A Hyperscaler’s operating system strategy: working with CentOS and Fedora
329 | Thu 31 Jul 4:30 p.m.–5:15 p.m.
Presented by
-
Michel Lind
https://michel-slm.name/
Michel Lind (né Salim) is a Fedora contributor in various capacity (proven packager, packager sponsor, serving in leadership committees) since almost the beginning; CentOS Proposed Updates SIG co-chair and Hyperscale SIG contributor. He is alao a Debian Maintainer, and has contributed to openSUSE, Fink and MacPorts in a previous life.
In his day job, Michel is a Production Engineer on the Linux Userspace team at Meta, which is responsible for the CentOS Stream deployment on the production fleet.
-
Davide Cavalca is a Production Engineer at Meta on the Linux team. Davide has worked in the systems space for more than 15 years, always with a strong focus towards open source and automation. Davide serves on the CentOS Board of Directors, co-chairs the CentOS Hyperscale SIG and actively participates in a number of other SIGs to drive the project forward. Davide is also involved in Fedora, where he sits on the EPEL Steering Committee, and has helped drive the development of several major distribution features. Davide also sits on the Asahi Linux Governance Board and is actively involved in the project, where he helps develop Fedora Asahi Remix within the Asahi SIG in Fedora.
Michel Lind
https://michel-slm.name/
Abstract
When an organization's Linux deployment gets large enough, a curious emergent property appears - the incentives start to favor developing in-house Linux expertise, as opposed to outsourcing operating systems support to external vendors.
At the same time, given the scale involved, such organizations tend to prefer having a stable base to build on - thus the appeal of slower-moving enterprise distributions, except with in-house customizations on top.
In this talk we are going to discuss some of the choices we made at Meta for our Linux fleet, and the thought process behind that. We hope that organizations in a similar situation can benefit from our experience, and that community members whose interests are aligned can benefit from our contributions and consider participating in the community projects we are involved in.
When an organization's Linux deployment gets large enough, a curious emergent property appears - the incentives start to favor developing in-house Linux expertise, as opposed to outsourcing operating systems support to external vendors. At the same time, given the scale involved, such organizations tend to prefer having a stable base to build on - thus the appeal of slower-moving enterprise distributions, except with in-house customizations on top. In this talk we are going to discuss some of the choices we made at Meta for our Linux fleet, and the thought process behind that. We hope that organizations in a similar situation can benefit from our experience, and that community members whose interests are aligned can benefit from our contributions and consider participating in the community projects we are involved in.