Extended Distance Replication With Mirror View A Mirror

Extended Distance Replication With Mirror. View / A

Mirror. View / A Asynchronous Replication via Delta Sets • Delta Set technology lowers bandwidth requirement – Uses delta technology to create “Delta Set” for incremental update – Replicates only the changed blocks during Mirror. View / A update cycle • Automatic update – Disk-based Delta Set is sent to target periodically – Customer-defined update frequency • “Gold Copy” protection – Target automatically protected via “point-intime” golden copy during the update cycle – Ensures atomic update – Protection against partial update scenarios Periodic Transport of Delta Set Source Delta Set Target Gold Copy

Mirror. View / A – Features • Data replication is decoupled from host I/O – Eliminates I/O latency overhead associated with synchronous solutions • Maintains consistent copy of data at all times – Guaranteed consistency across a set of LUNs in the system via consistency groups – Gold Copy protection on the secondary side • Extended distance data mirroring solution – Incremental tracking and update capability – Resilient to high latency networks • Minimum network link supported is T 1 • Supported systems: CX 3 -80, CX 3 -40, CX 700, CX 600, CX 500, CX 400

Mirror. View / A – Implementation Host I/O ( A D -> -> A’) D’) 11 A. M. 10 9 AM AM Source LUN A’ A B C D’ D Reserved LUN Pool COFW* 9 A. M. 11 A. M. COFW Target LUN A’ A BB C D D’ D Tracking Map 1 0 0 1 Delta Set A’ D’ Update Freq = 2 hrs A C B D Gold Copy (9 AM) Transfer Map MV/A 1 0 0 1 11 A. M. Source Site (Orlando, FL) * COFW : Copy On First Write IP/SA N MV/A 9 A. M. Target Site (Anaheim, CA)

Mirror. View / A – Storage System based Consistency Groups • Mirror. View / A mirrors are grouped for – Operations: Synchronization, promote, fracture – Applications • Database consistency • Prevents rolling disaster scenario • Atomic update across all the members of the consistency group – A Gold Copy of all the secondary mirrors is maintained for protection during group update • All the members (primary or secondary) must reside within the same storage system

Understanding Consistency The Dependent Write I/O Principle • • The logical dependency between write I/Os; that is embedded by the logic of an application, operating system, or database management system – A write I/O that will not be issued by an application until a prior related write I/O has completed • A logical dependency, not a time dependency • Inherent in all Database Management Systems (DBMS) – Page (data) write is dependent write I/O based on a successful log write Ensuring ‘dependent write consistency’ is the basic principle behind all of EMC’s Consistency Technology solutions – Enterprise SRDF Consistency Groups – Enterprise Time. Finder Consistent Split – EMC Consistent SNAPs and Consistent Clones – SRDF/A – Mirror. View/A

Mirror. View/A Storage System based Consistency Groups & Databases • Mirror. View / A Consistency Groups (CG) – During periodic update • Does not require the application to be stopped – Supports database software restart feature after a power failure • On source side – Creates consistent split across all the members of the CG • Momentarily freezes I/O for all the members for the CG to create a point-in-time view • Prevents dependent I/O from getting out of sync • On target side – Apply Delta Set update atomically across all the members of the CG – All the members are protected with Gold Copy – Creates a restartable point-in-time view of the data

Mirror. View/A Storage System based Consistency Groups – Implementation Atomic Operation Across all Members DB Src 1 DB Src 2 Log Src 1 Log Src 2 DB Trgt 1 DB Trgt 2 Log Trgt 1 Log Trgt 2 Gold Copy Delta Set MV/A Source Site (Orlando, FL) IP/SAN Mirror. View/A Consistency Group MV/A Target Site (Anaheim, CA)

Mirror. View / A – Consistency Group Operations