Advantages of Mainframe Replication With Hitachi VSP
1 like4,273 views
This document outlines the capabilities of Hitachi's mainframe storage replication solutions, emphasizing their importance for business continuity and disaster recovery. It discusses features of products like Universal Replicator and TrueCopy, as well as how Brocade's solutions enhance these offerings. Additionally, it provides insights on trends shaping business continuity requirements and includes case studies demonstrating successful implementations of the solutions.
Advantages of Mainframe Replication With Hitachi VSP
- 1. MAINFRAME STORAGE REPLICATION JOE AMATO, HDS PRODUCT MANAGER FOR REMOTE REPLICATION PRODUCTS MARK DETRICK, BROCADE GLOBAL SOLUTIONS ARCHITECT FOR RDR SEPTEMBER 19, 2012
- 2. WEBTECH EDUCATIONAL SERIES Mainframe Storage Replication In this session, learn how the Hitachi Virtual Storage Platform's mainframe replication capabilities can address your business continuity and disaster recovery requirements. Also, learn how Brocade switches and directors complement HDS mainframe replication capabilities and add value to HDS solutions. Specific Learning Objectives - By attending this webcast, you’ll learn: 1. Trends that are driving changes to business continuity requirements, and how HDS replication products such as Universal Replicator and Hyperswap integration capabilities with Business Continuity Manager are best positioned to address them. 2. Examples of how companies have implemented complete HDS solutions to solve their mainframe BC/DR needs. 3. The key features and functions of Brocade FCIP switches and fibrechannel/FICON director inter-datacenter connectivity that provide additional value to HDS replication solutions.
- 3. UPCOMING WEBTECHS Mainframe Series Why Networked FICON Storage Is Better than Direct-attached Storage, Oct 3, 9 a.m. PT, 12 p.m. ET Other Storage Analytics, Sept 20, 9 a.m. PT, 12 p.m. ET Maximize Availability and Uptime by Clustering your Physical Datacenters within Metro Distances, Oct 24, 9 a.m. PT, 12 p.m. ET Check www.hds.com/webtech for Links to the recording, the presentation and Q&A (available next week) Schedule and registration for upcoming WebTech sessions
- 4. AGENDA Introduction HDS mainframe replication portfolio Key HDS solution highlights Customer case studies Brocade FCIP switches and key features for replication Summary and Q&A
- 5. KEY TREND: SHRINKING RTO FOR TIER-1 APPLICATIONS Source: Enterprise Strategy Group, 2008 & 2010 Percent of ‘Enterprise’ companies (> 1000 employees) with minimum RTO for most-critical applications
- 6. BUSINESS CONTINUITY & DATA PROTECTION ALIGNING RECOVERY OBJECTIVES AND ARCHITECTURES Operational Recovery Local replication ‘crash consistent’ copies (clones/snapshots) Application aware copies Backup/Restore Disaster Recovery Remote replication clones w/ ‘crash consistent’ clones/snapshots Application aware copies Backup/Restore Regulatory Compliance Replicated Archive Solution Architectures Operational Resilience Server/Storage Clusters w/ Synchronous Replication 3rd SitePrimary Site Alternate Site RTO RPOTier-1 Tier-n Recovery Objectives Zero Zero Mins-to- Hrs Secs-to- Mins Hrs-to- Days Hrs-to- Days 3DC Sync-Async Dual-Async Multi-target Delta- Resync Cascade Data Center Relocation Less Resiliency More Resiliency
- 7. BUSINESS CONTINUITY – MAINFRAME PORTFOLIO In-System Replication Solutions Remote Replication Solutions Replication Management Solutions In-System Replication Solutions Remote Replication Solutions Replication Management Solutions Hitachi ShadowImage® Replication For full-volume clones of business data with consistency (ATTIME split with HUR) Hitachi TrueCopy® Synchronous, consistent clones at remote location up to 300km (~180 miles) Business Continuity Manager Replication management in z/OS environments Hitachi Compatible Mirroring for IBM® FlashCopy® Point-in-time volumes and datasets via IBM command set in z/OS environment Hitachi Universal Replicator (HUR) Any distance, unique use