Kairns

Softwares for the operation and study of computing infrastructures

Motivation

Kairns is a collective of computer science researchers and engineers working on the software for the operation and study of computing infrastructures. In particular, the collective focuses on computing systems such as High-Performance Computing supercomputers and Cloud infrastructures. In the last decade, members of the collective have developed various tools for the operation and for the studying of such systems. While these tools target each a specific problem, they can be used in conjunction with each other to cover a large spectrum of the operating and studying needs. The tools of Kairns collective are being used in production settings, research works, as well as in leading European research initiatives.

Back to the top

List of Softwares

Back to the top

OAR

OAR is a versatile resource and job manager (also called a batch scheduler or job scheduler) for HPC clusters and other computing infrastructures (like distributed computing experimental testbeds where versatility is a key). This is the third generation.

Back to the top

Batsim

Batsim is a scientific simulator to analyze batch schedulers. Batch schedulers — or Resource and Jobs Management Systems, RJMSs — are systems that manage resources in large-scale computing centers, notably by scheduling and placing jobs.

Back to the top

NixOS Compose

The goal of NixOS-Compose is to reduce the burden of setting up ephemeral distributed systems thanks to Nix functional package manager and Nixos The main common use cases are: Testing/developing for different targets (vm, container, baremetal). Rapid Infrastructure as Code prototyping. Evaluating and Benchmarking w/ reproducibility guarantee. Advanced experiment for research studies in DS, HPC, Cloud/Edge/Fog, ...

Back to the top

Alumet

Alumet provides a unified interface for gathering measurements with sources, transforming the data with models and writing the result to various outputs. The elements are created by plugins, on top of a standard framework.

Extensible Framework: Alumet can easily be extended in order to make new research experiments. Leverage existing plugins and only add what you need, without reinventing the wheel. Take advantage of the unified data model and parallel measurement pipeline. Operational Tool: the end result is (or aims to be) a ready-to-use measurement tool that is robust, efficient and scalable.

Back to the top

CiGri

CiGri is a lightweight grid middle-ware designed to run on top of a set of OAR (https://oar.imag.fr) clusters to manage efficiently large sets of multi-parametric tasks (also called bag-of-tasks).

With a powerful events handling, it manages automatic re-submission that is useful for best-effort jobs. With OAR and CIGRI, it is easy to reach 100% of usage efficiency on an heterogeneous grid in a local HPC center.

Back to the top

Kameleon

Kameleon is a simple but powerful tool to generate customized operating system images, based on traceable recipes.

Back to the top