Virtualizing Enterprises: Challenges Harrick M. Vin Vice President and Chief Scientist Tata Research Development and Design Centre, Tata Consultancy Services (TCS) January 2010
Dec 18, 2015
Virtualizing Enterprises: ChallengesHarrick M. VinVice President and Chief ScientistTata Research Development and Design Centre, Tata Consultancy Services (TCS)January 2010
January 2010
Enterprise Goals
2
Adaptive
• Adapt easily to different business contexts, regulations, M&A, … • Responsiveness to market and customer needs
Agile
• Rapid introduction of new products / services / business processes
Efficient
• Meet & exceed business and customer expectations (service levels)• Minimize total cost of operations
Enterprise Goals
January 2010
Enterprise Transformations
•Key message:– Many “how to transform” mechanisms … Yet, transformation is
difficult – Challenge: What, when and why ! (It’s all about policies)
• Analytics-led automation
•Key discussion topics:
3
Key Message
Why analytics-led transformations?
How does it work? What can it do?
Virtualization in Enterprise: Example
January 2010
Enterprise IT Transformation Challenges
1• No single optimization mechanism meets all
requirements
2• Many different realizations of each
mechanism• Different implementations are suited for
different environments
3
• Each optimization mechanism has cost, benefits, risks, and side effects• Key challenge: Deciding on what, when and
why? • “One-size-does-not-fit-all” Custom
strategy reqd for each environment• Suitability depends on “as-is” state
4 • Scale and complexity of “as-is” state
4
Why Analytics?
Today’s best practice: Manual, intuition-based strategy derivationImplication: Many implementation “surprises” Many failed transformation programs
January 2010
Analytics-led Transformations
Analyze
Plan
Transform
Prepare
5
Transformation planner; Address the “what”, “when” and “why” questions
Execute the transformation strategy using various “how” mechanisms & processes
Monitor an operational enterprise; data cleansingand preparation for analytics
Analyze “as-is” state; Joint analysis of business process, applications & infrastructureUnderstand operations support
Industry Best Practices
EmergingTechnologies
How Does It Work?
4
2
1 Analytics-led Transformation:
Fact-based & Automated Strategy Derivation
3
January 2010
Virtualization in Enterprises
The pain: High cost of operating enterprise IT
• Large numbers of dedicated, under-utilized servers
Virtualization benefit: Right-sizing of server environment
• Virtualization Consolidation of physical servers Higher utilization
Many “strategy derivation” questions:• What target server platform and configuration should one select?• Which servers should one consolidate together?• What will the target server utilization be?• What, if any, will be the impact on application performance?• In what order should one migrate to the virtualized state?• What will the benefits of server consolidation be? …
6
Virtualization
January 2010
Deriving Enterprise Virtualization Strategy
Many choices
• Many different target server platforms and configurations• Many virtualization technology implementations (w/ cost & applicability
constraints)
Many constraints• Business constraints
• Technology refresh plan and processes (depreciation and refresh intervals)• Data center consolidation strategy and schedule; application retirement• Restrictions on consolidating infrastructure for different business
processes• Technology constraints
• Restrictions on consolidating different architectural and OS platforms• Applications requiring special-purpose hardware
• Usage constraints• Separate development and production environments • Consolidate servers with out-of-phase resource utilization patterns
Requires systematic exploration of the solution space
7
Virtualization
January 2010
iTransform• What does it do?
• How does it do it?
8
VirtualizationDEMO
iTransform: A transformation strategy planner
• What to do? When to do? Why to do?
Example: Right-sizing and placing of servers• The what: • Select target platform(s) and virtualization technologies• Determine the consolidation plan: Technology, business and workload
constraints• The when: Derive a transformation schedule: Minimize risk … maximize rewards• The why: Derive a business case
Input data
• Application & infrastructure inventory, utilization, growth plans & cost (CapEx and OpEx)
Approach: Systematic exploration of the design space
• Codify constraints, benefits and side effects of transformation implementations
January 2010
Beyond Virtualization
From virtualization to clouds …
• Clouds offer different deployment options for virtualized environments• Many different cloud providers
• Different cost-benefit tradeoffs and constraints
… And beyond
• Mainframe, storage, network, desktop, application/tech stack, labor, …
Challenges
• Lots of optimization dimensions and mechanisms• Scale and complexity of as-is state
Research opportunities• Automate as-is state analysis• Formalize transformation strategies: cost, benefits, risks, side effects• Dependencies between transformation strategies• Systematize and automate strategy derivation
9
Concluding Remarks
For more information, contact:Harrick Vin ([email protected])
January 2010
Infrastructure Fact: Scale & Complexity
11
30,000+ servers 5+ OS w/ 20+ versions 10+ database products 2500+ app server instances25+ middleware envr.35+ programming languages200,000+ desktops 10+ petabytes of storage 197,000 e-mail boxes, …
Infrastructure Inventory
4000+ business apps
Average age: 3.5+ years
10-20 years: 190+
20+ years: 60+
Significant overlaps
25+ account opening apps
10+ authentication apps
12+ credit decision engines
15+ BI tools, …
Top-tier bank in the US
The Problem
January 2010
Infrastructure Fact: Scale & Complexity
13
Silo-based understanding Little understanding of the global picture !
Application Dev Teams
CapacityPlanning
OperationSupport
CompliancePlantMgmt
Security
The Problem
For more information, contact:Harrick Vin ([email protected])