of “pull” technology and journal volumes to accommodate link outages or interruptions Hitachi Replication Manager with mainframe BCM plus “enterprise-wide” GUI replication management and monitoring for mainframe and open systems Hitachi Compatible Software for IBM® FlashCopy® SE PiT volumes with virtual volume pools versus whole volumes as targets Hitachi Compatible Replication for IBM® XRC® Asynchronous remote replication via compatible with IBM XRC in z/OS environment z/OS Basic HyperSwap, GDPS HyperSwap certification and/or integration High-availability storage for z/OS environments Replication assessment, migration and implementation services Best-practice designs that ensure results and lower risks in implementation Replication scorecard (health check) services Periodic checkups that ensure continued solution value
- 8. TRUECOPY REMOTE REPLICATION (SYNCHRONOUS) (1) Host write (4) Write complete (2) Synchronous remote copy (3) Remote copy complete PVOL SVOL Previously 2 Round-trips Now, 1 Round trip! • Zero Data Loss possible with Fence-level = Data • Performance: “dual-write” plus one (1) round-trip latency and a little “overhead” • Up to 300 km supported – A couple of customers with up to ~ 140 km • Consistency groups supported for open and mainframe • Consistency groups supported across open and mainframe • zOS support is PPRC command compatible
- 9. HITACHI UNIVERSAL REPLICATOR HIGHLIGHTS Hitachi Universal Replicator and Data Resilience ‒ Protects production performance during replication anomalies ‒ Protects Recovery Point Objective (RPO) with replication resilience during replication anomalies ‒ Leverages HDS virtualization for lower-cost BC/DR solutions ‒ Can enable reductions in bandwidth requirements Key Features ‒ Asynchronous replication of customer’s data ‒ Unique cache and disk based journaling ‒ Unique pull technology to move data to target storage system ‒ Multi-volume and multi-storage-system consistency groups ‒ Consistency across mainframe and open systems supported ‒ Thin (dynamic) provisioning awareness ‒ Advanced configuration support (e.g., 3DC, 4DC, delta-resync)
- 10. RPO ESTIMATE WITH LESS THAN PEAK BANDWIDTH
- 11. Journal (HUR) vs. Cache (Bitmap) RPO Recovery Bitmap RPO HUR RETURN TO NORMAL OPERATION: DISK JOURNALS AVOID SUSPEND IMPACTS TO RPO * Modeled data only Outage Repaired No suspend, decreasing RPO Suspend, increasing RPO
- 12. THREE DATA CENTER REPLICATION SOLUTIONS TrueCopy S-VOL shared as HUR P-VOL in intermediate site Primary volume is shared P-VOL for two remote systems VSP HUR JNL Group P-VOL VSP S-VOL P-VOL JNL-VOLTrueCopy (Sync) 3DC Cascade* S-VOL VSP JNL-VOL JNL Group VSP JNL-VOL HUR JNL-VOL Journal Group P-VOL TrueCopy (Sync) or HUR* 3DC Multi-Target S-VOL VSP JNL-VOL JNL Group Optional Delta Resync VSP S-VOL S-VOL VSP JNL-VOL JNL Group * 3DC with two HUR (async) links, without Delta-Resync, is currently supported on VSP platforms only in both cascade and multi-target 1x1x1 configurations. Mainframe supports up to 12x12x12 controller configurations (4x4x4 cascade) Requires Disaster Recovery Extended and BCM Extended
- 13. HUR INTEGRATION WITH HYPERSWAP HIGH-AVAILABILITY WITH DISASTER RECOVERY TrueCopy Duplex HUR Duplex Delta Pairs Hold X HyperSwap Delta Pairs Hold HUR Duplex BCM detects HyperSwap via system messages and drives automated change of HUR states
- 14. MAINFRAME REPLICATION MANAGEMENT “REFRESHER” Hitachi - IBM Compatibility ‒ Hitachi storage has essential compatibility with IBM z/OS ‒ IBM replication management compatibility ‒ Hitachi storage controllers support PPRC, XRC and FlashCopy commands (assuming applicable software licensing) without additional HDS-provided software ‒ TrueCopy synchronous is compatible with PPRC ‒ For full-volumes, IBM commands operate on individual replication pairs ‒ But, IBM commands do not support Hitachi Universal Replicator Hitachi Business Continuity Manager (BCM) value-add ‒ Reduce complexity, simplify management ‒ Provide enhanced control and visibility ‒ (e.g., support configured and unconfigured devices) ‒ Reduce risk
