AWS CodeCommit | Hacker News Reader

AWS CodeCommit

aws.amazon.com

136 pointsrjsamson60 comments

In my opinion, GitHub has demonstrated how a good monopoly[1] can operate. They're quick to respond to problems, they act ethically for the most part, their pricing is a far-cry from the type of gouging most monopolies get away with, and their product continues to get better and better. GitHub also makes it easy for new developers to get started with Git, and their UI has many awesome features that keep getting better.

As Peter Thiel would say, because they have a monopoly, they have money to spend on things (e.g., improving their product) other than competing themselves out of business.

All that being said, where is the market gap that AWS seeks to fill, other than hosting of large-sized files (scientific / big data applications, etc.), and perhaps the user management, which I feel is a bit outmoded and cumbersome?

1. I think it's fair to say GitHub has a monopoly on paid Git hosting

I don't think the only store in town selling widgets has a widget monopoly unless there's something preventing another store from selling widgets. Until then they're just first-movers.

But I will agree about Github. They had time to build lock-in and didn't.

A monopoly exists when a specific person or enterprise is the only supplier of a particular commodity[1], so it doesn't really matter that another provider could join the market.

To be fair, GitHub is really part of an oligopoly, comprised of themselves, BitBucket, and GitLab, but the winner of a majority of market share in this situation is typically referred to as a monopoly, in the same sense that Google is referred to as a monopoly, despite the existence of Bing, Yahoo!, and other search engines.

For context, according to Compete, GitHub had 24 times as many unique visits as BitBucket last month. GitLab has about half to one third the traffic of BitBucket.

1. https://en.wikipedia.org/wiki/Monopoly

From your reference, I think this supports the other meaning:

> The verb "monopolise" refers to the process by which a company gains the ability to raise prices or exclude competitors.

As I see it, without that ability to exclude, it's not a monopoly.

In a market without lock-in, or lock-out depending on your PoV, I can start selling widgets at any time and will do so whenever your profit margin gets too sweet. You're competing against me even if I'm not actively selling yet.

> 1. I think it's fair to say GitHub has a monopoly on paid Git hosting

I don't know... depends on how precisely you want to define "monopoly". Bitbucket are certainly popular in some circles. We use them here at Mammoth Data, as their pricing model is more economical for what we do.

Yeah, this is a fair point. Anecdotally, most of the agencies I've worked with use Bitbucket due to their pricing being based around users as opposed to the number of private repos - the lock-in becomes even tighter if they're running JIRA. Github's private repo limit even on organizational accounts can become constricting pretty quickly.

The one thing that is appealing to an "AWS shop" and something that GitHub and other services have a harder time emulating is the integration with IAM (user authentication and provisioning) and the CloudTrail (audit trails). For any reasonably sized organization, ensuring that users get provisioned / deprovisioned is a tricky, error-prone task. Having IAM as a single toggle is nice. Using their SAML integration into something like Okta, OneLogin, or other IaaS-providers ... even better. Getting user activity into centralized logging via CloudTrail (Splunk, Sumo, ELK, et al.) is best practice for security teams. AWS makes all of this pretty easy, but at a price.

That's not just it. They are planning to add post-receive hooks. Imagine doing a commit and writing a post-commit hook to auto-deploy to appropriate hosts or spawn up an EC2 instance to run integration tests.

I don't see myself switching to this, but I'm glad it exists. After Google Code was shut down, it was clear that GitHub would remain miles ahead of any competition, making it the default choice for most orgs. That worries me quite a bit.

Agreed, and I'm glad companies like Amazon with deep pockets are deciding to jump into the pool with github, as adoption of these competing products are probably going to be very slow short of some sort of event like what happened with SourceForge.

I mostly use GitHub these days (at work), but I think Bitbucket is pretty similar in terms of features and may even be better in some ways (e.g., I like Bitbucket's design better, and they give you unlimited private repos). I don't think you could say GitHub is miles ahead of Bitbucket, except maybe in terms of users.

Regardless, I agree that more competition is good.

I meant miles ahead in terms of users, dev mindshare, and that special community feeling.

Bitbucket seems cheaper than CodeCommit if you have repositories under 2GB. However, once you get over this size, for example with lots of binaries, Bitbucket stops being an option and CodeCommit looks very interesting.

