Rapid Recovery from Logical Corruption

Transcription

1 Rapid Recovery from Logical Corruption Brett Quinn DellEMC 1 Nov 2016 Session DD

2 Topics Physical vs Logical Recovery SnapVX zdp: Data Protector for z Systems Isolated Recovery Solutions

3 Logical vs Physical Data Recovery Synchronous Primary Secondary DC2 DC1 Asynchronous Tertiary (Out Of Region) DC3

4 Physical Recovery SWAP Synchronous Primary Secondary DC2 DC1 Asynchronous SWAP Tertiary OOR DC3

5 Logical Recovery Primary Synchronous Secondary DC2 DC1 Asynchronous Tertiary OOR DC3

6 Logical Recovery Primary Synchronous Secondary DC2 DC1 Asynchronous Tertiary OOR DC3

7 Data Interdependence Data Lake Mobile Data Batch Flat Files DB2/IMS/ Other

8 8:15 SnapVX 10:00 9:15 8 AM 8:45 9:30 9:45 8:30 9:00

9 Incorporate Snaps for Data Copies New TimeFinder SnapVX Production Volume REDUCED IMPACT Target-less Snapshots Snapshot INCREASED AGILITY Up to 256 Snapshots per source Up to 1024 Linked Targets (snaps and/or clones) per source EASE OF USE User-defined name/version number Automatic expiration if desired Snapshot Linked Target Snapshot

10 TimeFinder SnapVX Pointer-Based SNAPs Source Device (TDEV) CREATE & ACTIVATE TERMINATE Snapshot (up to 256) 10 A.M. LINK / UNLINK Target Device(s) (TDEV) 12 Noon Host volume Snapshot Pointer Data structure Storage Resource Pool

11 zdp Data Protector for z Systems

12 zdp A z/os based solution that automates SnapVX snapshot creation/deletion Enables rapid recovery from a spectrum of risks to data: Processing error Human error Malicious intent Think: a time machine for the mainframe Scheduler

13 Why zdp? Logical vs Physical Protection DC1 Replication solutions focus the physical: availability with data integrity Planned and Unplanned outage events S Asynchronous P Multiple copies EVERYTHING gets physically replicated Even the logical corruptions! DC3 Synchronous S DC2

14 Why zdp? Logical vs Physical Protection DC1 Replication solutions focus the physical: availability with data integrity Planned and Unplanned outage events Multiple copies EVERYTHING gets physically replicated Even the logical corruptions! Today 1 additional full PIT copy Copy 1: near real time Copy 2: new PIT created every 24 hrs. 24 hours to find corruption Up to 24 hours of data loss 1 2 DC3 Asynchronous Synchronous DC2

15 Why zdp? Logical Protection zdp focuses on the logical: recovery from loss of data integrity Continual point in time copy creation Automated Selectable recovery points Brings Point in Time recovery capability to database and non-database systems Provides applications with recovery capability when not designed-in Provides cross-application recovery point Supports both FBA and CKD data Source Target Snapshot 1 Snapshot 2 Snapshot N

16 Rethinking DB2 Image Copy and Recovery processes To Image Copy or Not to image copy, that is the question System Level Backup in DB2 supports use of individual objects for recovery Creation of an image copy of an object from an SLB Use the SLB for object recovery and improve concurrency during recovery of multiple objects Benefits of using zdp for DB2 recovery Recovery Acceleration due to: Elimination of data restore time More recent copies of data mean less log to be traversed Reduced Image Copy processing Saves CPU and I/O resources If you use Virtual Tapes, the capacity requirements are minimized If you use Physical Tape, you need fewer tapes If you use Disk, the storage capacity requirements and network capacity (if replicated) are reduced

