Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Towards a true multi-cloud offering to deliver on broken promises (scaleway.com)
36 points by luu on Aug 28, 2024 | hide | past | favorite | 24 comments


Multi-cloud is simple, just don't use anything that's unique to a particular provider. (Losing weight is simple, just eat fewer calories than you burn.)


> Multi-cloud is simple, just don't use anything that's unique to a particular provider.

Using provider-specific services is fine. You only need to be mindful of what it will take to migrate that functionality out of that specific cloud provider,and what work will be required to replace it.

Everything is a tradeoff. The time and cost required to roll out a feature is balanced by the cost of avoiding a tie-in. It makes no sense to spend ages and tons of cash writing a cloud-agnostic implementation when a one-liner of a cloud-specific service does the same job.


We're moving an on-prem Windows app to the cloud, and that's the approach I've taken so far when staking out our approach.

We use Azure blob storage, but just the parts that reasonably maps to say S3. We're using Azure Service Bus, but keep to the AMPQ bits. We use Entra ID, but rely on OAuth and SCIM. And so on.

So far we've not needed special functionality for anything critical, so while it wouldn't be just a flick of a button, the hope is we're avoiding painting ourselves in a cloud provider corner.

Time will tell, hadn't done any cloudy stuff until a year or so ago.


Simple but inconvenient. And it doesn't take much to get sucked into cloud lock-in. As soon as you need to work on permissioning that is slightly less trivial you start to sink.


Could permissioning not be run along the packaging to keep things cloud agnostic?

Some days I wish there was an entire HN devoted to self-hosting and cloud-agnosticity lol.


the problem is there's not an exact 1:1 mapping between clouds, so there's complexity to hold in your head. gcp project, aws account, and azure subscription don't represent exactly the same thing, so at some level you can paper over the differences but you really can't, once you get into the nitty gritty of it, so then you're left building up expertise as a multi-cloud software engineer that knows those subtle differences.


It's a post by Scaleway (from 2022). Now which cloud provider has a huge interest in offering their customers true "multi-cloud capabilities"? I guess the criticism of GAIA-X is fair as it is probably a multi-million euro deadend (or the usual kind of hidden subsidies to the huge players). I can't figure from the blog post what Scaleway does in terms of "multi-cloud offering" apart from the fact they are taking part in EUCLIDIA.


> I guess the criticism of GAIA-X is fair as it is probably a multi-million euro deadend (or the usual kind of hidden subsidies to the huge players).

Projects like GAIA-X are aimed at eliminating the dependency on US companies for critical infrastructure. A few million are peanuts in the scope of shedding a multi-million dependency.


+ Chinese


Well not really. China is not a player in cloud computing. All the established cloud providers are US companies - AWS, Azure, GCP, etc. Alibaba cloud is currently one of the top 5 providers but it's market share is heavily skewed by the chinese market and even then it's only single-digit.

Among the top 10 cloud provides, 7 are from the US and below the top 3 all other players are barely in the single-digit scale.


They specifically mentioned Huawei. The dependency of telecom networks has been discussed repeatedly in Europe. This inititative isn't only about cloud computing.


> They specifically mentioned Huawei.

Not as a cloud provider, but as a secondary shareholder (i.e., is a company really EU if a minority shareholder is non-EU?) and as a reference to Huawei dominating the networking infrastructure market. The last, all on itself, is a strategic concern.


Yes, exactly.


What is the current state of GAIA-X?


There was a lot of discussion recently in another thread https://news.ycombinator.com/item?id=41348659

On their website you can see the "milestones" and there's a "Launch of the new Gaia-X Release (Loire)" mentioned but I couldn't find anything about that "release". So mainly summits, websites and so on.


From Bert Hubert's post: "The actual and official Gaia-X product is a set of standards so we can communicate about (as yet to be written) cloud/data standards. This is two steps removed from being useful - not a cloud, not a standard for a cloud, but a way to exchange authenticated data about standards compliance." - https://berthub.eu/articles/posts/gaia-x-is-an-expensive-dis...

This is such a European way of approaching problems. We were colonized by them, and this kind of vapor work reigns in all their ex colonies.


This is from 2022, any idea what eventually happen with it? Does anyone care about OpenStack any more?


The article doesn't mention OpenStack at all but apparently it's still going: https://www.theregister.com/2024/08/28/geico_vmware_openstac...

Gaia-X appears to have delivered nothing useful. Related newer rant: https://berthub.eu/articles/posts/gaia-x-is-an-expensive-dis...


Proxmox seems to be a way forward.

The IaaS that OpenStack was interesting but I'm not sure that it's made enough strides to keep up to options in the market.

With that being said, I definitely have used OpenStack and wish it was more well known.


> Does anyone care about OpenStack any more?

Yep, it's still widely used in a bunch of orgs (if I'm not mistaken it's popular at telcos, for instance). If you want a full blown private cloud (like the hyperscalers, but in your own datacenters on your own hardware), it's literally the only option. All alternatives just do Virtual Machines and that's it.


"Cloud by Committee" - sort of a predictable outcome


The article seems a little out of sync with the title.


(2022)


"The first few months of GAIA-X’s existence were characterized by a significant lack of understanding about its scope and objectives"

This feels like a commitment to misunderstanding.

A cloud's objective is only to acquire, and deeply retain customers including making switching cost and efforts unbearable.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: