Pushing the Limits of Scalability

When a major Australian university had to upgrade its BYOD platform from remote desktop server to cloud instances to meet growing user demand and accessibility, Envisian Cloud expertise showed how the upgrade could be leveraged into a long-term business changing opportunity.

We proposed to make the BYOD connectivity so attractive that it would replace the need for on-site computer labs – and wrote the code to deliver.

The University was already struggling to construct new student labs fast enough to meet demand and had voiced concerns that this might undermine its ability to offer courses.  Envisian’s virtualisation and cloud architect, working with the University as part of its enterprise innovation team, did a cost/benefit analysis.  He compared the per user cost of owning computer labs, including the sunk and future real estate investment, against the per user cost of implementing a new cloud-based BYOD platform. The projected four-year comparison, based on anticipated enrolments/users, not only revealed the scale of potential savings, but also identified opportunities for new courses and modes of teaching. Approval was given to implement this ‘bold’ solution as quickly as possible.

The challenge would be in finding the optimal cloud-based solution – one that would deliver the benefits of increased capacity, high degree of concurrency and improved accessibility without increasing support requirements and costs.  The true capability of a commercial off-the-shelf (COTS) product is severely tested when you deploy it to meet the scale of users found at Australian universities.They are some of the biggest in the world. Our client had over 140,000 user accounts (staff and students).

It soon became clear that none of the COTS solutions were built for the required hyperscale-sized cloud deployments (except AWS Workspaces, which was deemed too costly) and none provided an out-of-the-box functionality to allow major changes, such as patching of servers, without impacting users.

Our consultant proposed that he could write the code for a life-cycle automation layer that would hyperscale, as needed, and refresh/repair the server pool without impacting users.  He chose Citrix on AWS Cloud because it best met the University’s objectives and because its technology allowed the platform to turn resource off or on, based on scheduled and instantaneous demand.

Over two weeks in December, he built the Minimum Viable Product (MVP) and wrote the code for the automation layer to the Citrix application, without impacting established user sessions.  This was enough to demonstrate that the new system could provide all users access to some basic applications straight away. Following an online pilot in January 2018, the platform was sufficiently robust to offer to the rest of the University by the start of the academic year in March.  Our consultant continued to work to improve and stabilise the offering and ensure the platform was ready for BAU handover by August.

Capital investment in new structures for computer labs has been halted and plans are in place to re-purpose space from existing labs. User experience has improved dramatically with the new features, such as higher frame rate, audio, dynamic categories, and the ability to adapt to user’s internet quality.  The more responsive and feature-rich service, and its ability to seamlessly access hosted Windows and Linux applications from any device, prompted a re-think of some courses and the introduction of new ones. In the process, vendors were persuaded to change their licensing agreement to support the University’s new delivery model, which in turn led to them to rethink how they work with large tertiary organisations, in general.