HN Explorer

Ask HN: Why not #MultiCloud?

1 points by magacloud ↗ HN
Key factors that are holding you from adopting #MultiCloud?

Some example factors: 1) Unnecessary or I don't mind Vendor-Lockin 2) Security Concerns 3) No single pane of glass / DevOps Overheads 4) Management Decision 5) No source of truth about Multi-Cloud pricing and features 6) Free Cloud Credits on one-provider 7) Others

4 comments

[ 2.1 ms ] story [ 18.9 ms ] thread
abby_cohen_221
What is MultiCloud? Can you provide a link where people can read more about it?
magacloud
Sorry for the jargon.

MultiCloud refers to "using more than one cloud (Infrastructure As A Service) to run their business or production application.

Example: - I could run my production workloads on AWS for their increased availability and range of services they offer. But might be pricy. - I could run all of my QA or pre-production workloads on cost-effective/cheaper clouds such as DigitalOcean, Linode, or OVH.

That way I own my destiny. I build my application in such a way that application portability is maintained.

twobyfour
Even if we were concerned about our current vendor lock-in or displeased with our current vendor, maintaining our site on a single cloud already eats up more devops overhead than our team of three and a half server-side engineers can really afford. And doubling our monthly hosting costs in the name of redundancy (we do already have some redundancy within our existing cloud) isn't a good investment given our business model.

Yes, we had downtime when S3 died last year. So did half the Internet. Our customers understood and our business was impacted more by being unable to use Slack for a few hours than by our site going down for that time.

And if we became dissatisfied with our vendor, we would just make it a project and switch clouds. It would take a month or two, but so would setting up a multi-cloud setup. And once we switched we would be done. No ongoing additional overhead.

And if you think there's any sort of API or unifying panel that would eliminate the additional work of maintaining multiple clouds, I've got this really lovely bridge to sell you....

If we had a business and team the size of, I don't know, Uber, or equivalent availability requirements ... sure, it might be worthwhile. But for a small or medium sized startup, the problems that multi-cloud could solve are the wrong problems to be worrying about.

magacloud
Thanks.

"we would just make it a project and switch clouds. " and "And once we switched we would be done."

That makes a lot of sense.

I'm aware of few startup that use 8 clouds - including barematels from providers that you may not have heard of. Their devops team's expertise overshadows the challenges in managing multiple clouds. it is not practice to expect such level for startups.

>And if you think there's any sort of API or unifying panel that would eliminate the additional work of maintaining multiple clouds, I've got this really lovely bridge to sell you....

Yes - lib cloud APIs is one for Python. Also we at ActOnMagic, tried to support multiple clouds via www.actoncloud.com portal.

So what we are saying is that "Uber scale kind of companies would like MultiCloud?