Hitachi Dynamic Tiering For Mainframe And Hitachi Tiered .

2y ago
23 Views
2 Downloads
4.08 MB
38 Pages
Last View : 1d ago
Last Download : 3m ago
Upload by : Dahlia Ryals
Transcription

Hitachi Dynamic Tiering formainframe and Hitachi TieredStorage Manager for MainframeRos SchulmanHDS (ros.schulman@hds.com)March 11th, 2014Session Number (14811)

Hitachi Dynamic Tiering (HDT)PoolData Heat IndexTier 0HighActivitySetLeastReferencedPagesTier 1NormalWorkingSetTier 2DynamicTieringVolumeQuietData Set

Improved Performance at Reduced Cost:Data Locality and Pareto DistributionsClassic Pareto DistributionsActual Volume Workload(Also Known as the 80/20 Rule)Pool25I/O %20151050Capacity 1Capacity nnnLocation on VolumeWhy does it work so well?§ § 3Skew – At any time, only a small address range is activePersistence – When an address range is accessed it tends toremain so for a while

Hitachi Dynamic TieringAutomated Optimized Tiered Storage Management§ Before: Tiered storage andprovisioning Labor intensive Data classification before tieringHDT PoolTIER 0Flash Complicated management ofmultiple storage tiers§ Now: Dynamic tiering andprovisioning Controller-based automation Single, self-managed, self-healing,efficient pool of data All the benefits of tiered storage All the benefits of dynamicprovisioning No need for data lifies operations and data management§ Reduces opex, capex, and TCO§ Data Moves in 38MB Pages§ Datasets can span multiple Tiers§ Frequent AccessesInfrequent ReferencesTIER 1SASTIER 2Nearline

Hitachi Dynamic TieringSupports Virtualized Storage With HDT, Hitachi VirtualStorage Platform (VSP)provides automated tieredstorage management andperformance acceleration formultiple tiers of heterogeneousexternal storage- As an SSD accelerator- As a storage virtualizationcontroller- Heterogeneous storagebusiness continuity anddisaster recovery solution- Leveraging the breadth ofofferings on VSPNAS MainframeVTL VirtualCASServersVirtualizedStorageVirtualization !Services!DynamicTieringVolume

Relationship between numberof tiers and Tiering policy Configurable DP Volume Tiering PoliciesTieringNumber of Tiers in PoolNotePolicy2 Tier3 TiersAllAll tiersAll tiersDefault value; data isstored in all tiers.Level 1Tier 1Tier 1Data is always storedin the highest-speedtierLevel 2All tiers*Tier 1 and Tier 2Level 3All tiers*Tier 2Level 4All tiers*Tier 2 and Tier 3Level 5Tier 2Tier 3Data is always storedin the lowest-speedtier* Data is stored in all tiers as in the case of All specified in the tiering policy

New Page Assignment policyCONTROL WHICH TIER NEW PAGES ARE ASSIGNED TONew page allocationtier levelDescriptionHighThe new page is assigned from the higher tier of tiers set in the tiering policyMiddleThe new page is assigned from the middle tier of tiers set in the tiering policyLowThe new page is assigned from the lower tier of tiers set in the tiering policy

HDT CyclesHOW TIERING LEARNS YOUR WORKLOAD Cycle time set at the HDP pool level Manual mode User can start and stop performance monitoring using any interval up to7 days Scripting can set complex schedules to custom fit to priority work periods Automatic mode Customer defines strategy; it is then executed automatically Continuous monitoring followed by relocation cycles Monitor interval from 30 minutes to 1, 2, 4, 8 or 24 hours (default)Monitor 1Calc1T4T3T2T1T0Relocate1Monitor 2Calc2Relocate2Monitor 3TIMECalc3Relocate3Monitor 4Calc4Relocate4

HDT Monitoring ModesCONTROL HOW TIERING LEARNS YOUR WORKLOAD Monitoring modes set atHDP/HDT pool levelPeriodvs.Con nuous120 Period mode Continuous mode The value used in thecalculation cycle is theweighted average of multipleprevious monitoring cyclesfor DP volume page Reduces page trashing May slow migration to uppertiers100DP- ‐VolumePageIOPH The value used in thecalculation cycle is theactual I/O load on DPvolume page from previousmonitoring cycle11090Tier 1807060504030Tier ModeCon6nuousMode1,0091,010

