Hello,
I'm a middle technical manager at a largh-ish private agribusiness conglomerate. Not a tech company, but of course we use a lot of tech. My area is platform, i.e. Windows, Linux, middleware, and cloud. Out of our 3K+ servers, probably 90% are currently Windows,but that's somewhat by choice (we could probably run 50/50 with some effort over time). My observation, based on recent hiring experience, is that the highly technical and stoic Windows workforce may giving way to a highly energetic and creative Linux workforce. My question: would it be wise to force that ratio closer to 50/50 in order to better position for the future workforce? Disclosure: we have a few top-talent Linux admins that could take on a lot more servers. Thanks for your answers.
Ask HN: Corporate workforce IT Tech skills for future
1–3 of 3 posts
Re: Ask HN: Corporate workforce IT Tech skills for future
#2Why change a working systems just to keep up with the cool kids? Windows isn't going anywhere. What's the business case?
From a development standpoint, you could host any new development on Linux but the "energetic and creative Linux workforce" who prefer developing on Linux are using languages and frameworks that work on almost any operating system.
I've been developing on Windows for 20 years and until recently I saw no reason to stop. What changed my perspective is cloud hosting. Every project that I either develop myself or that I'm responsible for have resources attached to it and the cost of choosing Windows over Linux is easy to discern just by running a report. The Windows Tax has now become part of OpEx for the development department instead of capex that's someone else expenses.
I still think Windows is easier to manage (and cheaper to find sys admins for) but almost everything I would have used a Windows Server for is a managed service now.
Re: Ask HN: Corporate workforce IT Tech skills for future
#3I've seen corporate IT environments with memory and CPU utilization highly under utilized. People tend to order a lot more compute power than they need because of the bureaucracy involved with getting it from another department and getting it setup. Where I was before it took at least 6 weeks to get a VM, months if you had any decent sized disk requirements.
Do a proof of concept with your top talent Linux admins to see if there are apps they can reasonably fit into containers running Linux. You may find that the server footprint to actually run things is much smaller. Start with something easy they would recommend to put on Linux w/o development time.
There are probably cyclical applications in the environment that go under heavy load at certain times but then are mostly idle other times of the year (e.g. financial apps at quarter/year end). Prime candidates for this type of thing. If you can get some container orchestration framework running, it will make scaling apps up/down easier.
Through a POC you may find a business case that will save the company some money, or you may find it is a wash. Right now you're at a point where you don't know if there is an opportunity there. Worst case is your top talent Linux admins get to have a fun project to work on.