Caspar Herzberg |
This article was originally published in the Autumn 2018 issue of The Record.
Applications and infrastructure rely on a combination of desktop and server-based software, as well as cloud services. As a provider of hands-on learning solutions, we turned to Microsoft Azure to help us transform our Lab on Demand engine, which is responsible for automatically building and deploying millions of lab environments, into a tool that could deploy environments that spanned the data centre and the cloud.
The transformation began with changing the very definition of what a learning environment is. Traditionally, these were limited to a small set of virtual machines that could be cloned and deployed for individual students. To add cloud to this, we extended this definition to include things such as Azure compute virtual machines and much more. Instead of using networks and switches for isolation, we relied on cloud concepts such as resource groups and subscriptions.
What we underestimated, however, was the management of customer expectations and communication. As more labs were designed and deployed on cloud fabric, we found we had significantly less control over the experience of the student. Additionally, we needed to create scenarios that addressed the perception of problems where, in fact, none currently existed.
Our entire organisation underwent a fundamental change in how we approached building and deploying learning products as well as how we support customers. New roles and practices emerged. For example, we defined a role specifically dedicated to monitoring for upcoming change, analysing the impacts not just on our own software and infrastructure, but on the learning products people build on top of those.
Though these changes were challenging, the end result is that we are now far more agile and flexible than ever before. For example, we are no longer limited by data centre hardware; we can deploy labs to the scale of the cloud, leveraging Microsoft Azure to instantly expand our footprint. We are no longer limited to virtual machines as learning environments; we can provision lighter, faster, and cheaper environments consisting of containers, resource groups, or subscriptions across services such as Azure. Azure IaaS enabled us to add additional layers of redundancy and failover and solve some aging hardware issues without investment in additional metal.
After substantial transformation of our technology suite, our platform, our operations and support team, and our general approach to building learning environments, we have emerged with the ability to do more, in less time, and at a scale that was not previously possible.
Corey Hynes is CEO at Learn on Demand Systems