CADFEM UK and Ireland Ltd was founded in 1997 by managing director Derek Sweeney, a mechanical engineer. CADFEM, the ANSYS Channel Partner in UK and Ireland, specializes in computer-aided engineering (CAE) for a range of sectors, including companies working in renewable energy, aerospace, and civil and mechanical engineering. The company’s motto—“simulation is more than software”—reflects the fact that CADFEM goes beyond supplying the tools for simulation. Its experts also provide consultation, training, and high-performance computing (HPC) to help firms develop accurate models of their products using ANSYS.
CADFEM was finding it hard to compete with larger organizations when bidding on major contracts, where computational needs are extremely high. “Anyone involved in simulation will tell you the industry has a constant need for faster and faster computers,” says Sweeney. “We have some very powerful workstations in our office, but when we were up against a firm with a 500-core, high-performance cluster, we were at a disadvantage.”
Small firms often don’t have the budget to build such large clusters, but when cloud-computing services started becoming more widespread and affordable, new opportunities arose for CADFEM.
“The cloud offered a cost-effective way to get access to high-performance computing,” Sweeney says. “We saw the opportunity to switch clusters on and off as required and pay only for what we use.”
CADFEM started using Amazon Web Services (AWS) in 2010 to bolster its internal compute resources when doing consulting work for clients. Soon after, more clients wanted to know about AWS, and CADFEM began sharing Amazon Machine Images (AMIs) with clients who wanted to do the same. “We soon realized this could open up a new revenue stream for us,” says Sweeney.
In 2015, CADFEM became an AWS Channel and Advanced Consulting Partner, allowing it to resell Amazon AMIs with a full suite of ANSYS simulation software, backed up by the expertise of technical specialists from AWS. "So, along with the software licenses we were selling, we started giving customers the option of purchasing AWS offerings for simulation computing as a scalable alternative to traditional workstations, " Sweeney says. "This will help us take our offering into other markets with great potential like the UK and Germany. "
CADFEM uses Amazon Elastic Compute Cloud (Amazon EC2) instances, along with Amazon Elastic Block Store (Amazon EBS) and Amazon Simple Storage Service (Amazon S3) to build its HPC environments. “Setting up in AWS was quite easy,” says Sweeney. “Once we got started, there was very little difference for end users between using one of our on-premises workstations and AWS. It’s just like adding a new machine to our local area network.” CADFEM also makes extensive use of the AWS API to tailor its self-service portal for simulation needs. For example, it lets users launch a Linux cluster, connected to a Windows instance for pre- and post-processing, at the push of a button. “The AWS API is powerful and streamlines the way we and our customers work,” says Sweeney.
Sweeney believes using AWS to support simulation has “levelled the playing field” between small, specialized engineering firms and large enterprises. One of these specialized firms is OpenHydro, which designs and manufactures marine turbines to generate renewable energy from tidal streams. It needed to simulate the flow of water on new blade designs, but with 200 runs for different tidal conditions, and each run taking 13 hours, the total study would have taken two years on the company’s existing workstations. CADFEM configured a 128-core, high-performance Linux cluster on AWS, reducing the time of a single run to less than one hour, allowing the study to be completed in two months and reducing the cost per run by 75 percent. “Using AWS we're helping companies like OpenHydro create better products faster and more cost-effectively, in some cases halving their time to market. This helps them stay ahead of the competition,” says Sweeney.
Uptake of CADFEM’s simulation solutions—which are built on AWS—has been very high. “Pretty much all our new ANSYS customers sign up for it, because if you can have this additional compute capacity on a pay-per-use basis, why wouldn’t you?” says Sweeney, who expects three-quarters of customers will be using the solution within one year.
“It’s also really easy to deploy in other regions, like Germany, because of the coverage of AWS data centers,” he says. “Plus, we can give new customers a trial environment within 30 minutes by providing them with access to one of our AWS instances. And once they’ve decided to buy, setting them up takes about a day, as opposed to two weeks with a physical workstation.”
Another advantage of the AWS Cloud for CADFEM customers is the fact that the technology is always evolving. “Often a company will buy expensive, high-performance hardware, and it’s obsolete within three years,” says Sweeney. “But AWS constantly upgrades the chipsets it uses, and the cost of computing comes down all the time, so customers get the fastest performance available.”
To learn more about how AWS can help run simulations in the cloud, visit our high-performance computing details page.