- 15. Automatically discovers Hitachi storage systems and internal copy volumes Uses a single command to manage groups of copy volumes with common attributes Receives auto notification of key copy “state transitions” Supports advanced TrueCopy, ShadowImage and Universal Replicator features Real-time view of Universal Replicator Single, consistent interface based on Time Sharing Options (TSO) / Interactive System Productivity Facility (ISPF) Standard REXX scripting to customize and automate Optional Integration with Replication Manager for combined mainframe and open systems replication management BUSINESS CONTINUITY MANAGER™ HIGHLIGHTS
- 16. CUSTOMER CASE: INSURANCE COMPANY (VIRTUALIZATION & REPLICATION FOR MAINFRAME AND OPEN SYSTEMS) Customer originally implemented on USPV and is now moving to VSP! Case study available on SHARE Proceedings - Anaheim 2011 Production Site Recovery Site
- 17. CUSTOMER CASE: LARGE FINANCIAL INSTITUTION (MULTI-CONTROLLER CONSISTENCY/HUR AND SHADOWIMAGE) Converted from XRC! Case study available on SHARE Proceedings – San Jose 2008 Much larger configuration today! PRODUCTION PRODUCTION PRODUCTION Dual OC-192 M I R R O R E D D A T A Fiber Channel connections between control units Primary DASD Production Site Recovery Site Secondary DASD Tertiary DASD Monitoring/Management
- 18. Extraordinary Extension Brocade Brocade 7800 & FX - Brocade CONFIDENTIAL 18March 2012
- 19. Next-Generation Extension Products March 2012Brocade 7800 & FX - Brocade CONFIDENTIAL 19 High-touch services, product flexibility, scalability Solutionpricerange(hardwareandsoftware) Global 1000/ Large Enterprise Director-ClassStandalone 7800 Base Small/Medium Business SMB/Enterprise Brocade 7800 + PoD license FX8-24 blade for Brocade DCX 8510 DCX 8510-4 DCX 8510-8 FX base GE - 10 Gbps FX 10GE license 10GE - 20 Gbps Brocade Network Advisor
- 20. Remote Data Replication over IP Networks Business Challenge • Create nonstop IT operations for business continuance • Application failover and data access Solution • Leverage secondary site for failover • Deploy direct-connected FX8-24 Extension Blades over MLX IP network • Provide protected data, very small RPO and RTO March 2012Brocade 7800 & FX - Brocade CONFIDENTIAL 20 Disk Array with RDR Mainframe and/or Open Systems IP WAN Mainframe and/or Open Systems Disk Array with RDR Brocade DCX FX8-24 Extension Blades Brocade DCX FX8-24 Extension Blades Secondary Data Center Primary Data Center FICON FC MLXe L2/L3 Switch MLXe L2/L3 Switch Arrays replicate data between each other using FC
- 21. Brocade Extension Solutions Accelerates and optimizes replication, backup, and migration over any distance using advanced Fibre Channel and FCIP networking technology • FCIP Trunking creates an ISL spanning multiple physical ports for load balancing and network failure resiliency • ARL dynamically adjusts bandwidth limits to ensure efficient utilization and sharing of available bandwidth • PP-TCP-QoS prioritizes handling of initiator-target flows within an FCIP tunnel to optimize bandwidth and performance by application • Advanced Compression architecture provides flexibility to optimize compression ratios and maximize throughput • 200 ms Round-Trip Time (RTT) of latency supports distances approaching 17,500 kilometers (nearly 11,000 miles) • Brocade Network Advisor FCIP management unifies management of Fibre Channel and FCIP network infrastructures Features Introduced with 7800/FX March 2012Brocade 7800 & FX - Brocade CONFIDENTIAL 21 Brocade 7800 Channel Extender Brocade FX8-24 Extension Blade for the Brocade DCX 8510
