SAN Volume Controller Benefits of using it with

IBM Storage Systems IBM Drives Storage Virtualization to the Next Level with IBM System Storage SAN Volume Controller 6. 2 Virtualize IBM and non-IBM disk systems to increase up to 30% storage utilization, reduce the costs associated with disk storage and simplify administration. Designed to pool storage volumes into a reservoir of capacity for centralized management. SVC helps to hide the boundaries among disk systems, which simplify management and enables customers to focus on managing storage as a resource to meet business requirements and not as a set of boxes. Get ready to increase storage efficiency without complexity © 2011 IBM Corporation

IBM Storage Systems Traditional SAN Disk Storage Infrastructure changes lead to high effort and possible downtime Multipath Driver Server to storage logical connections are static Copy services unique to each storage system Multipath Driver SAN islands require multiple management points SAN Adv Function Low capacity utilization, difficult to pool volumes © 2011 IBM Corporation

IBM Storage Systems IBM System Storage SAN Volume Controller (SVC) A more flexible Storage Infrastructure Simplified interoperability planning (Server-SVC; SVC Multipath Driver SVC-Storage) Virtual Disk SVC Multipath Driver Disk infrastructure changes and data migration without application downtime Virtual Disk Centralized management SAN Common set of advanced functions SAN Volume Controller Advanced Copy Services Disk Storage Pool IBM Efficient use of resources ‘gold’ HDS ‘silver’ HP ‘bronze’ EMC Single pool of (tiered) disk storage © 2011 IBM Corporation

IBM Storage Systems SVC Virtualization in More Detail Volumes (Virtual Disks) Vol 1 Vol 2 Vol 3 Vol 4 Vol 5 Vol 6 Vol 7 Vol 8 Vol 9 Vol 10 SVC Virtual-to-physical Mapping Service Class 1 Storage Pools Managed Disks RAID protected Logical Drives MD 1 MD 2 MD 3 MD 4 Service Class 3 Service Class 2 MD 5 MD 6 MD 7 MD 8 MD 9 MD 10 MD 11 MD 12 SVC ‚insulates‘ servers from disk systems, enabling ü Simplified interoperability matrix ü Higher utilization rates LD 1 LD 2 LD 3 LD 4 Disk System 2 LD 1 LD 2 LD 3 LD 4 Disk System 3 ü Dynamic volume expansion / shrink ü Online data migration Hard Drives ü Thin provisioning In this sample, it is assumed that Disk System 1 & 2 and their hard drives have similar performance characteristics ü Tiered storage ü Multi-vendor strategy © 2011 IBM Corporation

IBM Storage Systems SAN Volume Controller Architecture § Highly scalable – capacity & performance – 2, 4, 6 or 8 node cluster – up to 192 GB cache per cluster – Up to 32 PB max virtual storage § Flexible host connection – FC and i. SCSI SVC Cluster § High availability – Shared cluster design – Internal disk mirroring & Split Cluster option § Disaster recovery – Synchronous and asynchronous replication with consistency groups © 2011 IBM Corporation

IBM Storage Systems SAN Volume Controller Storage Engine § 6 th generation SVC engine 2145 -CG 8 – Based on IBM System x 3550 M 3 Server • Intel Xeon 5600 2. 53 GHz quad-core processor • 24 GB of cache (with future growth possibility) • Four 8 Gbps FC ports & two 1 Gbps Ethernet ports – 10 Gbps Ethernet support (2 ports) optional* • For i. SCSI host attachment and future FCo. E support up to 4 node pairs (8 nodes) – Support for Solid State Drives (SSDs) optional* • Up to 4 146 GB SSDs per SVC engine – May be intermixed in pairs with other engines in SVC clusters – Non-disruptive upgrade & maintenance § SVC Virtualization Software 5639 -VC 6 (v 6) preloaded – Licensed per terabyte with cumulative tiers • Base virtualization: total capacity virtualized • Flash. Copy: total capacity for Flash. Copy sources • Metro/Global Mirror: total source and total target capacities *Support for EITHER SAS-attached SSD or 10 Gb. E for each node © 2011 IBM Corporation