GitHub doesn't explicitly states that they will disable pushing once the repository size crosses a threshold, but I don't think that they will allow multi-gigabyte repositories either.

We have a soft 2 GB limit: https://help.github.com/articles/what-is-my-disk-quota/

In reality this has been very much a non-issue. Users we contact are generally okay with not bloating repos, and we point out Releases / Git LFS as alternative places to store binaries / large assets.

Pricing: https://aws.amazon.com/codecommit/pricing/

$1 per user/month, with plenty of free storage (10GB) each. 2k git requests/month sounds a bit low though if you're doing anything automated. There is a good free tier.

And again, no Mercurial. Have we lost?

Hg seems to have lost the SCM war (of popularity) unfortunately - i quite like hg.

Same here.

I think if it was not for the convoluted Revlog [1] format (which makes it impossible to do a clean-room reimplementation of Mercurial Core), we could see a bit more of adoption by having third-party libraries to interface with Hg repositories. It _might_ have been one of the reasons Git, not Mercurial, was added to TFS [2].

1: https://mercurial.selenic.com/wiki/RevlogNG

2: https://hglabhq.com/blog/2014/1/17/mercurial-support-in-tfs-...

>if it was not for the convoluted Revlog [1] format (which makes it impossible to do a clean-room reimplementation of Mercurial Core), we could see a bit more of adoption

Alternatively, we could see a bit more adoption if it were available under a more permissive license than the GPL.

Git is also under GPL, so I am not sure how much of an argument that is.

You can theoretically reimplement Git Core by only referencing the documentation, which, according to some sources, does not "infect" the new code with GPL.

could you not have replaced git with hg above, and the argument still holds? Revlog's format itself isn't under copyright (and i doubt it exists, but only with a patent can you stop a _format_ spec from being used). If you wrote a parser for the revlog format, you aren't "infected".

Also, hg plugins are much easier to write (being in python and all), than messing around with git's internals imho (ala, facebook's modification to hg to make it massively scalable https://code.facebook.com/posts/218678814984400/scaling-merc..., which would've been almost impossible in git).

It's not realistically possible to write a Revlog parser by only using whatever documentation is available; you _will_ have to reference the Hg source code, which will turn your project into "derived work".

I'm afraid so. :(

I tried using Mercurial on Kiln for private, personal projects. When I wanted some of those to graduate to public, GitHub projects it was a pain in the ass. So I switched to Git for all Kiln projects. And Kiln even has either/or magic.

Damn.

My question is how long is GitHub will remain a git-only platform?

It seems only natural to me that they should branch out. (heh... branch)

Having invested so much into Git-oriented projects and whatnot, I think they will remain Git-only till the heat death of the universe.

And then again, Hg repositories on _Git_ Hub?

Hey, Rap Genius bought genius.com! :)

Unless one day they rebrand their name like "genius" did, maybe codehub or something.

I think it's incredibly unlikely Github will ever support another SCM, unless a new one becomes available which all the devs switch to.

They've supported Subversion for a few years now: https://help.github.com/articles/support-for-subversion-clie...

They (kinda) support subversion already.

https://help.github.com/articles/support-for-subversion-clie...

This is a shame since Mercurial is so much better than Git. Git is downright painful in comparison.

It's actually discouraging to see so many developers choose to work with something so problematic as Git, to be blunt.

Amazon seriously needs a new front-end dev.

Amazon going into Github's and Gitlab's space. I wonder how they will react to this and how this might affect them.

I don't think they need to worry; AWS continues its potato quality UX with CodeCommit. Unless you're hard core into using git in terminal for practically everything I can't see this taking much if any business from the existing big players.

uhhh....yeah.....I would say the vast majority of devs in the non-MSFT space use terminal more than 99% of the time...

using the terminal for staging lots of code, or comparing diffs is painful. i personally use sourcetree for that or open a PR on bitbucket/github

> I would say the vast majority of devs in the non-MSFT space use terminal more than 99% of the time

That's a pretty bold claim. Do you have any data behind it?

It's only anecdotal but in my experience it's almost the exact opposite; having worked in multiples of both MSFT and non-MSFT shops it seems like almost every one uses a UI to handle git. I don't blame them; it's not always easy visualizing all of the changes (in my opinion at least) just through a terminal window.

