AWS Architecture and Networking Proof of Concept

The Challenge

The legacy Quality Assurance workflow for AccessData’s nightly application builds was lengthy and required a number of manual steps. For example, a new deployment might require the QA engineer to copy the 10GB-plus build from a shared drive, copy it across a VPN connection to a set of static Windows EC2 instances, manually click through all of the component installers, and then manually modify configuration files to tie everything together and update the application to point to the database. Once the installation was complete, the QA engineer would then kick off the test suite. The process could take upward of two hours per build, and the Windows instances being static meant that software had to be uninstalled to test newer builds. Moreover, the builds weren’t tested “from scratch” as components would have been installed and therefore might not accurately represent real-world customer installations.

Why Amazon Web Services

As AccessData’s customer-facing product existed on AWS and already supported Auto Scaling and other AWS native services, it made sense to update the QA process to leverage AWS managed services. Additionally, moving the QA process to AWS enabled AccessData to take advantage of automation and CI/CD tools, such as AWS CodePipeline.

“As the first software company to offer a digital forensics platform in a cloud environment, we know firsthand the power of AWS to efficiently deliver professional software tools in the cloud,” said Shon Harris, principal cloud architect for AccessData. “AWS has a very robust infrastructure and product offering that we have not been able to find from any other provider due to their mature product offerings, the support and guidance to make our products successful, and the fact that our customer base is very vocal about preferring to run their forensic and e-discovery workloads in the AWS cloud. This is why we have put a lot of our development efforts and product strategy into our partnership with AWS.”

According to Harris, when AccessData identified the opportunity to update their QA workflow by moving it to AWS, it was clear that the company needed to work with technology experts who had deep experience leveraging cloud technology for quality assurance and with AWS in particular. “1Strategy came highly recommended by Amazon, so we were pleased to collaborate with their team,” Harris said.

The Benefits

1Strategy began by working with the AccessData QA team to create a CloudFormation template, which could accept parameters to create Windows EC2 instances with varying specifications based on use case. The QA team’s tasks included installing prerequisites and the Quin-C™ software and then triggering a collection of automated tests. This configuration and testing process could take a QA engineer two to three hours to complete. Testing was previously done using on-premises hardware. Since AccessData supported two configurations of their product (an All-in-One and a Distributed installation), the CloudFormation template was built to support both scenarios, as well as other customizations.

The use of AWS CloudFormation—built to provision one or many Windows EC2 instances integrated with AWS Systems Manager—enabled AccessData to take advantage of the flexibility and cost savings available to them in the cloud. The template also decreased the time a QA engineer previously dedicated to setup and configuration from as much as two hours to as little as 30 minutes, a 75% decrease. It is now available to the AccessData team for automated deployment, as well as on-demand. For example, just 30 minutes after the CloudFormation template is triggered, a cluster of three Windows EC2 instances can be up and running, having installed prerequisites on each instance (such as Node.js, Git and .NET), provisioned an SQL Server database and installed AccessData’s proprietary software. Since multiple stacks of the template can be spun up simultaneously, a QA engineer can test multiple configurations at one time and not be restricted by available machines as they were with the previous on-premises model.

AccessData has a development cadence of pushing a nightly on-premises Jenkins build. 1Strategy supported AccessData in integrating this nightly build to AWS. Once in AWS, this build was available to integrate into a code pipeline. 1Strategy assisted AccessData in streamlining this process by building a pipeline using Amazon S3, AWS CodeBuild and AWS CodePipeline, which was triggered by a build pushed to master. With this pipeline and CloudFormation template, the whole process is now completely automated.

As a result of the new QA workflow, AccessData has been able to save time and money and enjoy flexibility that was unavailable to them with on-premises infrastructure. Thanks to the outputs of this engagement, AccessData can focus less on provisioning and optimizing infrastructure for QA testing and more on what they do best: developing and delivering world-class digital forensics software.

Finally, the combined team architected and implemented the Windows-based CertMetrics application on top of this new foundation. To support high availability and automatic scaling in response to load, the web servers were implemented as an autoscaling group fronted by an application load balancer. These web servers utilized an FSx file system, which in turn leveraged AWS Managed Microsoft Active Directory to govern access. An EC2-based SQL Server setup using availability groups provided the right blend of support for database customizations, as well as high availability for the database tier. The result was that the CertMetrics app could then realize the benefits of AWS: stability, high availability (HA), automatic scaling and ease of management.

“The 1Strategy team included a number of talented, high-quality engineers who are extremely adept at working with the most advanced cloud technology,” said Harris. “They used a forward-thinking approach to offload the on-premises computing work into a cloud environment, generating significant efficiencies for our IT team and for our customers. This collaboration has helped us better leverage the power of our tools by automating their deployment in the cloud while helping us increase by 3X our software QA processes and automating functions we had never really been able to realize prior to our partnering with 1Strategy.”