IBM Storage Systems SAN Volume Controller Connection to IBM i FC conneciton vscsi IBM i disk drives VTSCSI Hdisks in VIOS Volumes in SVC (Virtual Disks) IBM i VIOS Storage Pools Managed Disks RAID protected Logical Drives Hard Drives © 2011 IBM Corporation

IBM Storage Systems Innovative Scale-out SSD Implementation § SSD support fully integrated into SVC system – Copy services, data migration, management same as with HDDs – Move data to/from SSD without disruption – SSDs in one engine may be accessed through any engine in the cluster § SSD data protection options – Volume Mirroring (SVC v 5) – RAID (0, 1, 5, 6, and 10) protection for SAS-attached SSDs (SVC v 6. 2+ release) Add SSDs to scale capacity (without distruption to engines) SVC designed to deliver maximum I/O capability of SSDs Add SVC I/O Groups to scale throughput and add capacity – Up to 50, 000 read IOPS per SSD – Up to 200, 000 read IOPS per SVC I/O Group – Up to 800, 000 read IOPS per SVC cluster © 2011 IBM Corporation

IBM Storage Systems SAN Volume Controller Graphical User Interface § Based on the XIV UI interface design to make SVC much easier and more productive to use § Newly designed user interface that not only delivers many more functional enhancements and greater ease of use, but can be accessed from anywhere on the network with a web browse © 2011 IBM Corporation

IBM Storage Systems SAN Volume Controller Functionality Highlights § Online data movement / migration § Sophisticated caching § Quality of service with I/O governing § Thin Provisioning § Virtual Disk Mirroring – Local high availability solution § Point-in-Time Copy - Flash. Copy (optional) § Remote Mirroring – Metro/Global Mirror (optional) § Solid State Drive support & IBM Easy Tier – Smart data placement and efficient utilization © 2011 IBM Corporation

IBM Storage Systems Online Volume Migration § Move data from one storage pool to another without losing access § Managed disks can be different types § Source and destination pool must have the same extent size* ü Improves application availability Storage Pool 1 Storage Pool 2 *Use Volume Mirroring to migrate a volume between storage pools with different extent sizes © 2011 IBM Corporation

IBM Storage Systems Volume I/O Governing and Thin Provisioning § IO Governing § Limit maximum amount of I/O activity a host can perform on a volume over a specific period of time § Governing rate can be set in I/Os per second or MB per second § An I/O budget is expressed as a number of I/Os or MBs over a minute and evenly divided between all SVC nodes that service that volume (the I/O group) max I/O or MB per second Volume ü Can be used to help satisfying a Quality of Service requirement § Thin provisioning § Options to manage and control physical disk allocation: Thin Provisioned Volume § Auto-expand / manual (dynamically switchable) § Report on physical disk usage at Volume and Storage Pool level § Helps save disk space and simplifies storage Physical allocation Storage Pools © 2011 IBM Corporation

IBM Storage Systems Volume Mirroring § SVC stores two copies of a volume, usually on separate FC-attached disk systems – SVC maintains both copies in sync and writes to both copies – If one virtual disk fails, SVC provides continuous data access by using the other copy (automatic resynchronization after repair) – Copies can be split – either copy can continue as production copy – Either or both copies may be space-efficient Volume’ Volume ü Protects critical data IBM i against failure of a disk system or disk array – A local high availability function, not a disaster recovery function Storage Pool 1 Storage Pool 2 © 2011 IBM Corporation

IBM Storage Systems Volume Mirroring & SVC Split Cluster ßFC Long Wave (up to 10 km) SVC Split Cluster Disk failure transparent failover High availability solution for IBM i SVC or SAN failure transparent or path switch Location 3 Quorum © 2011 IBM Corporation

