Praise Amazon for raising this service from the dead

For years, Google has seemingly indulged a corporate fetish of taking products that are beloved, then killing them. AWS has been on a different kick lately: Killing services that frankly shouldn't have seen the light of day. The first was Amazon Sumerian (tagline: "You're making that service up, right?"), which was followed …

[Dwarf]

Great news

When people can't trust that a product will be here next week, they will get all touchy when something new and shiny comes out.

Additionally, with the current approach of "everything as code", a modern using source control system is a critical component of the build pipeline, so why would someone like Amazon need to rely on an externally hosted service ? That would make no sense whatsoever.

At least AWS seems to have listened to someone and fixed what they got wrong. Other vendors might want to take note.

[cyberdemon]

"Praise Amazon for raising this service from the dead"

Just stop NOW with the clickbait headlines. You could have said "Praise Amazon for raising CodeCommit from the dead" instead of "this service" and saved those of us who couldn't care a jot the trouble of clicking on it to find out which effing service.

[FIA]

Re: "Praise Amazon for raising this service from the dead"

LOL, I think that ship has long sailed.

We live in a world where people expect quality journalism for free. That means advertising. That means clickbait.

Especially on a site like el Reg where the target demographic is probably more aware of things like adblocking than most.

[Hawkeye Pierce]

Re: "Praise Amazon for raising this service from the dead"

To be fair "CodeCommit" is close to being about the only AWS service whose name gives a broad enough indication of what it does!

[tiggity]

"Snowball Edge ("Please never Google either of those words")."

.. does it say something about me that I know the slang meaning of snowball & edge without using a search engine (don't like the use of "Google" to impky running a search, there are alternative search engines)

[FIA]

there are alternative search engines)

Are there? I thought they all got hoovered up in the early 2000s?

[yoganmahew]

Large files are in git precisely because it's git - they exist, can be monitored for changes, can't be lost. Documentation as code (asciidoc, markup and the like) mean the documentation system git has to take all sorts. Once you accept that, size and content type limitations becomes arbitrary. Big E has been let down by big Tech for Enterprise grade document management forever. Everyone hates Sharepoint. Everyone.

[Nate Amsden]

I've never used SharePoint (though annoyed that so many orgs seem to want to use it as a general file storage not the CMS that is is, when there are cloud things like Azure files that are built for file shares - note I haven't used Azure anything either), but at my first SaaS gig two decades ago we did use Visual Source Safe For version control of our internal documents. Fortunately I never had to deal with the fallout of that. There was one person we had who knew VSS and was tasked with recovering our VSS on the many occasions it decided to corrupt itself.

[doublelayer]

But have you ever found someone who decided that was a good thing but didn't know about LFS? I have. Size of the files you were going to use: about 2 GB. Size of stuff you'd get when actually cloning it, more like 45 GB. That was annoying to clean up. And just because LFS exists doesn't make this all better, because there are times when massive objects don't need to be in source control and costs to having them there which might be lower or zero if they were somewhere better suited to them.

[JimmyPage]

Will they pay for that wasted effort ?

So not that sorry then.

[Always Right Mostly]

Even with El Reg's audience, this article seem pointless at best and some kind of "Please give Amazon some love' bullshit.

[FIA]

Even with El Reg's audience, this article seem pointless at best and some kind of "Please give Amazon some love' bullshit.

It's about long termism.

If you're developing 'enterprise' apps then they're generally long lived. Often much longer than the services offered by the various cloud providers. Companies don't want to have to re-write everything every few years if it fulfills a busines need. You don't rebuild your offices every 5 years, it should be the same for well written software.

If you're building this kind of stuff then considerations like 'Will google just randomly remove this service the second we go live?' (The answer is usually 'Yes'), become important.

I've used two of the 3 big cloud providers over my time and would activly push against Google because they do like to kill stuff on a whim and I wouldn't want to base my business around a company that does this.

AWS, as an example, deprecated a service we used at a previous company. They gave us many years notice, and I think the service still worked long after we'd moved away from it. This meant we could plan the deprecation in along with all the other work we had to do on the product, which, as a small team really helped. We simply didn't have the capacity to be constantly updating the software for structural changes and also add and maintian new features.

This approach from AWS (and googles reputation) have driven descisions I've made in my career since.

[Rich 2]

No

I don't care what goody-goody things Amazon have done. I still think it's an abhorrent company

[FIA]

Re: No

It's a big company... it can have both good and bad aspects.

Also.. remember... abhorrant companies don't get where they are by magic.... we all like the shiny shiny...

"What? They make people shit in bags... that's terrible.... OOOO BLACK FRIDAY...."

Abhorrant companies get where they are because by and large people, en masse, don't really care.

"I'm part of the problem??? Don't be silly... I just like to buy cheap stuff without considering the social ramifications as it's getting expensive out there..."

<Wanders off to wait for the Amazon delivery man to bring stuff>

[cd]

They need fodder for a vibe AI.

[Claptrap314]

It's not just auditing

My last company was typical--Github for code, AWS for the servers. Which meant that a major Azure outage would take down development work, including anything needed for dealing with an outage.

That's a real operational risk. And security.

Sadly, I couldn't get them to start using HaveIBeenPwned, so the conversation about getting off Github didn't even get started...

[Chris Robinson]

First I Googled 'Corey Quinn':

Corey Quinn is a Chief Cloud Economist at The Duckbill Group, known for his work helping companies manage and optimize their cloud spending, particularly with Amazon Web Services (AWS). He is also the author of the weekly newsletter "Last Week in AWS", where he summarizes and analyzes news in the cloud computing space. Additionally, he co-hosts the podcasts "Screaming in the Cloud" and "AWS Morning Brief".

Then I Googled 'shill' and got a match.

Come on El Reg, you need to do better.