- 22. Brocade FX8-24 Extension Blade • Brocade FX8-24 Extension Blade • Twelve 8G FC ports, ten GE interfaces • Functionality enabled through optional licenses • 10GE Interfaces • Enables 10GE interfaces and doubles FCIP bandwidth to 20 Gbps • Supported 10GE interface configurations • Ten GE and one 10GE or • Two 10GE • Advanced Extension: FCIP Trunking and ARL • Adaptive Networking: Activates FC and FCIP QoS functionality • Enterprise Bundle: Advanced Performance Monitoring, Adaptive Networking, Fabric Watch, Brocade Trunking, and Server Application Optimization (SAO) • Advanced Accelerator for FICON: Enables high-performance FICON tape, XRC, and Teradata over distance • FICON Management Server: Control Unit Port (CUP) enables host control of switches in mainframe environments Blade Options March 2012Brocade 7800 & FX - Brocade CONFIDENTIAL 22 Twelve 8 Gbps FC or FICON ports Ten GE FCIP interfaces Two 10GE FCIP interfaces
- 23. Brocade 7800 Extension Switch • Two models • Brocade 7800 4/2: Four 8G FC ports, two GE interfaces • Brocade 7800 16/6: Sixteen 8G FC ports, six GE interfaces • Functionality enabled through optional licenses • Brocade 7800 16/6 Upgrade License: Enables all ports and OSTP • Advanced Extension: FCIP Trunking and ARL • Adaptive Networking: FC and FCIP QoS functionality • Enterprise Bundle: Advanced Performance Monitoring, Adaptive Networking, Fabric Watch, Brocade Trunking, and SAO • FICON Management Server: CUP enables host control of switches in mainframe environments • Advanced Accelerator for FICON: Enables high-performance FICON tape and replication over distance Platform Options March 2012Brocade 7800 & FX - Brocade CONFIDENTIAL 23
- 24. Brocade Network Advisor March 2012Brocade 7800 & FX - Brocade CONFIDENTIAL 24
- 25. FCIP Trunking • Is an FCIP Tunnel with more than 1 circuit • FCIP Trunking provides • Single logical ISL in routing table as a single link, ULP see single link • Bandwidth Aggregation of each circuit • Load Balancing per batch across circuits • Failover to remaining circuits • Lossless Link Loss (LLL) • Data in-flight is not lost when a link goes down, will be retransmitted same as with TCP • In-Order-Delivery (IOD) Does not require IOD to be enabled • Data in-flight will be delivered in the correct order, even after data lost in-flight • Works with both FICON and FC • Supports FastWrite, OSTP, and FICON emulation over multiple circuits • FCIP Trunking is a proven technology • Leveraged from widely deployed McDATA/CNT technology Brocade exclusive feature March 2012Brocade 7800 & FX - Brocade CONFIDENTIAL 25
- 26. VE_Port to VE_Port (Dedicated Ethernet Ports) March 2012Brocade 7800 & FX - Brocade CONFIDENTIAL 26 VE_Port 16 FCIP Trunk1 Logical Source IP Interfaces Logical Destination IP Interfaces Physical Ethernet Interfaces VE_Port 16 FCIP Trunk 1 Physical Ethernet Interfaces Circuits Circuits VE_Port FCIP Trunk IP Interface Circuit Ethernet Interface Unique IP Address Unique IP Address Unique IP Address Unique IP Address GE0 GE1 GE2 GE3 GE0 GE1 GE2 GE3 GE4 GE5 Unique IP Address Unique IP Address GE4 GE5 WAN IP Cloud • FCIP Trunks (multiple circuits) or Tunnels (single circuit) are logical entities • IP interfaces are logical entities • Circuits and Ethernet interfaces are physical entities
- 27. FCIP Trunking Concept Supervisor TCP Session • Encapsulates the circuits’ TCP sessions within a FCIP Trunk • Provides granular load-balancing • Retransmits lost frames due to circuit loss • For operational links, the TCP sessions within the circuit handle loss • Ensures FC frames are delivered in order (IOD) • Prevents IFCC on the mainframe • Supports FastWrite, OSTP, and FICON emulation • FCIP Trunk appears as a single logical ISL L3 Lossless Link Loss (LLL) March 2012Brocade 7800 & FX - Brocade CONFIDENTIAL 27 13 4 X X XCircuit 1 with TCP session Circuit 2 with TCP session FCIP Batches FCIP Batches 3? Supervisor TCP Session Data Received 3 FICON Emulation FastWrite OSTP FICON Emulation FastWrite OSTP 2