IBM Storage Systems Point-in-Time Copy: Flash. Copy § Volume-level local replication function – Copy available almost immediately for use – Background copy operation or ‚copy on write‘ – Source and target volumes can be thinprovisioned § Flexible, enterprise-class functionality – Thin Provisioned Flash. Copy – Multi Target Flash. Copy (256 targets) – Incremental, Cascaded, Reverse – Consistency Group support § Local within a SVC cluster system – Source and target may be on any SVC supported disk systems § Licensed per TB (Source) Source Storage Pool 1 Target Storage Pool 2 • up to 2048 FC mappings § Minimizes backup window of IBM i § Works with IASP or antire disk sapce © 2011 IBM Corporation

IBM Storage Systems Improving Application Testing with Flash. Copy (Sample) § Production systems often have many test systems that are replicas of production § Can be used for entire disk space of IBM i TP Production FA TP FA Test master copy TP TP Test FA: fully allocated TP: thin provisioned © 2011 IBM Corporation

IBM Storage Systems Remote Mirroring: Metro Mirror / Global Mirror § Metro Mirror – synchronous – Designed to maintain fully synchronized copies at both sides – Standard FCP: 10 km – Channel Extenders: up to 300 km § Global Mirror – asynchronous – Designed to maintain consistent secondary copy at all times – Up to 8000 km § Delivered as a single feature – Licensed per TB (source and target) ü Disaster Recovery solution for IBM i Source Target Source Storage Pools –Consitency Group support (up to 256) –Up to 8192 mirror relationships per cluster © 2011 IBM Corporation

IBM Storage Systems Easy Tier § A disk performance optimization feature – Operates on the extent level – a sub-LUN migration § Designed to automate data placement throughout a managed disk group onto two tiers of storage Host System – Automatically and nondisruptively relocates data at the extent level from one tier to another in either direction – Enables I/O to be exercised on the most appropriate storage tier to achieve the best possible performance Dynamic virtual volume – Continuous optimization cycle, no manual involvement – Dynamic data movement is transparent to the host server and application users of the data, other than providing improved performance SSD HDD ü Reduces administrative effort and costs to intelligently align the system with current workload requirements and to reduce response time © 2011 IBM Corporation

IBM Storage Systems Easy Tier - Concept § Automatic storage hierarchy – Hybrid storage pool with 2 tiers • Solid-State Disks (SSD) & hard disks Host § I/O Monitor keeps access history for each virtualization extent – 16 MB - 8 GB per extent – Every 5 minutes – over a 24 -hour period Host Volumes & Extents § Data Placement Adviser analyses history every 24 hours Hybrid storage pool § Data Migration Planner invokes data migration – Promote hot extents – Demote inactive extents § Operation modes – Off mode – Evaluation or measurement only mode – Auto data placment or extent migration mode SSD HDD Automatic extent migration © 2011 IBM Corporation

IBM Storage Systems Tier Selection for External Storage SSDs 21 © 2011 IBM Corporation

IBM Storage Systems Easy Tier - Measurement / Evaluation § Measurement functionality can be enabled without purchasing SSDs – Measurement / evaluation only mode collects volume extent usage statistics – Statistics summary file is created that can be offloaded for analysis – IBM Storage Tier Advisor Tool (STAT) can be used to create a summary report ðData produced will help to determine if SSDs may be beneficial © 2011 IBM Corporation

IBM Storage Systems Real Time Performance Statistics – GUI Screenshot © 2011 IBM Corporation

IBM Storage Systems Tivoli Productivity Center Trending and Planning Data Path Explorer Bottlenecks § More granular data (individual volumes and RAID arrays) § Collected every 15 minutes Disk Utilization Performance / Health Check § Stored in database for longterm analysis § Follow performance trends, enable capacity planning Changed Configuration History Energy Management Tools Longer-term comprehensive performance information © 2011 IBM Corporation