Hitachi Dynamic Tiering for MainframeHitachi DynamicTiering for Mainframeand DFSMS

§ 3390 volumes are "fixed" to a single tier§ To transition a volume's data to another tierrequires ACS work, then migration and recall3390-A§ Performance problems needintervention to migrate to"higher" storage group§ Host-based volumemovement has high overheadcost§ Stale datasets are treated thesame as active ones untilHSM migrationStorage Group “AS07”3390-AStorage Group “SAS10”Storage Group “SSD”LIMITS OF SMS Storage Groupsand ACS Routines3390-A

New! HDT Volume-provisionedstorage groupHDT VolumeMulti-TieredStorage GroupSSDSASSAS NL

DFSMSStorage Groups (SG) and HDT for Mainframe storage Now with HDT for Mainframe storage, allstorage tiers may be combined into a singleSG1SG2Subsystemstorage group HDT automatically moves data within astorage group to the correct tier basedon Workload performanceSG3SSD Tier 1SAS Tier 2NL SAS Tier 3HDT PoolStorage GroupSubsystemONE SGHDT Volumesare optimallymanageddynamically atthe page level

HDT Operational ImpactWorking with StorageBefore HDTWith HDTAdd physical capacityAdd 3390-X volumes intostorage groupsAdd capacity into poolBalance use over newcapacityManually use HSMmigration/recallNo actions are neededDirect specificapplications to specificstorage resourcesCode ACS routines,follow-up with HSMmigrations and recallsSet 3390-A to an HDTpolicy – use same ACSroutines but no HSMneededAddress performanceproblems by movingdatasets or volumesCode ACS routines anduse HSM migration/recallHDT relocation has likelyprevented the issue;otherwise use HDT policyMaintain SMS storagegroups and ACS routinesManual effort to keepupdated with rulesdescribing exceptionsFewer exceptions sinceHDT keeps tiers properlypopulatedDemote data to lower tiers HSM moves datasets toan ML “tier” that hasn’tbeen opened for a whileHDT automatically movespages that haven’t beenused

DFSMS Storage Groups and HDT for Mainframe StorageOnline SGBatch SGSubsystemFLASH Tier 0SAS Tier 1NL SAS Tier 2HDT PoolStorage GroupSubsystemSubsystemHSM Archive SG Simplifies integration HDT can be applied to selectedStorage Groups only DFSMS Horizontal StorageGroups and Dynamic Pagebased Tiering volumes can bealigned

DFSMSStorage Groups and HDT for mainframe storage With HDT for Mainframe storage policies, individual policies can bedefined for volumes mapped to different storage groups Policies are supported based on tier ranges, analysis/migrationperiods, initial tier page assignments and relocation prioritySSD Tier 1SAS Tier 2NL Tier 3StorageGroupOnline SGHDT CustomPolicy Defined forOnline DataAcross Top 2 TiersBatch SGArchive SGHDT PolicyRestricted to Tier 2ResidencyHDT PolicyRestricted to Tiers2 and 3ResidencyHDT Pool

HDT and DFSMS/HSM integrationExampleHDP (Pool 1)Primary SpaceLevel 0“Level 1”AdoredTier 1 - FlashTier 2 – SAS 10K or SAS 15K“All” with “RelocationPriority” Most Loved“All”Loved“Level 5”LikedHDP (Pool 2) HSM ML1 and Low-I/O Density DP VolumesUnlikedNear-Line SAS 7.5KHDP (Pool 3) HSM ML2 and Very Low-I/O Density DP VolumesUnlovedExternal Volumes

Hitachi Tiered Storage Manager for MainframeHitachi DynamicTiering for MainframeManagement

Options for Managing DynamicTiering for Mainframe Setup and maintenance operations are done via VSPService Processor (SVP) with Hitachi Storage Navigator Operational management can be done from eitherStorage Navigator or native from z/OS with Hitachi TieredStorage Manager for Mainframe (HTSM) HTSM features- Storage SLA management via HDT policies- Native mainframe operations- Group operations make it easy to manage target volumes- Linkage/integration with DFSMS- Customization via REXX scripting ISPF interface – April 2014NEW