- 28. Adaptive Rate Limiting High Availability Architecture March 2012Brocade 7800 & FX - Brocade CONFIDENTIAL 28 Channel Extender A1 Channel Extender A2 Channel Extender B1 Channel Extender B2 RDR Application ARL Min ½ OC-12 Max OC-12 ARL Min ½ OC-12 Max OC-12 IP Network OC-12 ½ OC-12 = 311 Mbps Site A Site B Fully Utilized Normal Operation • Link BW Number of Interfaces = BW of interface • Not oversubscribed or underutilized • Rate Limiting automatically changes with device or link failures •Optics, smashed cables, FW upgrades, configuration errors, defects… Offline Path Operation • Link BW Number of Remaining Interfaces = new BW of interface • Not oversubscribed or underutilized • Applies to the entire path from end device to end device ARL Not Operating ARL Operating at OC-12 50% 50%100% ARL Operating at ½ OC-12 ARL Operating at ½ OC-12
- 29. ARL Optimization after a path failure March 2012Brocade 7800 & FX - Brocade CONFIDENTIAL 29 0 1000000 2000000 3000000 4000000 5000000 6000000 7000000 8000000 9000000 0 50000 100000 150000 200000 250000 300000 350000 With one interface online, all available bandwidth is consumed up to the maximum rate When a second interface is enabled, the network experiences a congestion event and the first interface drops its rate down to the minimum configured rate. From here it will seek the ceiling again The second interface starts off by claiming its minimum configured rate Offline equipment/link results in available BW This interface ramps up to fully utilize the available WAN BW. It stops seeking a ceiling at the maximum configured rate
- 30. March 2012Brocade 7800 & FX - Brocade CONFIDENTIAL 30 FC ISL Virtual Channels PP-TCP-QoS FCIP Circuit FCIP Tunnel TCP Sessions DCX QOSH_<zone name> High Priority QOSL_<zone name> Low Priority 7800 7800 FC ASIC will put traffic into priority VC 7800/FX will map VCs into PP-TCP-QoS
- 31. IPsec • There is no license for IPsec • It is included in the base unit • It adds only 5 µs of latency • It operates at line rate • All it costs is the effort to configure • Uses PSK (Pre-Shared Keys) March 2012Brocade 7800 & FX - Brocade CONFIDENTIAL 31
- 32. Typical Four-Brocade 7800 Solution • Solution: •2 FCIP Trunks •1 FCIP Trunk per 7800 • 2 circuits in each trunk •Array sees 2 parallel paths • ARL used on all 4 circuits •Max is link BW or interface line rate, whichever is smaller •In this example: Min = Max 4 March 2012Brocade 7800 & FX - Brocade CONFIDENTIAL 32 FC path A FC path B FC path A FC path B 7800 4/2 7800 4/2 IP WAN Link GE FCIP GE FCIP GE FCIP GE FCIP Service Provider (Possibly carrier Ethernet) 7800 4/2 7800 4/2 MLX Core MLX Core MLX Core MLX Core
- 33. High-Availability Design • Four Brocade 7800/FX • Two IP Networks • Most RDR applications load-balance across multiple paths • Most RDR applications will fail over to the remaining paths • Supports FastWrite, OSTP, and FICON Emulation Multiple paths March 2012Brocade 7800 & FX - Brocade CONFIDENTIAL 33 FCIP FCIP Storage Local Storage Remote 7800 7800 7800 7800 Fabric 1A Fabric 1B Fabric 2A Fabric 2B WAN A WAN B
- 34. SUMMARY Final thoughts ‒ HDS is rooted-in and committed-to strong IBM compatibility ‒ HDS has complete replication solution portfolio for mainframe storage ‒ HUR is mature product (2004) with over 250 worldwide mainframe replication customers; and is on third generation of hardware platform support ‒ HDS storage provides one platform for virtualization, dynamic (thin) provisioning, and tiered storage management ‒ Brocade provides value-add to HDS replication solutions Benefits: ‒ Deliver SLAs that are better aligned with business objectives ‒ Reduced risk via proven, enterprise-class data protection ‒ Cost savings through: ‒ Best of breed asynchronous replication ‒ Seamless tiered storage management ‒ Protection of existing investments with external virtualization ‒ Dynamic provisioning and tier-ing for mainframe
- 36. UPCOMING WEBTECHS Mainframe Series Why Networked FICON Storage Is Better than Direct-attached Storage, Oct 3, 9 a.m. PT, 12 p.m. ET Other Storage Analytics, Sept 20, 9 a.m. PT, 12 p.m. ET Maximize Availability and Uptime by Clustering your Physical Datacenters within Metro Distances, Oct 24, 9 a.m. PT, 12 p.m. ET Check www.hds.com/webtech for Links to the recording, the presentation and Q&A (available next week) Schedule and registration for upcoming WebTech sessions
- 37. THANK YOU