HPE SimpliVity Sizing and Configuration

1 HPE SimpliVity Sizing and Configuration Marcus BurrowsAtrium ConsultingMarch 2019 Version confidence in Sizing HPE SimpliVity solutionsBe able to identify customers' needs to better recommend and size HPE productsKnow how to use HPE Sizing toolsUnderstand how to size a SimpliVity solutionUnderstanding HPE SimpliVity Sizing and ConfigurationWhat are the technical considerations in a SimpliVity Sizing exercise? Do NOT size using HPE QuickSpecsor OAC Effective Capacity, listed is per node and without the cost of HAHPE SimpliVity 380 Gen10*XL Configuration is available for 4000-series (TB) vs Tebibyte (TiB) units with the customer one unit and stick with it TiB(GiB) when Sizing for HPE SimpliVity Sizer does the math for you! NOT size using HPE QuickSpecsor convert TB to TiB=N-TB*(1000^4) / (1024^4)] = HPE SimpliVity Sizer does the math for you!HPE SimpliVity Performance Sizing Capacity Overview (QRS)SSD Workloads Primary technical difference is Lifetime Average Write Performance4000 Series 4000 Series is best for Read-intensive or typical read/write mixed workloads 4000 Series is a good fit for read/write mixed workloads when properly sized 4000 Series is not ideal for heavy write-intensive workloads6000 Series 6000 Series is best for high performance mixed workloads 6000 Series can play in both read intensive and IO intensive environments 6000 Series is not ideal for cost sensitive customersSeries 4000 Series 6000 ApplicationsBasic Business & OfficeAppsPerformance Driven, Enterprise AppsWorkloadProfileModerateIO IntensiveMedia Lifecycle Manager for value flash models (HPE SimpliVity Gen 10 -4000 series)All Gen 10 Series 4000 models use read intensive SSD drives with lower write endurance than Series 6000 drives.

2 This provides a lower cost option to customers, with the performance reliability associated with an all flash system. The expectation is that system Sizing aligns with customers workloads which will not exceed the product life span (SSD endurance). Media Lifecycle Manager will only adjust write I/O and not read Lifecycle ManagerHPE SimpliVity 380 Gen10 data at rest encryptionHPE SimpliVity 380 Gen10 data at rest encryptionHPE SimpliVity 380 Gen10 data at rest encryptionSimpliVity 380: Network interfaces1. 10 GbE 1 and 2 (numbered from right to left) HPE OmniStack Storage, Federation, and, optionally, GbE 1 and 2 (numbered left to right) HPE OmniStack Management3 and 4 Guest VM data (optional) SimpliVity 380: 10 GbE-only network configurationSimpliVity 380: Direct-connected network configurationSimpliVity 380: Switch-connected network Configuration Data AssessmentData GatheringArchitecturalDesignDesignReview 1234 The Process for SizingIntroduction to the MethodologyCollect data from your customerApply guidelines and formulas2 Build the solution for the customer Perform a review of your designData GatheringVirtualization / Compute-VMs-vCPU-Peak GHz-vRAM-Storage-Capacity-Used vs Allocated SANs lie, VMDKs lie, Datastores lie, beware In-Guest iSCSI & RDMs-Backup Capacity-Performance IOPS, %R/W, IO Sizes-GrowthUnderstanding HPE SimpliVity Sizing and ConfigurationFurther considerationsThe Data Assessment ProcessPutting It All TogetherData AssessmentData Gathering12 NodeClusterFederationBackupsCPUM emoryStorage CapacityStorage PerformanceFailoverSize WorkloadsAdjust for GuidelinesAdjust for SparesGrowthHPE SimpliVity Configuration Guidelines* & ** (Page 1 of 5)

3 Based on running version softwareHypervisorHyper-VvSphereUse CaseData Center Consolidation 380 Gen10380 Gen10170 Gen 10190 Gen 10 ClusterSimpliVity Nodes2161616 Compute NodesNA2:12:12:1 FederationSimpliVity Nodes8969696 Compute NodesNA2:12:12:1 VMsper Node40180120120per Federation320576057605760* The values are guidelines and not limits, please submit a DSR for exceptions** The values above are not a function of storage capacity but the number of objectsHPE SimpliVity Configuration Guidelines* & ** (page 2 of 5)based on running version softwareHypervisorvSphereUse CaseVDIROBO380 Gen10170 Gen10190 Gen10380 Gen 10 170 Gen 10190 Gen 10 ClusterSimpliVity Nodes161616161616 Compute Nodes1:11:11:12:12:12:1 FederationSimpliVity Nodes969696969696 Compute Nodes1:11:11:12:12:12:1 VMsper Node250250250180120120per Federation400040004000576057605760* The values are guidelines and not limits, please submit a DSR for exceptions* The values above are not a function of storage capacity but the number of objects* 96 nodes if using Enhanced Linked Mode to link multiple vCenter Servers.

