Mission Data Storage Consolidation MDSC

1
Mission Data Storage Consolidation (MDSC)

• Ray Hunnicutt
• David Garza
• Lockheed Martin

2
AGENDA

• Problem Statement
• Background
• Goals
• System Architecture
• Benefits
• Future

3
PROBLEM STATEMENT

• Mission Control Center (MCC) at the NASA Johnson
  Space Center (JSC) has a multitude of unique
  storage systems implemented with a variety of
  hardware and software elements. The physical
  distribution of these and their lack of
  commonality results in cost inefficiencies with
  respect to storage operations, maintenance, and
  sustaining.

4
BACKGROUND

• Equipment Maintenance
• Some equipment is over 10 years old
• Difficult to maintain
• Islands of Distributed Storage
• Storage is isolated to each system and excess
  storage cannot be shared
• System is not expandable to meet the current
  storage requirements or any future expansion
• Cost of Distributed Storage
• Labor intensive management
• Control and monitoring is not centralized
• Semi-manual backup process
• Multitude of COTS H/W and S/W Products
• Custom S/W
• Does not meet restore requirements (4 hours)

5
SYSTEM ARCHITECTURE CURRENT
6
GOALS

• Consolidate 2 Separate Systems onto a Single
  System (servers and storage)
• Must meet reliability and availability
  requirements
• Must meet failover requirements (within 60
  seconds)
• Must meet disaster recovery requirements (restore
  to operations within 4 hours)
• Automation of Storage Functions
• Backups defined once and then managed by storage
  software
• Separation of storage traffic (e.g. backups) from
  the OPS traffic
• Provide Initial Infrastructure
• Allows future data storage consolidation
• Meets new and growing storage needs
• Utilize open standards where appropriate
• Use commercially available products
• Reduce System Total Cost of Ownership (TCO)
• Consolidate operations personnel
• Reduce sustaining engineering cost

7
SYSTEM ARCHITECTURE NEW
8
SYSTEM ARCHITECTURE NEW (contd)

• Servers (IBM pSeries 6H1 Servers with 6-way
  processors)
• 3 servers operate as a cluster
• Provide NFS services
• Connected to Storage Area Network (SAN) via Host
  Bus Adapters (HBAs)
• SAN Fabric (IBM 16 port Fibre Channel Switches)
• 2 Fibre Channel Switches
• On-line Storage (IBM Enterprise Storage System,
  a.k.a. Shark)
• Prime and Backup Disk Array Storage Units
• Backup System
• 1 Backup Server (IBM pSeries 6H1 Server with
  4-way processor)
• 1 Linear Tape Open (LTO) Library (IBM 3583 with 6
  drives and 72 cartridge capacity)

9
SYSTEM ARCHITECTURE NEW (contd)

• System Management
• 1 Management Workstation (IBM Intellistation PC)
• Provides centralized management
• Out-of-Band Management Network Cisco Ethernet
  Switches
• Private Ethernet network
• Connects Management Workstation to each storage
  device via Ethernet switch
• COTS Software products
• IBM HACMP to provide clustering and failover on
  the servers
• Tivoli Storage Manager (TSM) to provide automated
  backups
• IBM FlashCopy to create snapshots of data
• IBM Total Storage Expert to mange the entire
  storage system
• Open Standards
• Fibre Channel
• LTO

10
ON-LINE STORAGE HIGH AVAILABILITY

• 2 Storage Volumes on each Disk Array Storage Unit
  to meet High-Availability Requirements
• OPS
• Nominal Ops storage
• Disaster Recovery (DR)
• Point-in-time copy of data for backup and
  recovery situations
• Snapshots copied from OPS to DR
• Backups copied from DR to LTO
• System produces two sets of tapes (on-site and
  off-site storage)
• DR volume is first level of disaster recovery
• Details
• OPS volumes on Prime and Backup Disk Arrays
  mirror each other
• All volumes are of equal size
• Snapshots allow data to be quickly copied from
  one volume to another

11
BENEFITS

• Turn-key COTS solution
• Maintenance and spares provided by the Vendor for
  the life of contract
• Reduced Sustaining and Operations Cost
• Consolidates storage for multiple systems onto
  same devices
• Lower hardware and software maintenance costs
• Lower engineering costs
• Reduces number of operators through consolidation
• Lower sustaining labor costs
• Return-on-Investment (ROI)
• Cost savings achieved in under 3 years
• New system provides much more capability than old
  system

12
BENEFITS (contd)

• Open Standards
• Defines standards and commonality
• S/W Products
• H/W Products
• Improved Backup and Disaster Recovery
• Reduces backup windows
• Provides LAN-free backups
• Reduces time required to restore system (meets 4
  hour time requirement)
• Expandability to meet future needs and include
  more systems
• Expandable disk capacity
• Defined technology growth path for LTO

13
FUTURE DEVELOPMENT

• Off-site disaster recovery unit connected to SAN
• Eliminates the need for moving tapes from on-site
  to off-site tape storage each week
• Server-less backups
• Reduces the workload on servers for the backup
  process
• Thin Servers
• Reduces the cost for future procurements
• Hierarchical Storage Management (HSM)
• Allows data to be automatically migrated from
  disk to tape and back to disk
• Archive Functionality
• Allows data to be moved to media for permanent
  archiving to be recalled at a later date