Sadly no, it was a terribly anecdotal claim with no official evidence to back it up. I guess the point behind my claim is what do you mean by "working with git"? To me it is about checking code in, diffing, etc... which most people I've ever worked with do from the terminal. Occasionally I'll see folks want a gui for diffs (not really surprising since it's kinda fugly from terminal), but do most of the "work" from term.

So I suppose I shall digress and admit I committed a total failure at an internet statement on a decently technical forum.

in the words of Homer....

doh!

Amazon is just releasing their own internal GitFarm product that's been around for a while, rather than building something new to "go after GitHub."

Amazon has clones of several major services that haven't really gone anywhere...I don't think it will affect github all that much.

GitHub's power is their UI and UX. Amazon is the exact opposite. That's why Amazon charges no more than $1 per user in teams, while GitHub can charge $25 for 10 repos at any team size.

We welcome Amazon coming into this space, there is room for multiple code hosting services. At GitLab we'll focus on the needs of our open source community: a complete, user-friendly and scalable product at a good price. We hope that other cloud platforms will consider running GitLab for git hosting. It is not a reaction to Amazon, but we did announce our seed round today https://news.ycombinator.com/item?id=9857901

I am logged in to AWS console and don't find CodeCommit. Am I missing something?

us-east-1 only perhaps?

Oooh, Jeff Barr states that it'll take a few hours to appear on the web console: https://twitter.com/jeffbarr/status/619155748392583168

I'm a little confused. Why would I use this instead of Github?

Because this'll integrate heavily with Amazon's other offerings in time. I'd expect click-to-deploy for Node apps to wind up in the interface fairly quickly.

Opsworks already has click-to-deploy. I imagine Elastic Beanstalk does too.

Really, I think the benefits of amazon git hosting is that they'll be more resistant to failure than github is currently.

Also I'd imagine companies with existing IAM account management structure and large teams would find this product somewhat attractive.

Lambda is pretty much there.

I view this more like Heroku and how they have git servers. Sure, Heroku has git servers but isn't trying to be GitHub (a web interface over git and development communication in general). This appears to just be a convenience for deploying your code.

Also, given Amazon's awful reputation for building UIs that people actually want to use, I doubt they could build something to compete with GitHub anyway.

> I'm a little confused. Why would I use this instead of Github?

I haven't actually done the math yet, but it looks like it may be more economical for large numbers of private repos. For Open Source projects, there's probably no reason to use this, but for companies and other organizations who want Git hosting for private code, it might make sense. And the IAM integration actually could be useful if you're already a heavy user of Amazon AWS services.

The main thing, at least for now, is IAM.

It looks like we're talking about these new AWS features whether we want to or not..

We're talking about them because HN users submitted the stories and other HN users voted them up.

Strange they both have the exact same number of upvotes (about 1 per minute too) and one was posted by an amazon employee. Not like there's ever been a corporate shill thread on HN...

edit: and look, we have a new user who just created an account to say they were "eagerly waiting for this!" on the other thread:

https://news.ycombinator.com/user?id=pragar

> and look, we have a new user who just created an account to say he was waiting for this on the other thread

Maybe he or she was waiting for this? Why does everything need to be a conspiracy to promote corporate America? Granted you could be right but I'd bet on the simple reason first (someone submitted it and it was upvoted).

Yeah, you're probably right. Just a funny coincidence.

They have the same upvotes because they were submitted at the same time. They were submitted at the same time because they were both just posted on the AWS blog. People are watching the AWS blog right now because there's an AWS conference happening right now, and new product announcements happen at conferences. There's a large segment of HN users that are interested in AWS and other infrastructure topics, and would upvote both stories. I'm one of them; I upvoted them from the /new queue when both were at 2 points. I don't work for Amazon and never have. You see the exact same pattern during Apple's conferences, Microsoft's conferences, etc.

So rather than just down vote this comment and the others, I'm going to point you to this guideline:

"Please don't submit comments complaining that a submission is inappropriate for the site. If you think a story is spam or off-topic, flag it by clicking on its "flag" link. (Not all users will see this; there is a karma threshold.) If you think a comment is egregious, click on its timestamp to go to its page, then click "flag" at the top."

If you have an issue with the content, commenting on it is not the solution. In fact, by commenting like you did, you are doing the wrong thing. You are becoming the problem you want to avoid.

Please use the tools available to you. Thanks!

Reply on news.ycombinator.com