With cloud reliability problems getting attention, are organizations likely to move compute back on-premises or into colocation facilities? What would make that worthwhile?
Could cloud outages push workloads back on-prem?
Answers
Outages alone probably won't settle it. On-prem systems can go down too, and power or internet failures can take out a whole site. A generator and redundant connections help, but there are still other dependencies in the chain.
The useful question is which failure modes you can afford to manage yourself. Building infrastructure that's resilient and accessible from everywhere is a big job.
Reply to this answer
Look at it workload by workload. A virtual desktop fleet with predictable, heavy usage might be worth comparing against owned hardware over its refresh cycle. Include staffing, maintenance, power and security, not just the server quote versus the cloud bill.
Hardware prices can make the upfront spend hard to swallow, though. Cloud providers buy hardware too, so rising costs there may eventually show up in their prices.
Reply to this answer
Hybrid is a practical middle ground. Keep workloads on-prem where the economics and dependencies make sense, and use cloud services where they're a better fit. Email and public websites are examples some organizations may prefer to leave hosted.
That doesn't automatically give you a backup location for everything. You still need a plan for what can move and what it would take to run it elsewhere.
Reply to this answer
Colocation isn't the same as handing the problem to a cloud provider. You may put your equipment in someone else's facility, but you still need people to maintain and secure the systems. Price the people and ongoing work before treating it as an escape from cloud bills or outages.
Reply to this answer
Add your answer
Answers are reviewed before publication.
Zelquari-3v:
Even if your own services stay up, an outage affecting other providers can still stop you reaching customers or partners.