It’s becoming ever more challenging for companies to remain competitive in their industries. In 1964 the average tenure of companies on the S&P 500 was 33 years. By 2016 that number shrank to 24 years, and by 2027 the average tenure will be a mere 12 years. The reason for such a drastic change is the ability of companies to innovate.
This truth was recently on my mind as I interviewed the HPC leader at an automotive company. He told me how their engineers are working tirelessly towards a launch that will revolutionize the auto industry. Success requires that he and his team hit strict requirements for performance and safety. This process consists of many iterations on each aspect of their design, then thorough testing.
“Ultimately, it’s our fundamental ability to innovate that sets us apart.” he said, “That’s what sells our product and builds our brand. It’s honestly what makes us a company.” With innovation in the front seat, his company invested in an HPC environment to support the design and testing process. As their team grew, so did their demand. Eventually, this business critical system could not keep pace and they frequently faced slow completions.
Unfortunately, static business critical systems like HPC can stunt progress within an organization. We often hear from engineers, “If I had additional resources, I could do more tests, more iterations, and better designs.”
So why don’t IT teams invest in more robust business critical systems? “Expanding our cluster is non-trivial, it requires time, space, resources, delivery time, and most of all, a lot of capital,” said the HPC leader. “And at times our demand is fairly unpredictable.”
All too common is this tension between R&D teams and IT managing business critical resources. Underutilized hardware is wasteful, and so is reduced cycles in engineering and design. Do you sacrifice innovation and competitive edge to appease the bottom line?
Cloud computing, on the other hand, is poised to eliminate this friction, delivering the competitive edge companies require to be a market leader. While an organization may have unmet needs with its on-premises infrastructures, the on-demand resources available through Microsoft Azure provide capacity and scale that solve critical business issues.
“In contrast to on-premises hardware, the cloud grows and shrinks based on our stage of product development. The resources are there when we need them and gone when we don’t,” he said. “We can accelerate development, react to changes in the market to capitalize quickly on new opportunities.”
Flexible and agile business critical systems like Azure offer additional advantages outside of scale, they also offer specialized hardware. “We found 3x faster speeds when running certain workflows on GPU’s,” I recently heard from an engineer in aerospace. Advantages like these enable organizations to be higher-performing, more resilient, and stable.
But not everyone is convinced that cloud is the end-all solution. “It’s so much more expensive than running on my local hardware,” said one researcher. And yes, if we look at core hour vs. core hour, cloud really doesn’t seem efficient. But is that an accurate comparison? Shouldn’t we compare the total cost of these solutions, apples to apples?
To properly evaluate, take into account all that goes into on-premises hardware; upfront capital, headcount to manage it, floor space, utilities, and other maintenance and replacement costs. Not to mention low utilization or project delays that might occur as a result of queuing or slower hardware. Once those factors are applied to core hour cost, the playing field becomes level if not in favor of on-demand resources.
“Especially at peak demand periods, bursting to cloud saves us valuable analysis time and ensures project delivery without a heavy investment for additional hardware,” the HPC leader explained. “It actually makes more economical sense for our situation.”
Economics aside, others see cloud HPC resources as adding an additional layer of complexity. Building out the platform and integrating software with a cloud API is not a weekend project. It requires a heavy lift from internal resources and constant maintenance.
Accessing Azure HPC resources on Rescale alleviates the complexity of migrating workflows to the cloud. “Our project had fallen behind and we needed a quick solution to accelerate our analysis,” said the lead engineer at an aerospace company. “Rescale onboarded us quickly, giving us the resources to finish our runs in a few days compared to weeks on our own clusters.”
Rescale provides a turnkey platform for engineers and R&D to quickly access Azure HPC hardware. Using workflow intelligence, it will recommend the best core type for your specific use case to deliver performance and cost-efficiency. And with a full suite of management tools, IT and HPC leaders retain full control and visibility over business-critical systems as if they were local.
“Rescale is ideal because we can use any hardware whenever we need it,” he said. “The only limitation is on our side – like how many iterations our imaginations can come up with.”
Regardless of the industry, robust and stable business critical systems can accelerate innovation. This gives companies a competitive edge to drive growth and secures their place as market leaders.
If you want to learn more see how the Rescale platform can increase your pace of innovation contact us for a free demo. Learn more about how you can run business-critical applications on Azure by visiting Microsoft’s website or reading this blog.