Best Practices to Manage Disruptive Technologies - PMI New … · 2019-02-27 · Best Practices to Manage Disruptive Technologies. Best Practices to Manage Disruptive Technologies

Steve Carter

The 451Group

451 Research, LLC

www.451research.com

PMINJ Chapter

May 5th Symposium 2014

Best Practices to Manage

Disruptive Technologies

Best Practices to Manage Disruptive Technologies

Steve Carter VP 451 Research, LLC

Project Management Demands for IT Industry

• Absolute NEED for PMO dedicated to IT organization

• Disruptive technologies have radically DECREASED speed to market timelines for innovation deployments

• Innovation lifecycle has also radically DECREASED. o Push something out and then we will have time to get it

right later? o Missing your product market requirements for REV. 1

innovation initiatives can lead to a DOA solution

• Global economic downturn of 2008 has moved ROI and associated project management best practices up the decision priority stack

3 05/05/2014

Industry Trends

• Hybrid Solutions – Mix of brick/mortar, COLO and modular o Drivers: Lower per deployment CAPEX spends based upon

ability to add capacity in smaller chunks

• Financial industry was a follower of IT optimization trends

• Utilizing compaction outcomes of IT optimization to enable effective multiple data center active/active availability scenarios

• MTDC market has been overbuilt since 2008 financial crisis. o Strong downward price pressure o Over last two years, general 20% price reduction

year/year.

4 05/05/2014

Disruptive Technology Drivers

Source: Cisco Systems

Average Enterprise Useable Data Storage Capacity Increased 53% in 2013. From 4.9 PB to 7.6 PB Source 451 Research

5 05/05/2014

Impact of Stranded Capacities and $

Overprovision 61%

Capacity Mgmt Tools

Virtualization Load Balancing

26%

No Variability 13%

External Cloud 2%

Outsource 2%

Leverage Disaster Recovery

1%

Hire Contractors 1% Other

5%

How do you currently deal with variable resource requirements due to randomness, time of day, and/or seasonal demand?

Successful PM must incorporate Modeling & Analysis to Save $$

6 05/05/2014

Digital Infrastructure Roadmap™

Data Center Consolidation

Optimization Planning

Everything as a Service

Transformation Planning

Third Party

Cost of Stranded Capacities

Modular Design

Return on Investment

Disaster Recovery

Answering the WHAT? WHY? and HOW? questions

CAPEX

Impact of IT Growth

OPEX

Efficiency

Availability

7 05/05/2014

Business Demand Alignment to Digital Infrastructure

• User Requirements

• Business Cycles

• M & A

Business Demands

• Tiering

• Function

• Growth

• Service Catalog

Application Rationalization

• Physical Refresh

• Logical Ratios

• Utilization

Technology Infrastructure

• Owned/Operated

• Third Party

• Hybrid with Cloud

Facility Infrastructure

• Business Demand Units Defined

• Availability Profiles

• Performance Profiles

Consumption and Allocation

Model

• Application Stacks

• Application Technology Model

Application Catalog

• Technology Stacks

• Facility Stacks

• Orchestration

Capability to Capacity Model

• $ / kW physical

• $ / instance cloud

• Decision criteria

Cost Model

8 05/05/2014

Journey of 5000 Servers

5000

2500

550 250 0

2000

4000

6000

8000

10000

12000

14000

16000

18000

20000

Pre 2006 2006-2010 2010-2012 Post 2012

Space SF

Physical Servers

Power kW

$0

$10

$20

$30

$40

$50

$60

$70

Pre 2006 2006-2010 2010-2012 Post 2012

Mill

ion

s

$ TIER 3 Data Center

$ TIER 3 Data Center

Improving with Virtualization but Still $Millions

9 05/05/2014

Metrics of Digital Infrastructure Capacity Planning – $ and Decisions

Traditional

•Own IT

•Owned/Outsourced Facility

•$ / Watt

Public Cloud

•Outsourced IT and Facility

•No view of kW

•$ / Instance

Capacity Utilization

and Requirements

Tech Cycles

Business Cycles

Orchestration

of Business and

Technology

Demands

Required

Client Example: Business Demand Metrics

10 05/05/2014

Metrics of Traditional vs. Cloud

Type $/kW $/Instance RU’s Required for 256 VM’s

HPDL380 $232 $580 32

HPC7000 $236 $589 10

EC2 On Demand NA $1,594 NA

EC2 Reserved Heavy

Utilization

NA $1,029 NA

Includes data center infrastructure costs (IaaS/compute as a service - baseline) Does not include network, data storage and application costs

Assumption: $200 per kW IT load annual internal data center facilities costs

• Single case study using our current cost comparison methodologies • Large enterprise client with data centers based in US West • Goal: compare traditional internal basic IT infrastructure cost across platform types,

power consumption and workload projections • Secondary goal: compare internal analysis to EC2. • SLA IS different between solutions!

