EDA in the Cloud: Key to Rapid Innovative SoC Design

Article By : Mahesh Turaga, Cadence Design Systems

More and more companies are turning to EDA in the cloud as they gradually overcome concerns about security and IP protection.

Simultaneous mega-trends are shaping multiple industries from aerospace and defense, automotive and high-tech to healthcare and others. These include 5G, autonomous vehicles, industrial internet of things (IIoT), electrification, hyperscale computing and artificial intelligence / machine learning (AI/ML). Add cloud to the mix, and we have another generational disruption that has driven business over the past decade and been further accelerated by our current global situation, changing the way we work, live, communicate and entertain. Cloud opportunities go far beyond flexible ubiquitous access.

In the preceding decade, the move towards cloud computing occurred primarily in sectors like retail and finance, with the advent of leading cloud vendors such as Amazon Web Services (AWS), Microsoft Azure, Google Cloud Platform and others accelerating the trend. In the electronic design automation (EDA) space, until recently, traditional concerns about security, protecting intellectual property (IP) and data outweighed the significant advantages offered by computing in the cloud — such as flexibility, scalability and productivity.

Mahesh Turaga - Cadence
Mahesh Turaga

That is now changing, and the cloud-enabled value of each of those industries is driving the need for intelligent systems. We now see that the only way such systems can be created is by using cloud-enabled computing tools and methods. It is the “systems” that are driving the need for massively parallel computing with close-to-linear performance growth and virtually unlimited scalability while maintaining the highest level of accuracy. Those results are possible only in the cloud. With leading foundries in the space adopting cloud and acknowledging the security of cloud infrastructure by having their process design kits (PDKs) in the cloud, security concerns have by and large diminished.

Multiple companies involved with EDA and systems development — from electronics to mechanical design — are increasingly recognizing the transformative power of the cloud and cloud computing to deliver fast-paced innovation to their end consumers, enabling users to come up with new business models that would not have been possible previously.

New startups driving innovation in emerging industries such as electrification are at the forefront of cloud adoption as a way to collaborate globally, deliver innovation at lightning speed and compress the traditional notions of product life cycles from decades to a few years — whether it’s coming up with a new electric vehicle or an eVTOL. These companies are proving to the broader EDA and systems space that the cloud is so much more than a delivery mechanism. The cloud is an accelerator of innovation, business reinvention and lightning growth.

To innovate faster in a digital transformation scenario, the five key pillars of innovation in an enterprise—people, data, processes, technology and tools — are reimagined in more efficient ways by a broadside cloud adoption. The advances brought by hyperscalers make this transformation even more seamless to companies of all sizes, from small and medium businesses to large enterprise, systems and semi companies. The question of cloud adoption is then, not about cost savings or even about technology per se, but about a fundamental business model transformation that companies can achieve to better serve their end customers.

Collaboration, innovation and cloud computing

In the EDA space, companies can benefit by collaborating in the cloud seamlessly across groups and geographies, thereby accelerating innovation from the chip level to the systems level. Chip designers can collaborate on complex chip and SoC designs instantaneously and communicate in real time with their mechanical counterparts through ECAD-MCAD collaboration in the cloud.

Real-time, dynamic collaboration across these two key domains enables many of the companies to come up with the most complex, next-generation cyber-physical products and systems that are at the core of the current generational trends. Industry leadership must take steps towards a next-generation cloud platform that enables this seamless collaboration across different disciplines, from the chip level to the systems level, from electronics to mechanical domains, thus enabling a true digital twin.

Beyond internal collaboration, industry collaboration is simplified and expedited via the cloud. Cloud collaboration brings together the foundry team, design team and EDA team for advanced process node adoption, proliferation and support. Machine learning (ML) is now required to overcome the technical and EDA tool flow complexity for processes below 7nm. And the cloud provides the necessary compute throughput for these ML workloads.

The cloud as a computer

The theoretically unlimited compute capacity in the cloud is another driver of this innovation brought upon by the cloud. While many large companies invested in on-premises server farms and today find it more economical to continue to run locally, there’s an increasing realization that a simple cost comparison is too short-term.

Thinking long-term and reimagining how moving to the cloud will transform their businesses and their end customers in the next 5, 10 or 15 years is going to reshape the dynamics in EDA going forward. Smaller and medium-sized companies are already increasingly moving to the cloud to take advantage of the unlimited compute capacity that helps them deliver complex SoCs in months versus years while collaborating seamlessly across remotely distributed teams.

This trend is going to accelerate in the coming years, and we are seeing users adapt to this reality. But they require a seamless transition to the cloud, new business models on the cloud, with more sophisticated collaboration capabilities.

Along with the move to the cloud come the challenges of optimizing a company’s compute needs in the cloud as a way to not only control costs but also to democratize innovation across the enterprise. Those taking advantage of EDA in the cloud are increasingly demanding business models that are prevalent in retail to run anything from a logic simulation to the most complex chip emulation to large multiphysics simulations on demand.

While the last decade was mostly about retail business transformation via the cloud, the next decade is going to be about the transformation of electronics and mechanical systems companies due to cloud computing.

This article was originally published on EE Times.

Mahesh Turaga is VP of business development for cloud at Cadence Design Systems Inc.



Subscribe to Newsletter

Leave a comment