Migrating Workloads to AWS – The DevOps Way

close up image of black servers


FlatWorld, a digital textbook publisher, provides a unified and cohesive user experience for instructors to
adopt, customize, and teach from more than 100 high-quality textbooks.

FlatWorld’s digital textbook platform is a set of custom-built services using a variety of open- source
technologies such as Drupal, Python, Ruby on Rails, and Sinatra. The infrastructure was hosted through
a managed data center provider, but costs were high and it was cumbersome to operate and maintain.
FlatWorld decided that migrating to a commercial cloud would reduce operational costs, deliver new features
and capabilities, and speed products to market. They selected Amazon Web Services (AWS) as their preferred
platform. However, significant challenges remained in executing the complex migration, including:

  • No Downtime: Students use FlatWorld every single day and contracts with institutions demand high
    availability, so the migration had to take place seamlessly without downtime.
  • Short Timeframe: The migration had to be complete before the start of the school year when usage and
    business peaks.
  • Retain Legacy System: Needed to operate the legacy and new cloud system in parallel to mitigate risk and
    assure confidence in the new services.
  • Constrained Resources and Budget: The available resources and budget for performing the cloud
    migration was limited.
  • Adding additional capabilities: Increased service level requirements were needed for the new system.


eGT employed an infrastructure-as-software approach in executing the entire system and data migration,
implementing all cloud provisioning, application configuration and orchestration tasks as a repeatable
software-driven process. We took full advantage of AWS programming interfaces and capabilities using our
Cloudamatic® tool, a 100% open-source cloud-orchestration framework. Cloudamatic is a core component of
eGT’s DevOps Factory® service.

eGT collaborated with FlatWorld to establish a consistent and common understanding of the business
requirements and timelines. Together we developed an overall cloud migration plan and schedule and
produced the target cloud architecture. We iteratively produced the necessary work products to facilitate
rapid, pain-free migration of FlatWorld services to AWS.

Test Environment
The first task was to establish a working test environment of FlatWorld’s digital textbook platform in AWS.
During the test deploy, Cloudamatic was used to define and automate a complex cloud software deployment.
Since the client requested AWS CloudFormation templates as a required output format, we saved time by
using Cloudamatic capabilities to machine generate the CloudFormation templates, which saved precious
labor hours.

Parallel Production Mirror
Our experts established a parallel production system of FlatWorld services in AWS to mitigate any risk of
disruption during FlatWorld’s peak business season. The parallel production system operated as a secondary
system and mirrored real-time data and transactions processed by the legacy production environment.
Parallel operations allowed for controlled testing of the FlatWorld AWS version in production, while
maintaining full business continuity and preparing for full production cut-over to be executed at the time and
schedule of our client’s choice.

Production Cut-Over
After completing production testing, we shifted the focus towards flipping the legacy data center-hosted
environment to secondary status, eventually moving to a full cutover to the AWS-hosted FlatWorld
infrastructure. After a month of uninterrupted operations, we decommissioned the data center instance and
completely switched over to an AWS-hosted FlatWorld digital textbook platform.


Established a functional test environment in less than 6 weeks and operationalized the production mirror
within 3 months.

  • Achieved a Continuous Delivery and Secure DevOps operating model through Cloudamatic.
  • Improved service-level agreement performance and system availability to over 99.9%.
  • Saved over 70% in operational costs by migrating to AWS.
  • Empowered team to manage at scale performing massive upgrades, patching, and deployments with zero
    downtime after migrating to AWS and adopting Cloudamatic.

“Our relationship with
eGT has allowed us to completely overhaul and modernize our web infrastructure
over the last year. Their tools, knowledge, and experience gave us the confidence to
begin new projects and their support as partners helped us implement and complete them, which
has significantly reduced our operating costs while allowing us to ship new features faster.”
Keith Fahlgren, VP of Product and Engineering, FlatWorld


Copyright 2018 | eGlobalTech | All rights reserved.