Journey to the Programmable Data Center

Journey to the Programmable Data Center Modernizing our Data Centers – Strategy

and ApproachHistory, Perspective, Trends

Toby Weiss, Sr. Director, National Infrastructure Practice, Visionary Integration Professionals

1Consulting | Technology | Outsourcing

Data Center Modernization Perspective

IT

Business

Real Estate

CFO

Image Sourced from Deepa Krishnan

The Hammer and the Nail

Small Business • Optimizing the workload already in house; • Building departments of expertise (that eventually lead to

silos)• Partnerships: likely focused on a single vendors solution• Leveraging Cloud for Applications and Infrastructure

Med Enterprise• Mastering centers of expertise into departments (silos)• Evolving from older platforms that started our business to

more robust and scalable IT• Beginning to manage technical debt• More reliance on multi-vendor environments

Larger enterprise• Breaking through the silos for lean efficiencies• Technology standardization• and adoption of converged infrastructure/cloud computing

Examples of Context:One size does not fit all:

Business sizeorganization needsapplication workloadsCulture, leadership, politics

Macro and industry economic trendsGrowth versus maintenanceIT budgets

Technology drivers like virtualization & converged infrastructureRates of adoptionROI for our unique business needs

“Computers in the future may have only 1,000

vacuum tubes and perhaps only weigh 1 1/2 tons.”

- Popular Mechanics, 1949

What Would You Predict?

The nearly 18,000 vacuum tubes and 6,000 switches of the ENIAC, the first electronic computer. Image Sourced from ExplorePAhistory.com

Evolution of Computers

Image Sourced from Computer History Museum

99.6% of our history without Computers!

First mechanical computer – difference engine designed 1837.

Today we each have a “smartphone”.

Evolution of Data Centers - What is Modernization?

Is the utility compute infrastructure our future?What is relevant to enterprise IT?

Industrialized Data Centers and IT Infrastructure

StandardizationModularityScalabilityGeo-redundancyAutomationOrchestration

1) Advances in Physical Technology

Enablers in Data Center Modernization

8

a) Shrinking Hardware Drive Lower Costs With More Power

b) Virtualized functions Converged Infrastructure Appliances

Two

Maj

or T

rend

s

2) Increasing Abstraction from Hardware

Driv

ing

DC

Mod

erni

zatio

n

c) Virtualization of Hardware Evolved de-facto & industry standards hypervisor interoperability workload variety

d) OS Overlay Concept GRID GoogleFS/Facebook Hadoop NextGen?

Historically, Two Approaches to Abstraction from the IT Infrastructure Layer

9

Driving freedom from hardware dependencies and pooling resources for greater utilization.

Top Down Bottoms Up

Workload (application)

Geographically dispersed physical compute assets.• Federation for aggregated capacity• Loosely Coupled with middleware libraries• Type of Parallel computing with common

networks as the “backplane”• Focused on specific large data set and processing

intensive workloads (scientific traditionally).

Compute (server) Asset Base

VM VM …

Small WorkloadsLarge Workload

Relatively centralized compute assets.• Un-federated physical capacity to divide and share• More tightly coupled due to domain of server, rack,

and data center• Higher speed server and data center network as the

“backplane”• Can support a wider variety of workloads.

Programmable or Software Defined Data Center… Simply Defined

• Complete Abstraction of Physical Assets• Hypervisors (software)

Control All Assets, not just servers• Top Down and Bottoms

up abstraction of assets• Lowers I&O costs and

increases agility

Compute Assets

VM VM …

Variety of Workloads

NIST Definition of Cloud• Essential Characteristics

• On-demand self service• Broad network access• Resource Pooling• Rapid Elasticity• Measured Service

• Service Models• SaaS, PaaS, IaaS

• Deployment Models• Private, Community, Public,

Hybrid

Characteristics of DC Modernization• Enables and Leverages Cloud

Computing• Abstracts workloads from

infrastructure• Enables programmability of

Virtualized Infrastructure• Modular, Scalable, Geo-dispersed• Embraces de-facto and industry

standards• Leverages new operational

process, new organizations, new roles, new skills

Char

acte

ristic

s of

Clo

ud

Characteristics of DC

modernization

This is not just about Cloud

What Would You Predict?“I think there is a world market for maybe five computers.”

- IBM Chairman Thomas Watson, 1943

The Pace of Change

Abstraction Continues to Enable IT Virtualization

Virtual Infrastructure Layer

Control Software Layer

Physical Infrastructure & OS

Applications

Business Process

Server Machines Storage Systems Networking/ Security Functions

GRID Middleware – OS Overlay

Quick Review of Recent History

Mainframe• Monolithic• Virtualized• Single Pane of Glass• Sized for Peaks• Centralized Control

Client Server• Lower Costs*• Spurs Sprawl• Processing power

distributes• Control moves to the

LOBs• Enables Growth via

LOB Agility• Creates Isolated

Islands of Storage• Management of de-

centralized computers

Consolidation• Virtualization drives

consolidation• Enables Virtual

Sprawl• Management

Opportunity grows• SANs simplify storage

via snapshots, dedup, etc.

• Modular approach evolves with DC Containers / PODS

• Data centers becoming more complex

Automation & Converged Inf.• 1st generation

solutions to simplify the controlled sprawl – HP Utility Data Center and Egenera

• Automation and management tools mature

• Hypervisors concentrate servers, SANs become bottlenecks

Cloud • Converged

Infrastructure at foundation but still expensive entry point.

