14
Our journey from manual deployment on Data center to container based CD on AWS Meital Morida Cohen & Genady Kofman

Our journey from manual deployment on data center

Embed Size (px)

Citation preview

Page 1: Our journey from manual deployment on data center

Our journey from manual deployment on Data center to container based CD on AWS

Meital Morida Cohen & Genady Kofman

Page 2: Our journey from manual deployment on data center

Who we are

, an Israeli start-up acquired in 2015 by Pitney Bowes.

CompleteTM Cross-Border is a scalable, end to end global ecommerce

solution that can handle all facets of enterprise retail cross-border expansion.

Pitney Bowes is a global technology company offering innovative

products and solutions that enable commerce in the areas of customer

information management, location intelligence, customer engagement,

shipping and mailing, and global ecommerce

Meital Morida CohenSenior Director, R&D

Genady KofmanArchitect, R&D

Page 3: Our journey from manual deployment on data center

Our Agile Practices

Advanced and established agile practices

Kanban – Delivery on demand

Minimum planning and estimations

Focus on delivery of business value

Continuous Integration

Inherent within our SDLC, based on the following principles:

Fast Feedback

High Coverage

Page 4: Our journey from manual deployment on data center

Why we started this journey

4

commit build test deliver

Instant 1-3 Days Days/Weeks Weeks/Months

Dependency!Someone

needs to create a job

Dependency!Someone

needs to provision

resources

Dependency!CM/Resources/Handover

Page 5: Our journey from manual deployment on data center

A lot to solve

Developer productivity

Elasticity

End To End Ownership

Reproducible environments

Common platform

High availability and resilience

Page 6: Our journey from manual deployment on data center

Our Solution

6

A new platform that took the dependencies away

Automate CI pipeline discovery & creation

Automate environment provisioning

Automated Delivery

commit build test deliver

Instant Minutes Minutes

Page 7: Our journey from manual deployment on data center

Innovation made the difference!

Cloud based solution

Infrastructure as a code

Fully reproducible

Elasticity

High Availability

Containerized infrastructure/application tier

Layered Docker Images

Jenkins Single pipeline

Scriptable pipeline, effectively making jobs definitions part of the application code

Convention based

ECS as Cluster management

Pets vs. Cattle

Auto scaling/ Auto healing

Page 8: Our journey from manual deployment on data center

A new way to run software

Page 9: Our journey from manual deployment on data center

A new development and delivery approach

Jenkins at the center of full software development lifecycle

Scans organization VCS to discover new projects/branches. Jobs are created for

repositories that include Jenkins pipeline scripts

All projects follow same convention: each repo covers full lifecycle, from build to deployment

It includes: Jenkinsfile (CD pipeline definition), Build scripts, Dockerfile, Deployment scripts

PB abstraction of ECS service to simplify and generalize deployment, using internal tool

ECS Tools – internal tool to save the complexity of ECS cluster

Page 10: Our journey from manual deployment on data center

Jenkins Pipeline

Page 11: Our journey from manual deployment on data center

Jenkins Pipeline

Page 12: Our journey from manual deployment on data center

Where we are today

Business Value

Fully implement a DevOps Culture

Accelerate Development and Deployment to be more rapid ,frequent and reliable

Innovation within the teams

Architecture Flexibility - shift to Micro Services

Still ahead…

Complete reproducible environment, including DB and management tiers

Handover process

Cloud vendor independence

Page 13: Our journey from manual deployment on data center

What we’ve learned

Executive backing - time and money

Resource Allocation – hiring dedicated strong people

R&D Skills – training engineers to mitigate the gap of Knowledge

Supporting organizational structure - Dev/Ops/Tech-Support

Standardizing company wide – “spreading” the platform among other

groups in the organization

Page 14: Our journey from manual deployment on data center

Meital Morida Cohen

Senior Director, R&D

[email protected]

Genady Kofman

Architect, R&D

[email protected]