High Performance Computing (HPC) Cloud Security Working Group
Introduction to the High Performance Computing (HPC) Cloud Security Working Group
Mission Statement: To develop a holistic security framework for cloud infrastructure architected for High Performance Computing (HPC) needs, with the aim of securing where the cloud environment and HPC cross paths.
Today, the increasing complexity of different types of workloads has resulted in the diversity of infrastructure architectures to serve them. The use of cloud to process HPC workloads is an illustration of this trend. For example, Microsoft’s Azure Batch leverages on the scale and flexibility of the cloud for big compute, and SAP HANA which AWS has a special HPC instance created for it. Google Cloud has also ventured into HPC with its massively multi-core instance to provide a powerful and flexible infrastructure. However, amongst all the demonstrated efficacies that cloud has brought about, researchers face certain challenges running HPC in a cloud computing environment.
One of the most prominent issue involve the architectural requirements that support most HPC workloads. Due to the high-performance requirements of these workloads, ‘close to metal’ operations are often demanded, stretching the processor’s core physical computing resource to its utmost capabilities. The apprehension lies in that running on a virtualized hypervisor may cause performance to suffer. Also, due to the nature of high-performance requirements, high-speed interconnect is an element encompassed by HPC for fast communication. Storage is also a factor that influences high-speed communications requirements. HPC applications used by researchers frequently handle a colossal volume of data which are pulled from consistent storage. For lower-latency access, data are often divided between multiple storage instances, further straining on I/O environments.
These technical concerns for HPC are further complicated by the complex and ever-evolving threat landscape. As we increasingly see cases of pure HPC bare metal infrastructure interacting with the cloud such as I/O interfaces and processes, it brings along more ‘opportunities’ for malicious attacks. While this should be considered and integrated into security policies and guidelines, performance face the peril of being compromised as precious resources are carved out for security protocols and processes.
‘Vanilla’ cloud environments were typically not made to handle environments like this. The crossing of cloud and HPC environments often leads us to questions of how security in an HPC cloud environment can be implemented, enforced and ensured without the need to compromise performance. This Working Group strives to provide recommendations that can answer these questions.
Download the High Performance Computing (HPC) Cloud Security Working Group Charter
High Performance Computing (HPC) Cloud Security Working Group Leadership
High Performance Computing (HPC) Cloud Security Co-chairs
Ong Guan Sin
ONG Guan Sin has over twenty years of diverse experience in the IT industry, spanning roles in IT management of end-user organisations,fast-paced startup businesses and IT vendors. The companies he has worked for include DBS Bank, National University of Singapore, i-Email.net/i-DNS.net, SCS Ltd (now NCS Pte Ltd) and Technology Reserve (Canada). He has experience working with the processes in IETF and SPEC.
Currently serving as head of new services and cybersecurity in National Supercomputing Centre, Singapore, he is chartered to promote the adoption of high-performance computing (HPC) by the industry with the appropriate security provisions and standards. He also sees opportunity in using HPC for security applications. He believes that IT security must align with the business goals and the appetite for risks of the organisation concerned.
Cloud Team Manager, National Computational Infrastructure Canberra Australia
Andrew has many decades of hands-on technical, diplomatic and logistics experience covering a wide range of standard and bespoke technologies, languages and applications within Industry, Government, Academia and Research nationally and internationally.
He is a member of the APAN Program Committee, Co-Chair of the APAN E-Culture and Asia Pacific Research Platform working groups and the Lead judge on the SCA19 Data Movement Challenge.
Through his work on the InfiniCortex and InfiniCloud projects, Andrew has designed and delivered the next generation of high speed data transfer and distributed computation throughout the Asia Pacific region and around the world.
High Performance Computing (HPC) Cloud Security Working Group Initiatives
Please contact Working Group Leadership for more information.