VHPC 2025 Program
For detailed timing and schedule information, please refer to the schedule page.
Tutorial
Writing a hypervisor from scratch
Presenter: Seiya Nuta
Affiliation: Vercel Inc.
This tutorial provides hands-on experience in hypervisor development, covering fundamental concepts and practical implementation techniques for building virtualization infrastructure from the ground up.
Abstract: Hardware-assisted virtualization technologies like Intel VT and AMD SVM have long been viewed solely as mechanisms for running virtual machines. This talk challenges the preconception - it's actually a hardware-assisted try-catch mechanism, where the hypervisor effectively behaves as a "catch" block for VM exits. Through this metaphor, we will revisit virtualization from scratch, explore recent virtualization-based applications, rethink what hardware-assisted virtualization is, and brainstorm what we could invent in the near future.
Accepted Paper Presentations
Enabling RDMA and GPUs in Rootless Kubernetes for Accelerated HPC and AI Applications
- • Lise Jolicoeur - University of Bordeaux & CEA, France
- • François Diakhaté - CEA DAM Ile de France, France
- • Daniel J Milroy - Lawrence Livermore National Laboratory, USA
- • Vanessa Sochat - Lawrence Livermore National Laboratory, USA
Abstract: HPC workflows increasingly integrate diverse technologies such as artificial intelligence (AI), machine learning (ML), data analytics, databases, and web services. Orchestrators like Kubernetes have been designed to facilitate deploying these heterogeneous workloads in cloud environments. Allowing Kubernetes to be launched and managed as a resource on HPC clusters would facilitate the deployment of modern workflows in HPC environments. To enable the deployment of Kubernetes by unprivileged HPC users, we evaluate the usability of a rootless version of Kubernetes, Usernetes. We analyze synthetic benchmarks as well as HPC and ML proxy apps to evaluate the overhead of Usernetes for HPC/ML workloads deployed on high performance networks and GPUs. While the results show that applications running in Usernetes can take advantage of InfiniBand networks and NVIDIA GPUs, some benchmarks incur measurable overheads at scale which warrant further investigation.
Performance Analysis of Container-in-VM Architectures: A Study on Hypervisor Isolation and Lightweight OS Integration
- • Vijayalakshmi Saravanan - University of Texas at Tyler & BNL, USA
- • Khaled Z Ibrahim - Lawrence Berkeley National Laboratory, USA
- • Manoj Kumar Patra - The LNMIIT, Jaipur & National Institute of Technology, Rourkela, India
Abstract: Container-in-VM architectures have become essential in modern cloud-native infrastructure, offering a secure yet flexible approach to deploying containerized workloads. This paper comprehensively reviews the architectural and performance trade-offs associated with deploying containers inside virtual machines, focusing on hypervisor-level isolation and lightweight operating system integration. We systematically examine a decade of research and industry innovations, synthesizing findings from benchmarking studies and architectural implementations. The paper introduces a comparative framework covering isolation strength, runtime overhead, boot latency, OS footprint, and orchestration compatibility, highlighting how solutions like Firecracker, Kata Containers, and unikernels deliver varying balances of efficiency and security. A significant contribution is the analytical contrast between general-purpose and minimalistic operating systems when used as VM guests, revealing substantial gains in performance and attack surface reduction with specialized OS choices. Similarly, hypervisor technologies are evaluated in terms of their ability to support microVMs and fast container lifecycle management, with observations drawn from real-world deployments and empirical studies. The findings demonstrate that combining lightweight OS stacks with hardware-assisted virtualization can significantly enhance container performance, though challenges persist in orchestration, observability, and standardization. This review identifies critical gaps and offers direction for future research in building secure, efficient, and scalable container-in-VM platforms suitable for cloud, edge, and serverless environments.
WebAssembly and Unikernels: A Comparative Study for Serverless at the Edge
- • Valerio Besozzi - University of Pisa, Italy
- • Enrico Fiasco - University of Pisa, Italy
- • Marco Danelutto - University of Pisa, Italy
- • Patrizio Dazzi - University of Pisa, Italy
Abstract: Serverless computing at the edge requires lightweight execution environments to minimize cold start latency, especially in Urgent Edge Computing (UEC). This paper compares WebAssembly and unikernel-based MicroVMs for serverless workloads. We present Limes, a WebAssembly runtime built on Wasmtime, and evaluate it against the Firecracker-based environment used in Spare. Results show that WebAssembly offers lower cold start times for lightweight functions but struggles with complex workloads, while Firecracker provides higher but more stable cold start times and better execution performance, particularly for I/O-heavy tasks.
Research Areas Covered
Hypervisor Development
- • From-scratch hypervisor implementation
- • Virtualization infrastructure fundamentals
- • Hands-on development techniques
HPC & AI Integration
- • RDMA networking in containerized environments
- • GPU acceleration for AI workloads
- • Rootless Kubernetes deployment
Container & VM Architecture
- • Container-in-VM performance analysis
- • Hypervisor isolation vs. performance
- • Lightweight OS integration strategies
Edge Computing Technologies
- • WebAssembly for edge serverless
- • Unikernel deployment strategies
- • Resource-constrained environment optimization
Workshop Format
Tutorial Session
50-minute hands-on tutorial on hypervisor development
Research Papers
3 technical papers covering virtualization, containers, and edge computing
Duration
3 hours of technical content with coffee break
General Information
The 20th Workshop on Virtualization in High-Performance Cloud Computing (VHPC 2025) will be held in conjunction with the International European Conference on Parallel and Distributed Computing (Euro-Par 2025), on Aug 26, 2025, in Dresden, Germany.
For more information, refer to either the VHPC 2025 homepage or the Euro-Par 2025 web pages.