Hitachi Tiered Storage Manager for Mainframe –Z/OS HDT managementHost-based software that provides:HTSMforMainframeDynamicTieringVolumeData Center Centralized and unified mainframemanagement of Hitachi DynamicTiering Automation Integration with DFSMS andstorage groups Online storage service level controls Increase application performance Improves problem avoidance Single, consistent interface Command based, script driven Auto-discovery eliminates errors Accelerates deployment Enables reporting and automaticnotifications

What is Tiered storage Manager forMainframe? Hitachi Tiered Storage Manager for Mainframe Manages operational aspects of HDT, including tiering policies Integrates with z/OS SMS (storage groups)IBM System z (z/OS )DFSMSSG ASG BSMSinforma6onSG h)Tier2(SAS)TierReloca6onPoliciesTier3(NL SAS)Hitachi Virtual Storage Platform – Hitachi Dyamic Tiering for Mainframe Pool

Why HTSM for mainframe?HTSM provides important benefits Native management from z/OS Control for data location from host’s point of view, not storage system’spoint of view – improves control and simplifies operations Ability to utilize HDT performance and relocation data with other data(such as SMF records) Reduce dependency on open server-based operations Control of storage service levels using HDT policies Linkage with z/OS SMS (storage group) speeds integration andreduces opex Simplify management in large-scale environments with groupoperations Flexible command-line interface (CLI) (TSO/E REXX) enables usersto get the most out of Dynamic Tiering

Linkage with z/OS SMS FeatureHTSM works with SMS construct (storage group). Each tiering policy group canhave 1 or more storage groups. This feature gives users the capability tomanage the HDT environment from a z/OS point of view. SMS integrationmakes it easy to add Dynamic Tiering to existing SMS operationsSystemzDFSMSTPG1SG ASG BSMSinfoTPG2SG CDP- ‐VolHDTPoolStorageSSDSASSATAHTSMforMF

Customization via REXX Scripting HTSM provides a CLI (TSO/E REXX) that enables users to manage theirenvironments. The CLI also has the ability to cooperate with OS services(such as TSO/E service) and other products. For improved manageability HTSM for MF has separated operationprocedures (REXX script) and configuration file-based objects (such asgroup erreloca6on*/:SystemzDFSMSHTSMforMFCCWCCWTPG1 DP- ‐VolHDTPoolStorageSSDSASTPG1- ‐:TPG1- ‐TPG2:- ‐:TPG2- ‐:TPG2SATA Defini ons

HTSM Mainframe Reporting TPG QUERY STATISTICS Sample Output

From SDSF using z/OS SYSREXX

Hitachi Dynamic Tiering and HTSM for MFwith DFSMShsm 2.1HDT Policy Level 2 by Storage Group:‘Hot’ data is moved to SSD, but ‘Cold’data is never allowed below EnterpriseClass storage.Primary Storage HierarchyHotWarmHDTSSD /Enterprise( )AllocateEnd-of-year TransitionCoolHDT Policy Level 4 by Storage Group:‘Hot’ data is not allowed higher thanEnterprise Class, and ‘Cold’ data isallowed to reside on NL.Migration Storage HierarchyColdHDTEnterprise/ Nearline( )RecallML2(VTS)Frigid32 Day Migration

DFSMShsm 2.1 AND HTSM FOR MF Datasets may be eligible for class transitions when certain criteriaare met specified by the management class Time since Creation Time since Last Use: Data set is eligible for a transition on or afterthis time Periodic: Data set is eligible for a transition on a specific date. Data sets could be allocated to a storage group using HTSM for MFthat uses Tier 1 and Tier 2 and then later transitioned to a storagegroup that uses less expensive class of storage for permanentretention After a certain period of time, can be migrated directly to ML2

Hitachi Dynamic Tiering for MainframeHitachi DynamicTiering forMainframeUse Case Slides

VSP – USPV Performance ComparisonStudy Objectives Establish baseline performance of VSP Determine performance level versus USPVenvironment Prove the capability of being able to use higherdensity drives Prove the feasibility of using HDT in mainframeenvironmentAre we maximizing the greatest number of IOs with the smallestamount of SSDs?

