DCLGApr 27, 2020

The Dark Side of Unikernels for Machine Learning

arXiv:2004.13081v14 citations
Originality Synthesis-oriented
AI Analysis

This addresses deployment challenges for ML practitioners using unikernels, but it is incremental as it focuses on tooling rather than fundamental changes.

The paper identifies shortcomings of unikernels for deploying machine learning inferencing applications and proposes a tool to manage dependent libraries, aiming to improve the build process and leverage unikernel security and performance benefits.

This paper analyzes the shortcomings of unikernels as a method of deployment for machine learning inferencing applications as well as provides insights and analysis on future work in this space. The findings of this paper advocate for a tool to enable management of dependent libraries in a unikernel to enable a more ergonomic build process as well as take advantage of the inherent security and performance benefits of unikernels.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes