On Fri, Jan 29, 2021 at 1:36 PM Matthew Miller <mattdm@mattdm.org> wrote:
> And if instead your company comes up with some process for _actually_
> managing everyone's active participation, well, okay, fine, but maybe that
> time would be better spent on figuring out how to actually just cope with
> CentOS Stream — or to skip all that and decide that you're at a scale where
> actually paying for RHEL isn't all that terrible after all and all the
> employees can concentrate on doing their actual jobs.
I'm trying to stay out of the weeds on many of the recent re-hashes of
known concerns and known responses to these concerns, but I want to be
clear about this point...
The larger the scale, the less attractive RHEL is. This is a
fundamental problem with the RHEL licensing model. Once the RHEL
licensing cost exceeds around $0.5M/year, you can now afford people to
work on this problem. If you were to exceed $5M/year, you can have a
team of people working on this problem. This is the problem I raised
to our Red Hat sales rep in 2016, and this is the problem that they
never provided a solution for. The discounts provided at scale were
insignificant as factors. This is why "Facebook" (as one example of
scale in context for this discussion) is using CentOS Stream, and not
using RHEL.
Depending on what industry you are in, you are looking at about 0.25-0.5 M/year per person in expenses spread through a company. Those sorts of companies have investors (owners) who want the company to only focus on their core competency which is usually making some other widget than an OS. If your company does not have those concerns, good on them and on you for being lucky to be in that spot. A lot of companies are not and that is where the scaling up proposition does work.