17 zdp Structural Elements Source Linked Target Snapshot 1 Snapshot 2 Versioned Data Group (VDG) Logical group of source volumes that associates volumes along with settings for creating multiple PIT images. Snapshot: Pointer based, point in time image of a single volume Snapset: Named point in time consistent image of all source volume snapshots in a VDG. Snapset capture: Continuous automated process of creating Snapsets Target set: Group of devices which have been defined to zdp as the desired LINK targets to a snapset. Snapshot 256

18 zdp deployment options zdp SCF source snap source snap source snap zdp supports snapset creation: Locally attached to VMAX3

19 zdp deployment options zdp SCF R1 snap SRDF/S or SRDF/A or PPRC R2 snap R1 snap R2 snap R1 snap R2 snap zdp supports snapset creation in: Locally attached VMAX3 Remote to VMAX3 via SRDF/S or SRDF/A (up to 4 hops) R1 array can be VMAX2 for sending SnapVX commands to a VMAX3 for remote execution

20 zdp deployment options R1 snap SRDF/S or SRDF/A or PPRC R2 snap zdp SCF R1 snap R2 snap R1 snap R2 snap zdp supports snapset creation in: Locally attached VMAX3 Remote to VMAX3 via SRDF/S or SRDF/A (up to 4 hops) R1 array can be VMAX2 or VMAX3 Direct attached to SRDF target array

21 Z Data Protector Summary Source Linked Target Snapshot 1 Snapshot 2 Host automation solution Frequent creation/deletion of snapshots Benefit: Fast recovery from logical corruption Features Summary Up to 144X more recovery capability compared to a daily snapshot Keep up to 256 versions of each volume Select snapshots to LINK by timestamp Automatic roll-off of aged snapshots Space estimation and monitoring Consistent scale (32,000 volumes) Snapshot 256

22 Building Business Resiliency Isolated Recovery Solutions

23 Cyber attacks get sophisticated Are you staying ahead of the criminal evolution? Traditional Threats Emerging Threats Cyber Theft Denial of Service Attacks Cyber Extortion Cyber Destruction Isolated Recovery Solutions protect against these types of attacks

24 REGULATORY CYBER DESTRUCTION PREVENTION GUIDANCE National Security Agency: best practices to protect information systems and networks from a destructive malware attack include... Segregate network systems Federal Reserve System: financial institutions should consider taking the following steps... such as logical network segmentation, hard backups, air gapping [and] physical segmentation of critical systems Federal Financial Institutions Examination Council (FFIEC): An air-gapped data backup architecture limits exposure to a cyber attack and allows for restoration of data to a point in time before the attack began. European Banking Authority: Competent authorities should assess whether the institution has comprehensive and tested business resilience and continuity plans in place Securities and Exchange Commission: A series of high-profile recent events involving systems-related issues further highlights the need for market participants to bolster the operational integrity of their automated systems in this area. National Association of Insurance Commissioners:... it is vital for state insurance regulators to provide effective cybersecurity guidance regarding the protection of the insurance sector s data security and infrastructure..

25 What are regulators Focused On? Executive Management needs to establish how much cyber and data breach insurance levels are appropriate for the firm Enterprise environments need to have a Cyber Breach Management Program that includes a life cycle approach to data protection Environments must implement multiple data replication technologies and methods to reduce the probability of data contamination Environments must deploy a multi-tiered approach to where data copies are stored to enable restore and recovery of business and environment data

26 WHAT IS ISOLATED RECOVERY? WHAT DOES IT PROTECT AGAINST? Characteristics: A Protected Copy of Data that is: Isolated Off-Net No Connections Cannot be directly accessed Multiple Roll-Back Points Maintained How: Periodic Copies of Critical Data Systems are isolated Data is Periodically Validated What does it protect against? Cyber Destruction Data Hostage BECAUSE YOU NEED TO THINK LIKE A HACKER!!!

