De La Pierre M1, Golovina E1, Han L1, Blasco J1
1Do It Now, , Aotearoa / New Zealand
Biography:
Marco De La Pierre is a Computational Scientist at Do IT Now, a global HPC services and consulting company. Building on a background in computational materials science and on 15 years of experience in High-Performance Computing, he enjoys investigating new software technologies to design and deploy HPC solutions that facilitate user experience and productivity. Active across various aspects of applications and user support, his interests include workflow automation, reproducibility and portability, scientific software stack management, and user training.
Abstract:
Research workloads are growing in scale, complexity, and urgency. At the same time, most HPC centres rely on on-premises infrastructure, which can become a bottleneck during peak demand and mission-critical operations.
Cloud bursting addresses these challenges by extending local infrastructure with cloud resources when needed. Rather than seeing the cloud as a separate environment, cloud bursting integrates it into the wider HPC environment, giving users access to additional compute power while preserving a consistent and seamless experience.
Cloud bursting has also become a key mechanism for providing a low-cost Disaster Recovery (DR) capability required by operational and mission critical workloads.
In this talk, we will explain how cloud bursting helps HPC centres support modern scientific, AI, and data-intensive workloads. We will describe the key elements required to make it work properly, including scheduler integration, user access, networking, storage, software environments, monitoring, and cost control. We will review practical use cases and common challenges, such as data transfer, security, and expected cloud costs. Finally, we will discuss cloud bursting from an operational perspective. A successful setup is not just about starting cloud instances. It requires clear processes, reliable automation, consistent environments, and good visibility across both local and cloud resources.
When implemented properly, cloud bursting can help HPC centres become more flexible, reliable, responsive and cost effective. With cloud bursting, organisations can better support user demand, reduce pressure on local infrastructure, reduce the cost of DR capability, and give researchers faster access to the resources they need.
