Efficiency vs' Availability:

1
Efficiency vs. Availability
Colin Rankine Vice President

• The IT Operations Dilemma Is Tougher Than Ever

2
The IT Operations Challenge 2003-2004

• Reduce costs (even more)
• Staff reductions
• Capital budget reductions
• Software expense management
• Make the IT infrastructure more robust
• Local HA demands
• More aggressive business continuity service
  levels
• In some industries, must comply with strict
  regulatory guidelines
• Support new technologies
• Platforms
• Architectures
• Systems management tools

3
How Are Successful IT Organizations Responding?

• Automation
• Standardization
• Consolidation

4
Automation, Standardization and Consolidation
5
How Automation, Standardization and Consolidation
Impact IT Operations
Efficiency As a Function of Operational Maturity
90
80
70
60
50
Efficiency or Cost Ratio
40
30
20
10
0
MIPS/FTE
Server/FTE
TB/FTE
SW Cost/MIPS
(K/year)
Metric
6
Automation

• Automation provides three primary benefits
• Reduces staffing requirements
• Reduces the opportunity for human error
• Institutionalizes technical and operational
  procedures
• Automation technologies come in different forms
• Scripts
• Packaged automation products (event management,
  storage management, job schedulers)
• Physical
• Systems automation for the distributed systems
  are still less capable, and the market more
  fragmented, but improvements are coming (clearly
  more focus on it)

7
Standardization

• Standardization enables operations cost reduction
  in at least four distinct ways
• Enables greater levels of consolidation
• Reduces skills inventory needed on staff
• Reduces complexity, in monitoring, automation
  and change control
• Enables infrastructure elements to be managed as
  collections, not as discreet instances
• Dont get carried away with it!
• Competitive bid
• One size does not fit all
• Applies to procedures, code levels (change
  management, etc.) not just HW and SW

8
Consolidation

• Consolidation applies to
• Physical site (data center)
• Server, storage, and network
• Logical image consolidation
• Dependency on standardization
• Dont force people and machines into the same
  physical location(s)
• Beware of hidden costs especially software and
  long-term contracts

9
Data Center Provisioning

• How many do I need?
• One or two per region
• Build vs. rent ?
• Internal vs. commercial recovery site?
• Recovery time service levels required?
• Cost
• Is managed space an option?
• Minimum distance between production and recovery
  sites?

10
Trends in Recovery Site Provisioning
Primary Enterprise Recovery-Site Strategies
60
50
40
30
20
10
0
Vendor Hot-Site
Internal
None
Mobile Hot-Site
Reciprocal
May 1999
May 2002
Source Disaster Recovery Journal Surveys
11
Site Separation Practices
lt 10 Miles
13
10 to 50 Miles
gt 100 Miles
45
28
11
50 to 100 Miles
Source Disaster Recovery Journal
12
Delivering on High Availability

• Delivering on availability service levels is much
  more about process than technology
• Operations cant do it alone
• In some geographies, local HA initiatives are
  converging with DR/BC initiatives

13
Findings and Recommendations

• The same techniques that deliver on efficiency
  also deliver on availability
• Be patient, it takes time and commitment to get
  results
• Consolidate first the rest will follow
• Dont feel compelled to put people and machines
  in the same building
• Recovery site distance policy should be
  responsible, not excessive
• If regional threats dont exist, consider
  combining HA and DR/BC initiatives

14
Efficiency vs. Availability
Colin Rankine Vice President

• The IT Operations Dilemma Is Tougher Than Ever