[alephnerd]

Consolidation is a good move, because it minimizes your overhead from a team management and procurement management basis.

Infra is a cost center at the end of the day.

> I thought the main lesson from the whole VMware fiasco would be that IT departments would not rely on a single vendor/hypervisor in the future

Can you justify spending 2x your hypervisor budget when that same pot of money could be used to hire more engineers who make the product the company is selling?

[karmarepellent]

You are absolutely right that consolidation minimizes overhead. I work in an IT department who relies on VMware and I certainly do not look forward to the prospect of updating all our tooling to work with multiple hypervisors. It works quite well as it is.

And it certainly makes sense in the short term to spend the money on the product instead of on infrastructre. I just wonder about the long term in the context of how we do business. Because on the one hand you have management pushing topics like "risk management" where I have to take responsibility for the most trivial of things in day-to-day operations. And then there is the hypervisor issue where "risk management" goes out of the window and we happily rely on a single hypervisor that could (from one day to the next, more or less) upend all our business.

[alephnerd]

Yep. It is a conundrum, and why a lot of my peers in F500s lead hybrid cloud or a multi-cloud migrations. This way you can auto-scale if needed within your cloud providers (which are easier to migrate from than on-prem due to better DevEx) or return back to on-prem if needed from a cost saving or infra standpoint.

That said, these are questions that are very organization dependent.

[mlrtime]

Why does risk management go out the window? Don't you highlight a single vendor in your BCP plan? Raise the risk, do some quick analysis on how to mitigate that risk and let someone sign off on it.

Accepting the risk doesn't mean 'going out the window'.

[markus_zhang]

However, would it be sensible to keep a small team on a different tech (in this case move everything to N but keep a small V team) so that when shit happens in the other side you have an immediate team to migrate everything to this side?

(I'm thinking about MSFT who maintained a small Windows team when the majority effort went to OS/2 back in the 80s)

[alephnerd]

The MSFT case was R&D spend, but Infra falls under the Finance&IT bucket, where the available capital is much less.

[markus_zhang]

I see. I guess the multi-cloud example you mentioned earlier makes more sense then.