27 WHY YOU NEED BOTH IR AND A DR SOLUTION ISOLATED RECOVERY IS NOT TRADITIONAL BC/DR IRS Systems are Isolated An isolated data center environment that is disconnected from the network and restricted from users other than those with proper clearance. Data is either Copied Periodically over an Air Gap or persistently replicated over a secure link infrastructure Software to perform data copies to a secondary storage and backup targets as well as procedures to schedule the copy over air gap between the production environment and the isolated recovery area or preserve a copy in the IR environment. Integrity Checking & Alerting Workflows to stage copied data in the isolated recovery zone and perform periodic integrity checks to rule out that it was affected by malware, along with mechanisms to trigger alerts in the event of a breach. Recovery & Remediation Procedures to perform recovery / remediation after an incident.

28 The most critical data first Compute Applications Validate & Store Protect the heartbeat of the business first Prioritize top applications or data sets to protect Usually less than 10% of data (this is NOT DR!) Start with a core set and build from there Highest Priority Data

29 Mainframe IR SOLUTION OVERVIEW Primary DC Corporate Network DR / BU Site FICON & Network Air Gap SRDF or Replicator Only Isolated Site Isolated data center environment Disconnected from the network Business data continuously received minimizing RPO Restricted from users other than those with proper clearance Planning: Business Critical Systems RPO/RTO Throughput Requirements Isolation: Isolated DC Environment Dedicated Link SRDF or Replicator Only FICON & Network Air Gap Hidden Local Copies that cannot be deleted Validation: Maintain gold copy Compare to incoming copy Tools application dependent Rate of Change Sentinel Record Injection Recovery: Standard failback mechanisms Additional remediation / validation

30 Concept: MF Isolated Air-Gapped Recovery System Preserves PiT copies on physically isolated VMAX Isolation from: Personnel Networks Primary Site zdp task FICON Production VMAX FICON either: 1. Not Connected 2. Not Genned Isolated Room Ensures recoverability from catastrophic large scale data corruption or deletion R 11 SRDF Persistent or periodic R2 zdp zdp Linked Targets Enhanced security for PiT copy Dual actors Access Control Lists SRDFA to Secondary Site For Disaster recovery Isolated Recovery VMAX

31 Concept: MF Isolated Air-Gapped Recovery System Preserves PiT copies on physically isolated VMAX Isolation from: Personnel Networks Primary Site FICON Production VMAX FICON either: 1. Not Connected 2. Not Genned Isolated Room Ensures recoverability from catastrophic large scale data corruption or deletion R 11 SRDF Persistent or periodic R2 zdp zdp Linked Targets Enhanced security for PiT copy Dual actors Access Control Lists SRDFA to Secondary Site For Disaster recovery Isolated Recovery VMAX zdp task

32 DLM Isolated Recovery Solution Components Production CEC R1 FICON & Network AIR GAP P DD or VNX Replication Isolated Recovery System S Snap zdp Changing Data Mgmt CEC Scheduled scripts creates several space efficient Fail- Safe library Images All TCPIP Sockets Closed on DLm storage except Replicator Separate dedicated storage replication ports Separate Dedicated Replication Network Backup Volumes / Restore Points Instant Access*** Retention Periods / WORM Cost Effective RPO = Hours to 1 Day* Primary Storage Backup Appliance zdp Isolated Recovery System RTO = Hours to Days** Can co-exist with and enhance an existing BC/DR Solution Tape Data is continuously periodically copied in to IRS environment File System Snapshot multiple space efficient Fail-Safe Llibrary images Management CEC is hardwired to IRS Solution Un-addressable Snapshot Copies allow multiple restore points Management CEC can be used used to periodically validate data * Adjustable based on desired security window ** Depends on # of Volumes, Data etc. ***Lower Performance

33 Summary Recovery from Logical corruption is the next wave of Business Continuity planning Operational errors and malicious attacks are real and emerging threats. Tools and procedures must be developed based on perpetual and persistent data copy technologies Consider the applicability of Isolated Recovery Solutions for your enterprise

34 Session feedback Please submit your feedback at Session is DD