all case studies

Arup: Paving the way for cloud-based engineering modeling

At a glance

Arup wanted to move its Windows desktop-based engineering simulation software, Oasys GSA, to the cloud. Working side by side with Oasys engineers, Slalom helped configure and run experiments on three cloud database solutions to see how each handled the large datasets produced by GSA.

What we did

  • Cloud transformation
  • AWS implementation
  • DevOps

A legacy of leadership

What happens when you cross the practical vision of a structural engineer with the abstract thoughtfulness of a moral philosopher?

Sir Ove Arup and the Arup Group Limited is what happens.

Established in 1946, Arup is a firm of designers, planners, engineers, consultants, and technical specialists who work across every aspect of the modern building environment. Today, Arup employs 14,000 people in more than 34 countries and continues to be fueled by the vision of its founder to shape a better world through its work with clients.

In 1976, Arup launched Oasys Ltd, the company’s software arm. Today, Oasys is recognized as a leading commercial developer of simulation software for structural and geotechnical engineering and pedestrian modeling.

Cloud rationale

Oasys initiated a journey to take GSA, its structural engineering simulation program, from the desktop to the cloud for a variety of reasons. First, with the rise of cloud-computing, today’s customers expect to be able to access and use software and applications anywhere. Shifting GSA to the cloud would allow for on-the-go access. Oasys was also hopeful that the shift to the cloud would help enhance performance and flexibility, while allowing the company to explore a subscription-based software business model.

There was just one problem: The Oasys team didn’t have experience with cloud computing. They needed backup. That’s when Slalom joined the project.

Analyzing the software

The re-platforming project unfolded in two distinct phases. First, the Slalom team worked with Oasys engineers to understand the ins and outs of how GSA was structured, how it functioned, and how different parts of the software related to one another.

In GSA, a user creates or imports a model that describes a structure (such as a building or a bridge) in terms of nodes and elements representing beams, slabs, and other structural components. Then the user assigns a loading scenario to the structure, like wind-induced loading or earthquake loading. GSA performs an analysis and produces a visualization predicting the response of the structure to the load scenarios.

Slalom's cloud expertise, combined with our engineering knowledge, allowed us to get things configured quickly. We could see how different areas of the software would communicate between each other, which was new to us.

Experiments with cloud databases

Between analysis and visualization, GSA creates and requires the fast retrieval of a vast amount of data. That’s why the second phase of the re-platforming project was so critical. The team ran extensive experiments on three cloud databases – AWS RDS Postgres, MongoDB Atlas, and TileDB – to find out which would most efficiently and effectively analyze, store, and retrieve the large datasets produced by GSA with a high level of responsiveness.

The Slalom team and Oasys engineers built an infrastructure as code environment with AWS and a set of test framework artefacts. Oasys primarily develops C++ software for the Windows platform, so native C++ and .NET Core were selected as the programming languages for the test frameworks. The infrastructure and test framework artefacts were then used to understand the performance of the three data storage solutions.

Throughout the process, Slalom introduced modern development practices to Oasys, providing guidance and implementation strategies related to user-centric software delivery techniques and operating strategies. Working with the Slalom team, Oasys engineers began the journey of up-skilling in modern software development tools and ways of working.

Next steps

The experimentation phase of the project revealed that TileDB was the best database fit in terms of speed and being able to pull and access data for GSA. However, AWS Postgres would allow more users to quickly pick up the database and begin writing their own queries.

“In considering databases, it’s about trying to balance both of these pieces,” says Stephen Hendry, associate, Oasys Ltd. “There’s a large community of people who like to get their fingers dirty, extract things, do their own processing. So there’s nothing definitive yet.”

As Oasys moves toward a decision, which may involve some combination of databases, the company has its sights set firmly on the next phase of its journey to the cloud—implementation.

Learn more in our whitepaper.