Agility, flexibility and security: The value of cloud in HPC

In today’s competitive business environment, firms are confronted with complex, computational issues that demand swift resolution. Such problems might be too intricate for a single system to handle or might require an extended time to resolve. For companies that need quick answers, every minute counts. Allowing problems to linger for weeks or months is not feasible for businesses determined to stay ahead of the competition. To address these challenges, enterprises across various industries, such as those in the semiconductor, life sciences, healthcare, financial services and more, have embraced high-performance computing (HPC).

With HPC, enterprises are taking advantage of the speed and performance that comes with powerful computers working together. This can be especially helpful amid a steadily growing push to build AI on a larger and larger scale. While analyzing massive amounts of data might feel impossible, HPC enables the use of high-end computational resources that can perform many computations rapidly and in parallel to help businesses get insights faster. At the same time, HPC is used to help businesses bring new products to market. It is also used to better manage risks and more, which is why an increasing number of enterprises are adopting it.

The role of cloud in HPC

Most commonly, enterprises that run workloads with surges in activity are finding that they exceed the compute capacity available on-premises. This is an example of where cloud computing can augment on-premises HPC to transform the business’s approach to HPC with cloud resources. Cloud can help address peaks in demand during product development cycles, which might last from a short duration to a longer duration, and enable organizations to get access to the resources and capabilities that they might not have a need for around the clock. Businesses using HPC from the cloud can take advantage of the benefits of greater flexibility, enhanced scalability, better agility, improved cost efficiencies and more.

Cadence uses IBM Cloud HPC

Cadence is a global innovator in electronic design automation (EDA) with over 30 years of computational software experience. It has helped companies across the world design electronic products that drive today’s emerging technology, including chips. The growing demand for more chips, along with the company’s incorporation of AI and machine learning into its EDA processes means that their need for compute power is at an all-time high. For organizations in the EDA industry like Cadence, solutions that enable workloads to seamlessly shift between on premises and the cloud, while also allowing for differentiation from project to project, are key.

Cadence uses IBM Cloud® HPC with IBM Spectrum® LSF as the workload scheduler to support the development of chip and system design software, which requires innovative solutions, powerful compute resources and advanced security support. By using IBM Cloud HPC, Cadence reports improved time-to-solution, performance enhancements, cost reductions and streamlined workload management.

Additionally, Cadence understands firsthand that moving to the cloud can require new knowledge and capabilities that not every company possesses. The Cadence Cloud comprehensive portfolio aims to help customers across the world use the possibilities of the cloud with Cadence Managed Cloud Service as a turnkey solution ideal for start-ups and small and medium customers, and with the customer-managed cloud option known as Cloud Passport to enable Cadence tools for large enterprise customers. Cadence is dedicated to giving its customers an easy path to the cloud by connecting them with knowledgeable service providers, such as IBM®, whose platforms can be used to deploy Cadence tools in cloud environments. For enterprises that want to drive innovation at scale, the Cadence Cloud Passport model can deliver access to cloud-ready software tools for use on IBM Cloud.

Taking a hybrid cloud approach to HPC

Traditionally, HPC systems were built on-premises. However, the large models and large workloads that exist today are often not compatible with the hardware that most companies have on premises. Given the high up-front costs of obtaining GPUs, CPUs and networking, as well as those of building the data center infrastructures needed to efficiently run compute at scale, many companies have used cloud infrastructure providers that have already made massive investments in their hardware. To realize the full value of public cloud and on-premises infrastructures, many organizations are adopting a hybrid cloud architecture that is focused on the mechanics of transforming portions of a company’s on-premises data center into private cloud infrastructure.

By adopting a hybrid cloud approach to HPC where cloud and on premises are used together, organizations can use the strengths of both, allowing organizations to achieve the agility, flexibility and security required to meet their demands. For example, IBM Cloud® HPC can help organizations flexibly manage compute-intensive workloads on-premises. With security and controls built into the platform, IBM Cloud HPC also allows organizations to consume HPC as a fully managed service while helping them address third- and fourth-party risks.

Looking ahead

By using hybrid cloud services through platforms like IBM Cloud HPC, enterprises can solve many of their most difficult challenges. As organizations continue to embrace HPC, they should consider how a hybrid cloud approach can complement traditional on-premises HPC infrastructure deployments.

Source: ibm.com