4 To reach the max of 96 you need three linked vCenter Servers. Max # of nodes per vCenter Server is 32. HPE SimpliVity Configuration Guidelines (Page 3 of 5)based on running version softwareHypervisorHyper-VvSphereUse CaseData Center Consolidation 380 Gen 10380 Gen 10170 Gen 10190 Gen 10 BackupsTotalperFederation100K250K250K250 KTotal concurrent local40100100100 Totalconcurrent inbound per node80100100100 Total concurrent outbound pernode40505050 BackupPolicies per Federation32128128128 Rules per policy16161616 Data stores per Federation32128128128 Data stores per Cluster16161616* The values are guidelines and not limits, please submit a DSR for exceptions* The values above are not a function of storage capacity but the number of objectsHPE SimpliVity Configuration Guidelines*&** (Page 4 of 5)based on running version softwareHypervisorvSphereUse CaseVDIROBO380 Gen 10170 Gen 10190 gen 10380 Gen 10170 Gen 10190 Gen 10 BackupsTotalperFederationN/A**N/A**N/A** 750K750K750 KTotal concurrent localN/A**N/A**N/A**100100100 Totalconcurrent inbound per nodeN/A**N/A**N/A**100100100 Total concurrent outbound pernodeN/A**N/A**N/A**505050 BackupPolicies per FederationN/A**N/A**N/A**128128128 Rules per policyN/A**N/A**N/A**161616 Data stores per Federation128128128128128128 Data stores per Cluster161616161616** The majority of VDI deployments use linked clones.

5 Therefore, backups may not be as important.* The values are guidelines and not limits, please submit a DSR for exceptions** The values above are not a function of storage capacity but the number of objectsHPE SimpliVity Configuration Guidelines*&** (Page 5 of 5)based on running version softwareHypervisorHyper-VvSphereUse CaseData Center Consolidation VDI380 Gen10380 Gen10380 Gen10170 Gen10190 Gen10 LatencyOptimal for storage network (stretched cluster)Not supported today<1msDSR OnlyDSR OnlyDSR OnlyMax for managementnetwork (WAN)300ms300ms300ms300ms300msClones concurrent44888* The values are guidelines and not limits, please submit a DSR for exceptions* The values above are not a function of storage capacity but the number of objectsCompute NodesExisting VM Hosts10 GbE1-1 compute node SimpliVity node ratio recommendedThe Competition for Resources -Best Practices for Architecture of Application Specific workloads User interactive apps should be isolated from other workloads to reduce potential latency.

6 Ask your customer if they have latency sensitive application when designing their systems virtual desktops require a separate cluster to isolate for performance. A DSR is required when designing a mixed virtual desktop and server environment EUC (End-User Computing) requires fast response (cached IO, DRAM, SSD) Examples: VDI, XenDesktop, Horizon view, Terminal Servers Citrix Desktop users expect fast response and are sensitive to delays Separate EUC (End-User Computing) into separate clusters Certain Apps can be less sensitive to latency Reporting Business intelligence File Services Exchange (assuming cache-mode enabled)DC1 EUC (End-User Computing)DC2 OLTP3rdParty BackupFor customers that need to continue to use a 3rdparty backup tool for archival purposes, propose the the the 3rdparty backup be tool for archival purposes and not for recovery Use HPE SimpliVity 's backup when 3rdparty archival backup, propose to move to a monthly or quarterly backup schedule and use HPE SimpliVity for more frequent daily and weekly backupsWhy use HPE SimpliVity backups Take advantage of dedupe Much faster than traditional 3rdparty backup Faster recovery of your applications Automated recovery with RapidDRHPE SimpliVity RapidDR requirements Works with HPE SimpliVity or higher All supported ESXi versions with OmniStack or higher Windows 8, 10, or Server 2012, 2016 system to run the HPE SimpliVity RapidDR executable requires the following (Please refer to RapidDR user guide for more details on prerequisites) : Dual CPU, Ghz (minimum) 1GB of free memory (minimum) 1GB of disk space for install Microsoft.