• Automation evolves to Orchestration

• Notion of IT-as-a-Service

• Early in the hype • Next Generation of

Converged infrastructure and Software Defined Data Centers Evolves

Increasing Abstraction between the Human and Machine

What is different today?

Standardized Infrastructure

Evolution of Converged Infrastructure

Software Defined Networks

Application Abstraction

Programmable Data Centers Defined by Software

Facebook and others show us Vanity Free Servers - Opencompute.org evolves. Hypervisor

Interoperability.

Lessons learned with control of hardware. DevOps, people, skills, IT organizing differently. Server

virtualization matures.

The Last Mile: increased flexibility to deploy networks virtually via NFV and SDN. Control Plane

& Data Plane concept evolve.

Evolution of top down Abstraction - OS Overlay concept: GRID, Hadoop, GoogleFS, etc.

Evolve architectures to simplified, modular, highly scalable designs

Appliances that simplify & consolidate hardware stacks

Local Storage Controller

Example 1

Typical Storage Stack Deployment Replaced with a local controller and Cloud Storage Service

Images Sourced from Nasuni

Evolve architectures to simplified, modular, highly scalable designs

Modular building blocks replace traditional hub & spoke designs (the Google and Facebook file system concepts)

Layer of Abstraction Decouples Storage

Example 2

Images Sourced from Nutanix

19

Software Defined Data Center Layer (Virtual Infrastructure or Data Center)

Vendors’ Product: Management, Automation, Orchestration of Physical Assets

Applications

Business Process

Standard Servers Standard Storage Systems Standard Network Components

IDS/IPSSAN NAS/DAS

Example 3

Evolve architectures to simplified, modular, highly scalable designs

Software Defined Data Center Solutions: Data Center Hypervisors

Management Management Management

Elements of Today’s DC Modernization Options

DevOpsManaged DemandIT Maturity in People and Process

Abstraction from infrastructure landscapeIncreased virtualization of servers, storage, and networking

Standardization of Infrastructure HardwareShared Resource Pools of servers, storage, networks

Converged Infrastructure Point SolutionsCloud-first or “Greenfield” strategies

Automated Deployment for legacy landscapesSoftware Defined Data Center Solutions

Cloud Deployment Models IT-as-a-ServiceConsume it, don’t build it yourself

What Would You Predict?

“640K ought to be enough for anybody.”

Microsoft Chairman Bill Gates,1981

The Journey for Each of Us is Different

Multi-Year Strategic Planning to Include SDDC

• Integrate DC Modernization Strategies with Application Portfolio Planning• Understand the pace of change for your company and look for entry points to next generation

programmable DC solutions• Drive out tactical efforts incrementally to lead to a transformation for the larger multi-year picture• Leverage an IT maturity model from your favorite vendor.

201520142013

201120102009

2012

201820172016

% SDDC Deployed

SDDC Pilots

Last Inf. Refresh

Balancing Tactical Efforts with Strategic Transformation

Transformational ChangesFocus on:

“Doing things differently in a big way”

Technology & Process ImprovementFocus on:“Doing things incrementally better and with less cost”

Pure Strategic

Pure Tactical

Finding the right balance

Example Planning Framework for Supporting Transform to SDDC

Business Strategy

Technology Strategy(Applications, Architecture,And infrastructure.)

Management of Change

Process Integration

Organization & People

Future SDDC Operations (FMO)Current Operations (CMO)Planning Layers Interim Mode Operations (IMO)

Ongoing Governance

Business Case

Management & Orchestration tools

Business and IT process discovery

Readiness Assessment Key Announcements and Communications

POC Design

Communications Planning

MOC Roadmap

Governance Plan

Discovery: Current Technology &

Workloads

Business strategyImplement Governance

Mechanisms

Define Services, Contracts and SLAs

Define new ITaaS process and

integration points to existing processes

ITaaS Org Design

IMO resource transition

FMO resource transition

IMO resource training

Skill AssessmentsSteady State Operations – adds, changes, decommissions, capacity expansions, etc.

Adopt/modify Operations Guide

IMO Team leads Workload Migration

Ongoing Refresh and Maintenance

Test Operations Guide

Updates to Operations Guide

Implement Service Catalog Portal

POC Buildout

Publish Service Catalog

Plan Outsourcing Transition

Transition to External Service

Provider/Operator

Full Buildout Migrate Application Workload

Line of Business and General IT Education/Training

Journey to Data Center Modernization

Alignment across layers

Don’t forget the human aspects

of change

Top down usually drives

success

Look for Greenfields

Best Practices

Summary of Approach to Modernize Data Centers

Embrace Multi-year strategic

planning for I&O

Big Picture on the Puzzle may be a common Vision

Balance Incremental with Transformational

Change

Align Pace of Change across layers in your

strategic planning

201520142013201120102009

2012

201820172016

Parting Thoughts

• IT infrastructure may become like a utility, but it is a long way off. Infrastructure and Operations still relevant today!• Embrace both on and off premise solutions. DC modernization is also

about leveraging someone else's modernized data center.• Explore emerging technologies leveraging the ingenuity of the Web

2.0 companies. They may not be right for all of us today, but when will it be right? Can you predict?• DC modernization should support your enterprise cloud strategy.

THANK YOU

Toby [email protected](503) 926-4710

28Consulting | Technology | Outsourcing