Red Hat Consulting Discovery Session Series: Migrating Workloads to - - PowerPoint PPT Presentation

Red Hat Consulting Discovery Session Series: Migrating Workloads to Containers

Eric Sauer Architect Red Hat Consulting

NA PaaS/DevOps Practice

1

First, Some Intros

What you should know about me...

Eric Sauer (like Sauerkraut)

• 6 years at Red Hat (all in Consulting)
• My job is to improve how Red Hat delivers services engagements around Containers

and OpenShift Container Platform

What I assume about you...

• You’ve heard of containers
• You believe (or are exploring the possibility) that containers could improve the speed

and reliability with which you deploy and scale applications

What I want to know about you...

What kinds of applications are you looking at migrating?

• Languages?
• Platforms?
• Architectures?

CONTAINER PLATFORM SOLUTIONS FROM RED HAT CONSULTING

6

7

MODERNIZE APPLICATION DELIVERY WITH CONTAINER PLATFORMS

Infrastructure DevOps Applications Program increment 0

Smart Start Discover Design

Program increment 1

Groundwork Discover Design Deploy

Program increment 2

Automation Discover Design Deploy

Program increment 3

Rollout Discover Design Deploy

Operationalizing container platforms Discovery Session Container Platform Smart Start Container Platform Design Workshop Continuous delivery Container- driven continuous delivery Migrating workloads to containers Container adoption for application development Microservice development for container platforms

MIGRATING WORKLOADS TO CONTAINERS

Environments:

• Existing pre-production environment

Tasks:

• Migrate traditional applications into containers
• Analyze applications to determine fitness for containers

and distributed architecture

• Containerize existing application platforms
• Analyze migration to determine level of effort
• Mentor on how to redesign (or not redesign) traditional

applications to run in OpenShift

• Decompose and migrate existing applications to

microservices architecture in a controlled, iterative manner

• Onboard migrated applications onto container platform

MIGRATING WORKLOADS TO CONTAINERS

Build Migration Case Determine Fit and Priority Plan Migration Execute Migration

We help calculate and analyze ROI based on: Technical:

• Application scalability
• Auto deployment
• Infrastructure agnostic

Business:

• Consolidation
• Flexibility
• Speed to market

We help prioritize migrations based on: Portability:

• Ease of switching servers
• Ease of switching cloud

providers Scalability:

• Ability to run multiple

instances Active Development:

• Ability to make changes

We help plan migration activity: Change code:

• Required: fix incompatibilities
• Value add: architect for cloud

Automate build:

• Required: S2I/Docker script
• Recommended: CI/CD

Customize image:

• Required: support

development languages and libraries

• Value add: support

shared libraries and standardized configuration We help execute migrations through an iterative process...

• 1. Perform all required tasks
• 2. Perform all recommended

tasks

• 3. Assess and perform

“value add” tasks individually based on need …and emphasize continuous improvement to drive down incremental migration costs

• ver time.

Let’s look at some migration scenarios

10

MIGRATING WORKLOADS TO CONTAINERS

Scenario #1: Web Scale Enabled Application

• Is not tightly coupled to its current server environment

○ I customize my application for its environment through environment variables or properties files

• Doesn’t care how many instances are running

○ I can run 2, I can run 10. Capabilities and behavior are the same

• State is loosely coupled from my core runtime

○ I have no state, OR ○ My state can be offloaded to an external cache, OR ○ My state is written to disk

• Deployment process is automated, or close to

MIGRATING WORKLOADS TO CONTAINERS

Scenario #2: Traditional Web App

• Can run on another server, but I need to make some changes and rebuild to do so

○ Would need to remove hard-coded references to things in my environment

• Can scale out in theory, but i’ve never tried

○ Currently run on multiple instances (3 or more)

• Manage state by some form of internal clustering
• Installed manually, but that work could be scripted

MIGRATING WORKLOADS TO CONTAINERS

Scenario #3: Legacy and Monoliths

• My app depends on system libraries that may or may not exist in a new server environment

○ Runs CLI Commands

• Designed under the assumption of a static number of instances
• Requires me to add CPU/Memory in order to handle more traffic

○ Containers are better suited scale horizontally than vertically

• My app performs many functions and takes up a very large footprint

What’s next?

14

THANK YOU

plus.google.com/+RedHat linkedin.com/company/red-hat youtube.com/user/RedHatVideos facebook.com/redhatinc twitter.com/RedHatNews

15