7 Net Framework 4 or above PowerShell version vSphere Power CLI release 28 Note: RapidDR client Systems running additional services such as vCenter may require more resources. Additionally,RapidDR client memory and CPU requirements depends on the number of VMs that are protected with RapidDR. 600 VMs recovery require 16 GB of free HPE and Channel Partner internal use onlyHPE SimpliVity RapidDR Configuration guidelinesConfiguration guidelines (recommended) :-Two vCenter Servers in (Enhanced) Linked Mode -One vCenter for primary site and one vCenter for recovery site-1000 VMs in the entire recovery environment-600 VMs per recovery plan-50 recovery groups per recovery planRemember: Guidelines are not limits. We can support customer requirements exceeding guidelines with a Deal Specific and Arbiter placement Combine Arbiter and vCenter onto single Windows server Install Arbiter and vCenter onto separate servers Can utilize vCenter Server Appliance vCenter running in managed SimpliVity Data Center Utilize SimpliVity native data protection for vCenter Server Minimize the requirements for non- SimpliVity hardwareWindows or vCSAvCenterCPU1 GHz, dual threadMemory1 GB for Arbiter applicationStorage16 GB spaceWindowsArbiterUsing HPE SimpliVity Sizing ToolsLanamark OneLanamark AccessData Collection and Analysis Journey Data Collection and Analysis Journey Data Collection and Analysis Journey Data Collection and Analysis Journey Data Collection and Analysis Journey Data Collection and Analysis Journey Analysis and ReportingAnalysis and ReportingAnalysis and ReportingUsing HPE SimpliVity Sizing ToolsHPE SimpliVity Online Sizing HPE SimpliVity Sizer does the math for you!

8 : Amount of data the customer needs to Sum of [ VMData + plus growth over 3-years + backup data(local and remote) ] 1: (10,000 GiB) to store, the value entered as input in the sizer! ; two-copies (Primary and Secondary) = VMData x 2 = 20,000 GiB Node1 10,000 = 6,060 GiB for Primary Replica Node2 10,000 = 6,060 GiB for Secondary Storage Efficiency = Default :1 ( Dedupe and Compression) = 20,000 = 12,121 GiB stored on-disk in Used Physical Capacity in cluster (12,121 GiB/1024) = Available Physical Capacity = (Total Free Space Used) = SimpliVity Performance Sizing Capacity -VMDataHPE Confidential | HPE and Partner Training Use OnlyHPE SimpliVity Online Sizing UtilityCreate a new Sizing Create new Sizing Name the new Sizing Create Sizing Create a new cluster Add Cluster Name the cluster Select the product family Add Cluster Populate the cluster with data Add Cluster Add environmental data Select compression, dedupe, and latency settings Add growth (optional) Select N+ 2 10% Growth per year for 3 Years and No BackupNote.

9 Not year over year growthHPE SimpliVity Performance Sizing Capacity OverviewVMData (GiB)StrategyGrowthLocal BackupRemote BackupLocalCluster(GiB)Remote Cluster(GiB)Total Stored VMData10,000 Growth-only10%/year for 3-yrs(1000 GiB x 3= 3000 GiB)0013,000N/A13,000(GiB)Recommended Solutions Select Recommendations to view options For all models, view nodes required to satisfy design Select your model View Configuration info Select More to view other options for that model Select Compare to view multiple configs together Click Select to view Configuration for OCAA dvanced Options Add a second cluster Select Advanced Mode to enable backup options for your clusters To define Backup Policies, select the button at the upper left cornerBackup Policies Name your backup policy Select frequency, count, and retention time Repeat for each rule Verify policy at the summary box and click Create Policy Repeat process for multiple policiesBackup Policies Enter your change rate Select the backup policy for both local and remote clusters Select remote target cluster Click Recommendations and select your optionsSet VM Groups Change RatesIf you cannot determine the expected change rate.

10 Consider the reference chart below to choose a representative values imply that SimpliVity backups are effectively excluded from the Sizing Recommendations Final Recommendations will take into consideration backup retention for local clusters only To size for full DR, you will need to integrate both requirements together. This is an upcoming featureIncrease confidence in Sizing HPE SimpliVity solutionsBe able to identify customers' needs to better recommend and size HPE productsKnow how to use HPE Sizing toolsUnderstand how to size a SimpliVity solutionThank youAtrium Consulting