Assumption: 3 Year Life of Hardware/Software Costs (3 Year Cost Model)

11 05/05/2014

Digital Infrastructure Dashboard

“Metrics That Matter” must be set up, facilitated and tracked by the PM & Project Team

1. Monitoring

2. Management

3. Modeling

Monitoring

• Compute: CPU, Memory & I/O

• Storage: Raw vs. Provisioned, Provisioned vs. Utilized

• Network: Bandwidth, Latency, Utilization, Errors

Management

• Capacity vs. Utilization

• Availability & reliability

• Reporting, Alerting & Tracking

Modeling

• On premise vs. Off Premise DIRECT Comparisons

• Future State Transformation (what if scenarios)

• Align DI Requirements to Dynamic Business Demands

12 05/05/2014

Digital Infrastructure Efficiency Metrics

13 05/05/2014

Cost per Transaction with Facility Cost vs. Without

0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Transactions Per DI Dollar with Facilities MEP

Biz Trans per Month per DI Dollar Biz Trans per Month per DI Dollar w/MEP

14 05/05/2014

Digital Infrastructure Analysis Case Study

Financial Client

Case Study Introduction

• Client Roadmap was to have Active/Active environments in Mission Critical Applications

• Initial Analysis Requirements

• 1,100kW of IT load spread over 12,000 SF of Raised Floor • ~$40-$80M Capital Cost

• Client wanted to validate the capacity assumptions with a neutral party, so engaged with 451 Advisors

16 05/05/2014

Project Overview

• Client Requirements

• Understanding Data Center “A” Current State of Digital Infrastructure

• Define the Future State Scenarios

• Create Guidelines for Historical and Future Growth Assumptions

Results by Scenario:

1. Data Center “A” Baseline Model

2. Data Center “A” Cloud Optimization Model

3. Data Center “B” Full Replication Build Model

4. Data Center “B” Mission Critical Replication

• Recommendations/Comments

• Next Steps

• Peer Benchmarking

17 05/05/2014

Data Center “A” Future Growth Model (Optimized State)

18 05/05/2014

Conclusion – Transition

• Initial Analysis Requirements

• 1,100kW of IT load spread over 12,000 ft2 of Raised Floor - ~$40-$80M Capital Cost

• Refined Requirements (451 Advisors)

• 400kW of IT load spread over 3,500 ft2 of Raised Floor - $14-$21M Capital Cost

• Over provisioned facility projections

• 75% Lower on IT KW projections

• 80% Lower on ft2 projections

• 65% - 75% Day 1 Capital Savings

19 05/05/2014

Client Case Study Secondary (DR) Data Center Demand Analysis

Digital Infrastructure Roadmap Case Study

Options Evaluated

• Space and power In ADC cannot accommodate entire Drop Ship List

• Requires alternative site MEP upgrade (possibly expanded space)

Stay in ADC – Drop Ship to new site

• Move Drop Ship to the Cloud

• Incur a nominal monthly fee for test/dev and storage ADC & Drop Ship in the

Cloud

• Utilize existing facility; would be about the same as Greenfield (may be less if facilities space is available & demolition costs are minimal)

• Alternate Client site may not be viable for full ADC build out (not enough space)

Brownfield

• Build new structure on Client Campus – future cost savings long-term

• Risk – two data centers in close proximity Greenfield

• DRaaS for Most of ADC; No HA support

• Would not support all ADC technology (e.g., Tape) Managed Private Cloud

• Price/evaluate other City area Collocation facilities New COLLO

21 05/05/2014

Current 2012 2013 2014 2015 2016

Storage

Server

Network

Model Results: ADC Power In Recovery Mode (with Drop Ship)

Capacity Estimated Production Load with Increased CPU Utilization

22 05/05/2014

Ten Year CAPEX and OPEX Cost Estimates

Notes: ADC + new site includes $500K upgrade & NO increase in OPEX Managed Private Cloud is DR Only Greenfield & COLO includes migration costs

$-

$5,000,000

$10,000,000

$15,000,000

$20,000,000

$25,000,000

1 2 3 4 5 6 7 8 9 10

Ten Year Cumulative Approximation

ADC + Warrendale

ADC + Cloud

Greenfield

Managed Private Cloud

Colo

23 05/05/2014

Deciphering the Digital Infrastructure Journey: An Agile, Automated, Adaptable Infrastructure

24 05/05/2014

Effective Delivery & Demonstrated Value

Metrics Identified Metrics Collected Metrics Correlated

25 05/05/2014