleLatencyFirst Clock Cycle
VMware SchedulerCPU 1CPU 2CPU 3CPU 4CPU 5CPU 6CPU 7CPU 8Dual Core 4-P1 vcpu, 100%2 vcpu, 150%1 vcpu, 50%1 vcpu, 100%First Clock Cycle4 vcpu, 400%Second Clock CycleLatency2 vcpu, 200%
AgendaSizingThe elements of sizingIBM’s insights into sizingIBM PortfolioPortfolio PositioningCompetitive AdvantageSystems Management – IBM DirectorWhat is itHow it helps virtualizationSummary
x86 Servers Made Better With ion servingServer consolidationIBM BladeCenter IBM System xTMTwo-processor rack & towerSystemxTM Four IBM IBMSystemx Fourprocessorprocessorrack rackDescriptionIntegrating server, networkingand storage resourcesCombining a few applicationson a single server for greaterutilizationConsolidating large numbers ofunderutilized servers forgreatest TCOKey attributesHardware usabilityPackaging densityUnified managementPower/cooling savingsServer mobilityHigh availability/disasterrecovery capabilityIndustry-standard designPrice/performanceCompatibilityMultiple apps per serverHigh alabilityStrong reliability featuresApplication isolationReduction of physical systemsRapid system deploymentnotiRecudnoti
IBM differentiates itself from the competitionedlaCerentx m oret smet ss ces eryS pro towtwBo-s ceyS pro& Compatibility across chassis Innovative design withstandard parts Comprehensive ecosystem More memory DIMMS per Power managementprocessor than competitors Two redundant high-speed More I/O slots perfabricsprocessor than competitorskcarrosxkcra rufo IBM X3 chipset Up to 32-socket scalability Mainframe-inspired reliability PFA on more componentsthan competitorsCommon across the portfolio IBM CoolBlue innovations IBM Director Largest IBM Virtualization ManagerOEM vendor(1)Note (1) : Based on bundled license revenue
IBM and VMware : A Relationship Of Firsts!First VMware system vendorFirst joint development partnerFirst to leverage VMware SDKFirst to offer comprehensive supportMicrosoft Windows, Linux and IBM softwarein a VMware virtual machineFirst blade offering to include VMware First to integrate VMware into a Virtual ClientSolutionFirst system vendor to announce VMware Infrastructure 3
BladeCenter Competitive AdvantagedalBeCerent Up to 30% better power efficiency Fully redundant configurations for higher availability Largest ecosystem provides more flexibility and choice Innovative ‘snap-in’ scalable blade server that scales from 2P to 4P in seconds Open architecture allows 3rd party hw and sw vendors to provide more BladeCenter solutions Blade compatibility across chassis provides investment protectionCommon across the portfolio IBM CoolBlue innovations IBM Director Largest IBM Virtualization ManagerOEM vendor(1)Note (1) : Based on bundled license revenue
When do I deploy IBM BladeCenter Systems?When you want to:Consolidate infrastructure and centralize managementReduce the complexity of ‘scale out’Reduce power consumptionMake more efficient use of datacenter spaceReduce floor space consumedReduce weightAchieve maximum processor densityIt’s that Simple . With IBM
I/O Virtualization SolutionBuilding BlocksNetwork-embedded InfiniBandSubnet ManagerCisco VFrame ServerVirtualization SoftwareHost channel adapter (HCA)With upper layer protocolsCisco InfiniBandSwitch Module forIBM BladeCenter MPIIPoIBSDPSRPLinux, Windows, and Solarisdriver supportEmbedded system andfabric managementCisco SFS 3012 InfiniBand Switching InfiniBand to Ethernet InfiniBand to Fibre Channel