ENVIRONMENTAL COMPARISONUSPV – Total Cap 231TBSN(kg)(lb)(kW)BTU/hrkVASQ ,71379,89924.9324.3633.6433.6455VSP – Total Cap 682TB(kg)(lb)(kW)BTU/hrkVASQ )(lb)(kW) BTU/hrkVA SQ FT FramesVSP%Savings55.45% 55.40% 62.38% 62.46% 62.06% 47.04%50.00%With 3X Capacity

Frontend Response –RMF Type 74VSP 1Avg Response Time pre-migrationAvg Response Time post migrationClose to 2X Faster Response Time over USP VUp to 50% more IO

Backend IO Distribution vs PhysicalCapacityPhysical Capacity DistributionIO Rate Distribution18.37%51.02%30.61%

HDT and HTSM for MFSummary

VSP HDT Benefit Summary Excellent performance benefit with a small amount ofSSD Proven ability to take advantage of larger capacity drives 2X improved response time compared to USPV Increased workload of up to 55% more IO 50% Environmental Cost Savings Significant software savings through consolidation Ease of performance management

HDT for Mainframe and HTSM forMainframe Storage Benefits Enables automation and more efficient use of storage Data that is highly used is in high-performance storage Lower priority jobs can be kept in lower-performing, lower-coststorage Improves ability to manage SLAs High-priority jobs using more expensive, higher-performing storagecan be given it and charged accordingly Jobs with less-demanding SLAs can use lower-cost, lowerperformance storage Ability to manage via DFSMS storage groups Reduced overhead of migrations to ML1

Questions anddiscussion37

Thank you38

Service Processor (SVP) with Hitachi Storage Navigator Operational management can be done from either Storage Navigator or native from z/OS with Hitachi Tiered Storage Manager for Mainframe (HTSM) HTSM features -

Related Documents:

a policy on your on-premises NetApp filer to tier data to Amazon S3. In ONTAP 9.2 and 9.3, FabricPool supports tiering NetApp snapshot copies and data in data protection volumes to Amazon S3 by using the Snapshot-Only and Backup volume tiering policies. In ONTAP 9.4, an auto tiering policy was added. When a NetApp volume tiering policy

on mainframe and non-mainframe dev, test and maintenance tasks File-AID, Xpediter, Topaz Workbench 1. Empirical productivity metrics (ex. Delivery cycle times, etc.) 2. Positive anecdotal feedback 3. Motivate non-mainframe devs to work on mainframe-related activities Large, complex, undocumented mainframe apps impede transformation

Automated Optimized Tiered Storage Management! Before: Tiered storage and provisioning Labor intensive Data classification before tiering Complicated management of multiple storage tiers ! Now: Dynamic tiering and provisioning Controller-based automation Single, self-managed, self-healing,

Bruksanvisning för bilstereo . Bruksanvisning for bilstereo . Instrukcja obsługi samochodowego odtwarzacza stereo . Operating Instructions for Car Stereo . 610-104 . SV . Bruksanvisning i original

82 hitachi zx200-3, zx225usrlc-3 1033091 sprocket sp-zx200-3 83 hitachi zx240-3 1032489 sprocket 84 hitachi ex300-1/2, ex270-1/2 1010467 sprocket 80 sp-ex300-1 85 hitachi ex300-3,ex270-5 1017928 sprocket 80 sp-ex300-3 86 hitachi ex300-5, ex330-5, zx270, zx330, zx350, zx370 1022168 sprocket 85 sp-ex300-5 87 hitachi

applications. Mainframe transformation deals with options to move all applications off the mainframe. Mainframe-as-a-Service (MFaaS) supplements the modernization path, providing a pay-as-you-go (PAYG) business model. Those that do not plan to modernize consider outsourcing mainframe operations. These four quadrants help clients find the

Automated Sign-on for Mainframe enables an end user to automatically log on to a host application on a z/OS mainframe by using a terminal emulation client. Automated Sign-On for Mainframe is an add-on to Host Access Management and Security Server and requires a separate license. In this guide: How Automated Sign-on for Mainframe Works

Animal Food Nutrition Science Public Health Sports & Exercise Healthcare Medical 2.3 Separate, speciality specific listings providing examples of the detailed areas of knowledge and application for each of the five new core competencies required by Registered Nutritionist within these specialist areas have been created and are listed later in this document under the relevant headings. 2.4 All .