6 ms·

I would suggest running your own half rack or full rack at a local colo facility, and integrate amazon or azure into your external scaling strategy using APIs t

I would suggest running your own half rack or full rack at a local colo facility, and integrate amazon or azure into your external scaling strategy using APIs to spin up instances on an as-needed basis. You should always be in control of your core databases, and never allow an external vendor to be your fallback position or failsafe. The odds of both Azure and Amazon having an outage at the same time are very very very small, especially if you span your instances across multiple zones. Allow your local servers to be a fallback to the cloud, so that when (not if) there IS an outage, you have a controlled set of servers to rely upon.

chrislomax 14y ago
It's really hard to make the decision. We priced up co-lo at our local data centre and the costs are good and would save us money, it's just a massive headache. Network switches, redundant lines, internal vlans. They are all a pain in the ass, the only person at current who would manage that is me and I need to be developing really. Obviously you don't know our company structure so I'm not having a go there! We have gone through this process and pricing already and it's ideally an area I want to stay clear of. I would prefer the responsibility to be somewhere else. It's really hard making any decision, I just want a reliable solution where everything just works. Whenever I seem to settle on something then something happens at one of the companies and I'm deterred. I do appreciate the comment though and if I do ever get the day when there is someone dedicated to sorting this then I'll be dogging them to get this solution in place, the mythical 100% is the end goal