Infrastructure Simplification BenefitBefore Virtualization .14 x 2P BladesBladecenter chassis 21.9KHS20 Blade 3.8GHz w/4GBBlade cost 5.6K x 14 78.4KWindows SE cost 800 x 14 11.2KAverage CPU Utilization 8%Space Consumption 7UPower Consumption 2000W x 2 4000WTotal Cost 111.5KCost/Application 7,964Virtualization could provide upto a 2.25x cost improvement.After Virtualization .14 x 2P Blade56 ApplicationsBladecenter chassis 21.9KHS20 Blade 3.8Ghz w/4GBBlade cost 5.6K x 14 78.4KWindows SE cost 800 x 56 44.8KVMware cost 3,750 x 14 52.5KSupports 4 avg and 2 peak utilization VMs(1)Recommend 4 VMs/system (Zone 2)84 ApplicationsSpace Consumption 7UPower Consumption 2000W x 2 4000WTotal Cost 197.6KCost/Application 3,528Note (1) : Configuration performed after publication of Sizing Guide, 4 avg vs 2 peak for 4GB configuration.Web list prices found on www.ibm.com as of 8/1/06
BladeCenter Competitive AdvantageHP BL465cLS21LS41 scalableblade2248 DIMMs8 DIMMs16 DIMMs16GB32GB64MBHDDs21(4)*2(5)*I/O expansion slots223In futurenownowDual core AMD OpteronProcessorsDIMMs/Max RAMHigh speed (10Gb) supportNote : Specifications for the server hardware can be found on the following web sites : www.ibm.com,www.dell.com, and www.hp.com* With optional SIO expansion blade
IBM differentiates itself from the competitionxm oret skcras ces eryS pro towtwo-& Innovative design withstandard parts More memory DIMMS perprocessor than competitors More I/O slots perprocessor than competitorsCommon across the portfolio IBM CoolBlue innovations IBM Director Largest IBM Virtualization ManagerOEM vendor(1)Note (1) : Based on bundled license revenue
When do I deploy IBM Two Socket Systems?When you want to:Optimize highly demanding business applicationsRun applications that require high-speed memory subsystemsNeed broad and low-latency I/O choices for Ethernet and fibrechannelActively manage energy consumptionConsolidate legacy 2 Socket servers using virtualization When consolidating several traditional 2 way workloads When consolidating less than 18 traditional 2 way workloadsIt’s that Simple . With IBM
Application Serving BenefitBefore Virtualization .6 x 2P Serverx3650 3.0GHz DC w/4GBServer cost 4.9K x 6 29.4KWindows SE cost 800 x 6 4.8KAverage CPU Utilization 8%Space Consumption 12UPower Consumption 242Wx6 1452WTotal Cost 34.2KCost/Application 5,700After Virtualization .2P Serverx3650 3.0GHz DC w/8GBServer cost 9.5KWindows SE cost 800 x 6 4.8KVMware cost 3,750 x 2 7.5KSupports 8 avg and 3 peak utilization VMsRecommend 6 VMs/system (Zone 2)Space Consumption 2UPower Consumption 242WTotal Cost 21.8KCost/Application 3,633Virtualization could provide up to a 1.5x cost improvement.Web list prices found on www.ibm.com as of 8/1/06
How about 12 legacy 2 way workloadsAfter Virtualization .Before Virtualization .12 Servers2P ServersServing from .4P Serverx3650 3.0GHz DC w/8GBServer cost 9.5K x 2 19KWindows DC cost 2K x 2 x 2 8K*VMware cost 3,750 x 2 x 2 15KSupports 8 avg and 3 peak utilization VMsRecommend 6 VMs/System (Zone 2)Space Consumption 18UPower Consumption 246Wx2 492WTotal Cost 42KCost/VM 3,500x3850 3.0GHz DC w/24GBServer cost 24.4K 24.4kWindows DC cost 2K x 4 8K*VMware cost 3,750 x 4 15KSupports 24 avg and 9 peak utilization VMsRecommend 18 VMs/system (Zone 2)Space Consumption 3UPower Consumption 799WTotal Cost 47.4KCost/VM 3,950Cost per VM could be up to 22% less on the 2P PlatformNote : MS Windows Datacenter licensing – Estimated 2,000 processor, unlimited virtualization client. GA 9/28/06.Web list prices found on www.ibm.com as of 8/1/06
IBM 2 Socket Competitive AdvantageDellHP DL385 G2IBM x3655Dual core AMD OpteronProcessors?22DIMMs/Max RAM?8 DIMMs16 DIMMs32GB64MBHDDs?2(8)*2(8)*I/O expansion slots?44High speed (10Gb) support?nownowDell has NO mainstream 2U twosocket AMD offering!!Note : Specifications for the server hardware can be found on the following web sites : www.ibm.com,www.dell.com, and www.hp.com
IBM differentiates itself from the competitionkcarrosxmet ss ceyS pro rufo IBM X3 chipset Up to 32-socket scalability Mainframe-inspired reliability PFA on more componentsthan competitorsCommon across the portfolio IBM CoolBlue innovations IBM Director Largest IBM Virtualization ManagerOEM vendor(1)Note (1) : Based on bundled license revenue
