McGraw Hill accelerates its AMI build process with a cloud native solution on AWS

How we helped

The sheer size of more than 190 AWS accounts required the company to bring in an expert partner like Rackspace Technology to help manage costs with its Rackspace Optimizer+ solution and to provide a dedicated team that could optimize McGraw Hill’s cloud architecture and infrastructure.

“Many challenges accompany the management of 190 accounts,” said Koegler. “McGraw Hill has thousands of EC2 images. We have a heterogeneous development environment with different types of solutions, including Windows, Linux and variations of Linux. We sought a secure way to encapsulate all of this through an automation pipeline — without having to manually build the baseline for our AWS images.” 

While the existing process had served McGraw Hill well over the years, the Rackspace Elastic Engineering team sought an opportunity to create a cloud native solution that would address three goals:

• Increase automation and improve resiliency
• Broaden existing security and compliance measures
• Enable the reusability of build components

Because McGraw Hill’s legacy AMI process supported multiple core base AMIs, including CentOS, AWS, Linux and Windows, it experienced additional overhead costs stemming from complex governance processes that were required to generate alignment with different standards.

The current process lacked automation. Each step implemented various automated but separate mechanisms, such as building AMIs with CircleCI. McGraw Hill then used Jenkins for image distribution, which doesn’t easily integrate for a full end-to-end solution. Moving the build process to AWS’s EC2 Image Builder would allow for more seamless integration points with existing AWS services.

Rackspace Elastic Engineering started out by helping to manage many application-specific AMIs, but quickly realized there was a way to automate and standardize the build process. EC2 Image Builder’s component-based model offered a perfect opportunity to minimize duplication while supporting the development of an automated pipeline for all AMI builds.

The team determined that replacing core AMI build processes with EC2 Image Builder would provide and ideal platform on which to standardize the creation of custom AMIs for McGraw Hill.

The Rackspace Elastic Engineering solution features several key improvements, including:

• A decoupled architecture that supports asynchronous components and retry mechanisms
• Improved governance and security controls, including AMI encryption enforcement
• Automated security scanning integration that uses Okta to restrict approvals to only authorized individuals
• Added reusable components to support various architectures and versions of operating systems
• Automated Slack notifications with a custom Slack app that supports user interactions

For the Process Deep Dive, check out the technical case study here.

The Rackspace Elastic Engineering team also enlisted AWS services to boost compliance controls. The solution reduced the complexity of securing and managing multiple third-party components by keeping as much of the overall process on AWS as possible.

Building a new AMI pipeline used to be a process that took from two to six months. This cycle has been shortened to less than two weeks. This means that McGraw Hill’s developers can focus their time and efforts on growth priorities and creating applications that are reliable and secure for users.

“With Rackspace Technology, we were able to automate the EC2 images for each of the different platforms that we use for the various markets we serve, whether it’s PreK-12 education, higher education or professional development or continuing learner markets,” said Koegler. “Together, we were able to come up with an automated solution that secured the pipeline end-to-end, accelerated time to market and enabled development teams to focus on education problems for our learners and educators.”