Two papers from the WIDE team have been accepted for presentation at ACM Middleware 2024. The first paper concerns predicting the effects of a system-level patch on applications running inside virtual machines in the Cloud. The work entitled “𝑈𝑇𝑤𝑖𝑛𝑉𝑀: Reliable hints on the effects of hypervisor updates on VMs in the Cloud” addresses the problem above with a non-negative least squares model and carefully crafted applications that mimic the behaviour of applications running in VMs without introspection, thus preserving VM security in the Cloud. The work is led by Djob Mvondo of the WIDE team, Tong Xing, and Antonio Barbalace (both from the University of Edinburgh).
The second paper, “Horse: Ultra-low latency workloads on FaaS platforms,” addresses the problem of efficiently launching microsecond-scale workloads on FaaS, the fast-growing platform model on Cloud platforms. Horse introduces a novel algorithm, PPSM, which aims to improve the resume time of paused virtual machines by up to 142.84x. Djob Mvondo, Francois Taiani, and Yerom-David Bromberg, all members of the WIDE team, lead the work.
Both papers will be presented during the Middleware 2024 conference at Hong Kong, 2-6 December 2024.
