Skip to content

Microsoft Industry Blogs


In today’s modern technology marketplace, companies must innovate at breakneck speed to compete. The first gasoline-fueled automobiles were introduced to the public in the late 1800’s yet remained novelties until the vehicles became affordable and practical in the 1920’s. The automobile has gone through significant changes since – becoming faster and more efficient. But now, as technology accelerates the pace of innovation, it is no longer enough for manufacturers to build a faster, cheaper or more attractive automobile. Now, new electric cars flood the marketplace and advances in autonomous driving technology hint at a dramatic shift in how personal transportation will look in the coming years. As vehicles become connected, manufacturers are also given a new wealth of data to incorporate into the next generation of vehicle design and engineering. These trends are not confined to the automotive industry, and are putting pressure on manufacturers to evolve their product development and accelerate innovation. Manufacturers are being asked to iterate faster on design while also improving quality by simulating and replicating the real-world performance of physical products and incorporating feedback from the ever-increasing number of IoT-enabled devices.

In response to these challenges, manufacturers have turned to high-performance computing (HPC) and deep learning to help drive the product development lifecycle. New advances in HPC and deep learning technology are enabling manufacturers to eliminate these physical and financial constraints and realize benefits that make them more nimble, flexible, efficient and leading-edge. Initially, this solution may seem ideal, but manufacturers often run into additional challenges that impact productivity and scalability. Continued success in HPC infrastructure requires ongoing capital expenditure, straining engineering budgets. Once the infrastructure is in place, it quickly grows obsolete, creating a gap between existing capabilities and imminent next-generation hardware. In addition, lack of physical space leaves teams unable to easily expand on-premises infrastructure to meet their goals.

Big Compute Microsoft graphic

Empower collaboration by enabling distributed teams to work from anywhere, on any device, across multiple sites

As manufacturing grows more complex, an organizations’ designers and engineers are operating across multiple sites and geographies, with extended supply chains of OEMs and other 3rd party vendors. Engineers need the ability to be productive wherever they are, without being tied to on-premises equipment. Providing access to high-powered virtual workstations through consumer-grade mobile devices enables them to engineer from anywhere, whether they’re in the office or out in the field. With added mobility and flexibility, companies eliminate the need to create separate instances and time-consuming tunneling processes, paving the way for secure collaboration throughout the supply chain in real time.

Enhance product design iteration and optimization by leveraging digital twin

enhance car image

The convergence of physical and digital components in manufacturing has enabled a shift towards smarter, more connected products that are continuously refined. Innovation and iteration produce more sophisticated product designs that must be quickly trained, simulated, and validated to achieve functional quality and safety goals. Standard simulations such as Computational Fluid Dynamics (CFD) and Finite Element Analysis (FEA) can take hours to execute, leaving engineers with extended wait times to get results. Engineers want to iterate product designs quickly, without waiting in line. They also need to integrate IoT data from existing products into simulations to derive new insights and speed up the validation process. When coupled with AI, a control loop is created that provides feedback and helps to analyze and validate results. By creating a virtual model of physical products or production assets, manufacturers can use or leverage this digital twin concept to integrate IoT data and run numerous simulations, backed by the compute power of the cloud.

Support changing business requirements by scaling up or down as deadlines shift

In an ideal world, manufacturing organizations could schedule engineering projects to maximize compute resource utilization. The reality is that compute requirements come in peaks and valleys. Releases often overlap, schedules compress, and plans suddenly change, creating a backlog of jobs that trigger spikes in core usage. When business requirements demand more compute power than existing infrastructure can support, companies have two choices – invest in additional on-premises computing resources and hope it will arrive soon enough to meet demand, or turn to the power of the cloud. Given the inconsistent nature of compute requirements, companies are wary of being stuck with excess capacity following a spike in usage, which often results in a lack of compute power when it is needed most. Investing in the cloud eliminates this difficult decision; the cloud is more cost-effective and much quicker, enabling businesses to spin up virtual machines and scale into the cloud, or instantly extend their on-premises capabilities. With the cloud, companies can optimize for changing priorities and control costs, paying only for what they use.

Transform large-scale training and validation by engaging cloud resources on demand

Today’s connected assets have intricate systems of sensors, cameras and central computing systems. Each must be trained and programmed to perform as intended, a process that often spans weeks or even years. A report published by the Rand Corporation in 2016 estimated that 10 billion miles of testing would be required to produce a fully autonomous car.1 Multiply this training process by the number of assets in use across global worksites, and you rapidly discover a massive big data and big compute challenge. To assimilate vast and complex data sources and devices, manufacturers need storage, networking and compute capacity to either augment or replace existing systems and infrastructure – not only at a centralized R+D facility but at the edge as well.

By integrating cloud computing with open source software tools and deep learning frameworks, manufacturers can more efficiently execute large-scale programs that train the products of the future.

Introducing Azure Big Compute

Azure Big Compute is a portfolio of end-to-end HPC and deep learning solutions from Microsoft that accelerates innovation across the product development lifecycle – a portfolio that includes Microsoft’s Cloud Workstation, HPC Simulation and Analysis, Deep Learning and AI Training, and Cloud Rendering. With Azure Big Compute capabilities, manufacturers are empowered to do more, iterate faster, and take advantage of the scale and flexibility that the cloud brings to organizations of all sizes. Run Windows and Linux applications, solvers, and frameworks at scale on demand, from anywhere backed by Microsoft’s Trusted Cloud security.

Get started now

With Azure Big Compute, manufacturers are already realizing the powerful performance, scale, and flexibility of Microsoft’s global hyper-scale cloud. Don’t let a lack of computing power and networking/storage infrastructure be barriers standing between your engineers and the next generation of products.

Are you ready to unlock the full value of Big Compute for your business? Learn more about our Big Compute solution portfolio today.


1 Rand, 2016 http://www.dailymail.co.uk/sciencetech/article-3535368/Will-self-driving-cars-safe-hit-roads-Report-claims-vehicles-need-drive-billions-miles-hundreds-years-test-eventualities.html