When do I deploy IBM Scaleable Systems?When you want to:Run Applications that Multi-Thread SAP, SAS, DB2, Oracle, SQL, CognosRun Applications that have large local storage requirements Database, Large Collaboration and or MessagingConsolidate legacy servers using virtualization When consolidating more than 18 traditional 2 way workloads When consolidating more than 10 traditional 4 way workloads When partitions needs 2 or more virtual CPU’sIt’s that Simple . With IBM
Datacenter Consolidation BenefitBefore Virtualization .After Virtualization .18 Servers4P Serverx3650 3.0GHz DC w/4GBServer cost 4.9K x 18 88.2KWindows SE cost 800 x 18 14.4KAverage Utilization 5%Peak Utilization 47%Space Consumption 36U 1 RackPower Consumption 242W x 18 4356WTotal Cost 102.6KCost/Application 5,700x3850 3.0GHz DC w/24GBServer cost 24.4KWindows DC cost 2K x 4 8K*VMware cost 3,750 x 4 15KSupports 24 avg and 9 peak utilizationVMsRecommend 18 VMs/system (Zone 2)Space Consumption 3UPower Consumption 835WTotal Cost 47.4KCost/Application 2,633Virtualization could provide up to a 2.2x cost improvementNote : MS Windows Datacenter licensing – Estimated 2,000 processor, unlimited virtualization client. GA 9/28/06.Web list prices found on www.ibm.com as of 8/1/06
How about 36 legacy 2 way workloadsBefore Virtualization .36 ServersAfter Virtualization .Consolidated to .2P Servers4P Serverx3650 3.0GHz DC w/8GBServer cost 9.5K x 6 57KWindows DC cost 2K x 2 x 6 24K*VMware cost 3,750 x 2 x 6 45KSupports 8 avg and 3 peak utilization VMsRecommend 6 VMs/System (Zone 2)Space Consumption 18UPower Consumption 246Wx6 1476WTotal Cost 126KCost/VM 3,500x3850 3.0GHz DC w/24GBServer cost 24.4K x 2 48.8kWindows DC cost 2K x 4 x 2 16K*VMware cost 3,750 x 4 x 2 30KSupports 24 avg and 9 peak utilization VMsRecommend 18 VMs/system (Zone 2)Space Consumption 6UPower Consumption 799Wx2 1598WTotal Cost 94.8KCost/VM 2,633Cost per VM could be up to 25% less on the 4P PlatformPlus Lower Software & Management CostsNote : MS Windows Datacenter licensing – Estimated 2,000 processor, unlimited virtualization client. GA 9/28/06.Web list prices found on www.ibm.com as of 8/1/06
X Architecture Competitive AdvantageDell 6950HP DL585IBM x375544416 DIMMs32 DIMMs32 DIMMs64GB64GB128MB4 X 3.5” SAS8 x 2.5” SAS5 x 3.5” SASDual core AMD OpteronProcessorsDIMMs/Max RAMHDDsI/O expansion slotsHTX SupportPCI-X 0PCI-X 2(100mhz)PCI-X 2(66mhz)PC-E 2(x8), 5(x4)PCI-E 3(x8), 4(x4)PC-E 1(x16), 3(x8)NoYes – 1 slotNoNote : Specifications for the server hardware can be found on the following web sites : www.ibm.com,www.dell.com, and www.hp.com
Application Serving versus Server Consolidation2P Server4P ServerWorkgroup ConsolidationDrive server utilizationMultiple applications per serverDatacenter ConsolidationSmallest number of serversApplication IsolationPower ReductionManagement FeaturesPower ReductionManagement featuresPlatform selection driven by size, and number ofvirtual machines
AgendaSizingThe elements of sizingIBM’s insights into sizingIBM PortfolioPortfolio PositioningCompetitive AdvantageSystems ManagementIBM Virtualization ManagerIBM DirectorSummary
IBM DirectorFrom Deployment Through Maintenance Deploy MaintainRemote DeploymentManager Monitor & AlertHardware StatusSW DistributionEvent Action PlansUpdate AssistantInventorySchedulerResource MonitorRemoteControlAnalyze &TroubleshootDynamicSystem eManager SystemAvailabilityOptimize
IBM Virtualization ndustry standards-based virtualization mgmt interfaces (DMTF)PHyP
