Abstract: Data Processing Unit (DPU) is a System on Chip (SoC) based network interface cards (NICs) that can perform processing and analysis of network, security, storage for datasets including virtualization. In this tutorial, We will begin by providing an overview of how modern day Data Centers (DCs) work and introduce the key networking requirements (Virtual Machine bridging, Virtual Extensible LAN, SR-IOV, paravirtualization and Container networking) and the limitations with the existing host-centric networking applications. We will showcase the key bottlenecks that arise at the Network interface end of any node (noth physical host or virtual entity).
Next we will introduce the Data Processing Unit (DPU) and its diverse use cases in the field of high performance compute/network. We will describe each and every component that a DPU houses and how we can leverage the DPUs for basic compute, networking and storage use cases. We will focus on presenting how a data packet traverses into the DPU and how the network packet signatures can be processed and analyzed in the SoC, and how we can look for any security layers on the header section of the network packets.
Finally, we will introduce the audience to NVIDIA DOCA – a software framework that brings together APIs, drivers, and libraries, to simplify and speed up application development and deployment on BlueField DPUs. We will try to conduct the live hands-on session on an actual Bluefield-2 DPU 100Gbps, to showcase the capabilities of DPU, We will guide and make the audience build the packet inspection, and/or malware detection tools for building robust network security applications.
Duration: 3 hours
Designation and Affiliation:
1 System Architect, Indian Institute of Technology Delhi, Hauz Khas, New Delhi, Delhi 110016
2 Assistant Professor, Indian Institute of Technology Gandhinagar, Palaj, Gujarat 382055
Mr. Sudeep Narayan Banerjee is working at Indian Institute of Technology Delhi (IITD) as System Architect since Nov 2022 and with a total experience of 14+ years in the field of High Performance Computing, Virtualisation, Cloud Computing, Parallel and Distributed Computing. Previously he has worked with Indian Institute of Technology Gandhinagar (IITGN) where he was the Staff In-charge of the Supercomputing Facility. He received his M.Tech degree in Computer Science and Infrastructure from BITS Pilani, in 2021, and B.Tech degree in Electronics and Instrumentation Engineering from West Bengal University of Technology, Calcutta, in 2008. He has designed and commissioned several HPC clusters at IITGN and former institutions. He has published 2 papers in reputed journals. His research interests include High Performance Computing Cluster, GPU/DPU, Network Function Virtualization, Distributed systems, Cloud Computing, CyberSecurity. He loves to teach and is always ready to converse on topics to Physics, Computer Science and Education.
Dr. Sameer G. Kulkarni is an Assistant Professor in the Department of Computer Science and Engineering at Indian Institute of Technology (IIT) Gandhinagar. He received a Ph.D. degree in Computer Science from University of Göttingen, Germany in July 2018, and worked as a postdoctoral researcher at the University of California, Riverside. He received his M.S. degree in Computer Engineering from the University of Southern California, in 2010, and B.E. degree in Computer Science and Engineering from National Institute of Engineering, Mysore, in 2004. He has made active contributions in the field of cloud security and network virtualization. He has over 20 publications and has published papers in premier journals and conferences such as Transactions on Networking and Transactions on Network Service Management journals, and ACM SIGCOMM, Symposium on Cloud Computing, CoNEXT, IEEE DASC (Dependable, Autonomic and Secure Computing) conferences. He is the recipient of the IEEE Technical committee on scalable computing outstanding doctoral dissertation award 2019. His work focuses on the resource management aspects towards building Efficient, Scalable and Resilient NFV/Edge platforms. His research interests include Software Defined Networking, Network Function Virtualization, Edge Cloud Platforms, Distributed systems, and Disaster Management.