dc2447

Try ordering a couple of thousand Rome servers from Dell right now and you might be surprised.

8x is close to the top end, somewhere between 3x and 10x depending on a number of factors, mostly scale.

For the record I factor in:

  - Hardware depreciation (36mo)
  - Power
  - People 
  - DC rent + power
  - software licences / support

for on premise.

What we see now is that compute is dropping in price for on prem every year and density is improving. AMD Rome brings incredibly bang for your buck when buying at significant scale.

But it's not comparing Apples with Apples.

It's virtually impossible to accurately factor in the opportunity cost of doing all this yourself but you can potentially hire a bunch of engineers with the savings of going on prem, ymmv

You can never ever recreate the developer experience on prem regardless of your scale, on you can tell if on prem is good enough

It's difficult to put a value on being to pay as you go or suddenly be serving workloads out of a geo close to your users in Cloud where on prem there is always a lead time

Finally whilst the developer experience is better suddenly having to deal with new challenges takes a while to adjust in Cloud, outages out of your control, non predictable performance, poor support, no access to your hardware

TLDR: Cost is hard to define and isn't a zero sum game

Curious if you tested POC for simple COE?

Melatonin is now available from Chemists. You need to do a short online consultation but other than that it’s straight forward to buy

Looks great, would love and use something like this for GCP

Virtualised control plane is totally fine. Next step is containerised OpenStack.

Pros:

- economy of scale, if you are large enough OS is very cost effective - OSS - No vendor lock in

Cons:

- Complexity - Lack of features and higher level abstractions v public cloud

Definitely ahead of their time, Openstack wasn't formed until 2010

Absolutely. You get it.

It's frustrating to read people comparing container technologies or worse container schedulers with OpenStack. Apples and oranges.

If you are planning on running containers on premise you need something from managing the IAAS side of things. OpenStack is excellent at IAAS and is the foundation for running higher level abstractions and services on top.

In HTTP2 the opposite is true.

Single domain will help you with performance

There are a few factors here.

Depreciation: is typically three years for hardware purchased rather than leased. At the end of 3 years the item has zero value thus can be replaced.

Useful life: is hopefully longer and up to five years. So an item is without value but can still be used. More typical with networking hardware than compute.

The other poster who mentioned power consumption is totally correct. It can make sense to renew hardware even before it has fully depreciated in order to get better datacentre density and lower power draw.

This is what is driving most refreshes in my experience.

I love when the guys from Etsy present at conferences and they start with that joke:

We are a monitoring company with a side line in being an online marketplace