Session
FOSDEM 2021 Schedule
Cloud Computing

Let There Be Topology-Awareness in Kube-Scheduler!

Enhancing Kubernetes Scheduler
M.misc
SWATI SEHGAL
With Kubernetes gaining popularity for performance-critical workloads such as 5G, Edge, IoT, Telco, and AI/ML, it is becoming increasingly important to meet stringent networking and resource management requirements of these use cases. Performance-critical workloads like these require topology information in order to use co-located CPU cores and devices. Despite the success of Topology Manager, aligning topology of requested resources, the current native scheduler does not select a node based on it. It's time to solve this problem! We will introduce the audience to hardware topology, the current state of Topology Manager, gaps in the current scheduling process, and prior out-of-tree solutions. We'll explain the workarounds available right now: custom schedulers, creating scheduling extensions, using node selectors, or manually assigning resources semi-automatically. All these methods have their drawbacks. Finally, we will explain how we plan to improve the native scheduler to work with Topology Manager. Attendees will learn both current workarounds, and the future of topology aware scheduling in Kubernetes.
Kubernetes has taken the world by storm attracting unconventional workloads such as HPC Edge, IoT, Telco and Comm service providers, 5G, AI/ML and NFV solutions to it. This talk would benefit users, engineers, and cluster admins deploying performance sensitive workloads on k8s. Addition of newer nodes running alongside older ones in data centers results in hardware heterogeneity. Motivated by saving physical space in the data centers, newer nodes are packed with more CPUs, enhanced hardware capabilities. Exposing to use fine grain topology information for optimised workload placement would help service providers and VNF vendors too. We’ll explain numerous challenges encountered in efficiently deploying workloads due to inability to understand the hardware topology of the underlying bare metal infrastructure and scheduling based on it. Scheduler’s lack of knowledge of resource topology can lead to unpredictable application performance, in general under-performance, and in the worst case, complete mismatch of resource requests and kubelet policies, scheduling a pod where it is destined to fail, potentially entering a failure loop. Exposing cluster level topology to the scheduler empowers it to make intelligent NUMA aware placement decisions optimizing cluster wide performance of workloads. This would benefit Telco User Group in kubernetes, kubernetes and the overall CNCF ecosystem enabling improved application performance without impacting user experience.

Additional information

Type maintrack

More sessions

2/6/21
FOSDEM
K.fosdem
FOSDEM welcome and opening talk.
2/6/21
Community
M.community
How can users of your network be allowed not only to remotely access their files but also to collaboratively edit them? Docker, NextCloud, LibreOffice Online and LDAP are the pillars of the proposed solution. The talk will start describing the context where this proposal was born i.e. the FUSS Project (https://fuss.bz.it). The analysis of the problem will follow along with the development details of the solution and suggested deployment strategies.
2/6/21
Community
M.community
You've just had an idea for a great application but don't think anyone else is interested? You've used your weekends and free time to come up with something that actually solves someone else's problem? You'd love to open source your project so others can use it but you don't know where to start. We'll explain the way from an idea to an open source project using a step by step guide - including links, code snippets, and open source tools you can use to open source your own project.
2/6/21
Community
Benedict Suska
M.community
We are convinced that young people can make an important contribution to the Free Software community. They see many things from a different perspective than adults, which is worth looking at more closely, because new ideas and points of view can be integrated into the development of free software. Since most children and young people are already in contact with the digital world at school, we want to involve young people in the development of software. In our presentation, we will show how we ...
2/6/21
Community
Matt Yonkovit
M.community
2020 was a very bad year for most of us, under the shadow of all that was going on there was a troubling trend we all need to be aware of: the erosion of the classic open source model and values. For years vendors have been slowly chipping away at the freedom and openness provided by open source, this year saw unprecedented changes to how people view and value OSS. From licensing changes to as a service exclusive, what was open is no longer. Is this being driven solely by the cloud? or is this ...
2/6/21
Community
Matteo Bruno
M.community
Do you feel like you don't belong, you don't deserve what you achieved, everyone in your office is more talented than you? Do you have imposter syndrome... too? Imposter syndrome is common across all industries, but the increasing pressure to be successful in IT is taking its toll on employees, affecting more than half workers, me included :) After many years working in tech for a lot of companies (from startups to big corporations) in many business fields, I found a way to overcome self-doubt ...
2/6/21
Community
Bruce Momjian
M.community
This presentation covers the history of various governing structures and why democracy provides superior results. It then explains that open source is a form of democracy, compared to the methods used by proprietary software producers. It covers the many benefits Postgres has enjoyed using an open development model, and how its future remains bright.