IBM Virtualization ManagerKey Design PointsDiscover and visualize resources and relationshipsOut-of-the-box discovery, easily find resources and relationships, analyzeenvironment, etc.Show health and tasks from all resources and relationshipsDefine and monitor health, drill down on problems quickly to find root causeProvide common tasks that work across all resourcesProvide detailed tasks to perform platform-specific tasks in common waysGrow existing workload by expanding or migratingIncrease virtual server’s memory, storage, processing, networking capabilitiesGrowing virtual server workload can be moved to more powerful physical serverUse virtualization to help with repair and upgrade of serversAdd new resources and work into virtual environmentEasily allocate, configure, and manage virtual serversMake changes transparently without “change windows”
IBM Virtualization Manager and VMware VirtualCenterVMware VirtualCenter provides management of homogeneous VMwareenvironmentsDynamic (live) migration via VMotion Supports cloning and P2VSingle point of management for multiple virtual hostsFocuses on virtualization management, not physical systemsIBM Virtualization Manage
Title: Virtualization - IBM Insights in Sizing Servers for Virtualization Abstract: Come hear how IBM can help you build a better virtualized infrastructure. Learn how to use the new IBM Virtualization Sizing Guide, to help you understand the method and tools that IBM has used to
Modi ed IBM IBM Informix Client SDK 4.10 03/2019 Modi ed IBM KVM for IBM z Systems 1.1 03/2019 Modi ed IBM IBM Tivoli Application Dependency Discovery Manager 7.3 03/2019 New added IBM IBM Workspace Analyzer for Banking 6.0 03/2019 New added IBM IBM StoredIQ Suite 7.6 03/2019 New added IBM IBM Rational Performance Test Server 9.5 03/2019 New .
Chapter 1. Introduction to IBM Power Virtualization Center IBM Power Virtualization Center is designed to simplify the management of virtual resources in your Power Systems environment. This topic provides an overview of what you can do using IBM PowerVC. Using IBM Power Virtualization Center, you can do the following:
In this section, we give an overview of virtualization and describe virtio, the virtualization standard for I/O devices. In addition, we discuss the state-of-the-art for network I/O virtualization. 2.1 Overview of Virtualization and virtio The virtualization technology is generally classi ed into full-virtualization and paravirtualization.
IBM 360 IBM 370IBM 3033 IBM ES9000 Fujitsu VP2000 IBM 3090S NTT Fujitsu M-780 IBM 3090 CDC Cyber 205 IBM 4381 IBM 3081 Fujitsu M380 IBM RY5 IBM GP IBM RY6 Apache Pulsar Merced IBM RY7
Product Analysis for IBM Lotus Domino, IBM Lotus Notes, IBM Lotus iNotes, IBM Lotus Foundations, IBM Lotus Quickr, IBM Lotus Sametime, IBM Lotus Connections, and IBM LotusLive. This report is intended for Organizations, Vendors, and Investors who need to make informed decisions about the Email and Collaboration market. Figure 1: Worldwide IBM .
This guide also explains the advantages of virtualization and dispels some common myths that exist regarding virtualization. 1.1. Who should read this guide? This guide is designed for anyone wishing to understand the basics of virtualization, but may be of particular interest to: Those who are new to virtualization.
TU Dresden, 2009-12-01 MOS - Virtualization Slide 6 von 58 Virtualization – a hype A lot of interest in the research community within the last years, e.g.: SOSP 03: Xen and the Art of Virtualization EuroSys 07: a whole session on virtualization Many virtualization products: VMware, QEmu, VirtualBox, KVM
The modern approach is fact based and lays emphasis on the factual study of political phenomenon to arrive at scientific and definite conclusions. The modern approaches include sociological approach, economic approach, psychological approach, quantitative approach, simulation approach, system approach, behavioural approach, Marxian approach etc. 2 Wasby, L Stephen (1972), “Political Science .
