refresh_DR_lopresti by wuzhengqin


									                Storage Refresh and
                 Disaster Recovery

April 2, 2008            1
             The Current State

• IBM high end and EMC mid level fiber
  storage frames in the Primary Data Center
• Small network or direct attached storage
• Local storage
• McData fiber switches
• A 1 year old Clariion mid level storage frame
  in the Secondary Data Center

                 The Issues
• Age
• Lack of features that would better support
  the diverse needs of the University
• Our applications require a faster and more
  reliable storage infrastructure
• Critical need for Disaster Recovery
• We are running out of space!

                         OIT Solution
• An EMC high end storage array for each Data Center (DMX4-2500)
    • Can hold up to 2400 drives
    • Redundancy at the component level - if there is a failure 1/64 of the
      storage affected. Called Direct Matrix Architecture
    • Supports fiber channel, ATA drives
    • Supports tiering within the array
    • Non disruptive upgrades
• Two Cisco Fiber Switches for each Data Center (Cisco 9509)
    • >1000 connections possible
    • Supports 10 Gbps
• Control Software
    • Can set priorities on data delivery
    • Performance monitoring
    • Reporting
• Storage Based Replication software
• Snapshot software

             What’s Next
• Site preparation
• Installation
• Training
• Analysis of optimum data placement
  and policies
• Data migration(s)
• EMC services

Disaster Recovery

  Status Update


• Provide tape based recovery for audit-
  defined critical mainframe applications
  at a commercial recovery site
• Provide storage, server, and software
  based recovery for audit-defined
  critical open systems applications at
  the OIT Secondary Data Center

                Disaster Recovery Tests
• Award offsite mainframe contract to SunGard -        Summer 2007
• Award SAN purchase to Dell/EMC –                     February 2008
• Readiness assessment with SunGard -                  Fall 2007
• Conducted first disaster recovery test of basic mainframe systems
  and connectivity to the Secondary Data Center -      March, 2008
• Conduct second disaster recovery test for mainframe applications
  and critical mainframe application testing from the Secondary Data
  Center -                                             May 2008
• Conduct tape based recovery of selected applications in the
  Secondary Data Center –                              June 2008
• Full Storage Based Replication of Open systems in the Secondary
  Data Center –                                        Late Fall 2008

                                                   Oracle Replication (over IP)
                                                    AFS Replication (over IP)

                                                    Gig-Ethernet Campus Link

                                Local LAN Switch                                  Local LAN Switch

                                                                                                     1,2 & 4 Gbs
                  1,2 & 4 Gbs
                                                               4 Gbs ISL
                                                            (Longwave) x 2

                                                                                                     4 Gbs (x16)

                  4Gbs (x16)

                                               Exchange Replication via SRDF/A over FC

                                                                                                                   Secondary Data
Production Data
                                                                                                                    Center (SDC)
 Center (PDC)

            What’s Next

• Fiber Network Assessment
• Replication design
• Configuration
• Testing
• Start replication


To top