Aeron – Proof of the benefits of open development | Hacker News Reader

Aeron – Proof of the benefits of open development

gresearch.com

65 pointsimpish920812 comments

Not to be confused with the famous Aeron chair, this Aeron is "efficient reliable UDP unicast, UDP multicast, and IPC message transport."

https://github.com/real-logic/aeron

I’ve heard great things about it.

I’m going to be looking at it more closely, when I come up for air, sometime.

I'll add my approval here. Aeron is fantastic, exactly what you need when building a fast messaging infra. Reliable multicast is a really massive deal.

In a previous role I tried reliable multicast via ZeroMQ's support for PGM and it was a fucking disaster.

Aeron does indeed actually work.

Yeah I tried this as well a few years earlier, didn't get anywhere. Aeron I got to work immediately.

zmq is a challenge to get to work for sure. It's super low level.

In my experience, it’s entirely possible to have a “Too many cooks spoil the soup” situation, as well.

I think any project has stages and architectural contexts. Each may or may not benefit from increased participation.

I personally believe that one of the reasons that Linus Torvalds’ projects work so well, is that he’s a 1000X architect and coder, and the parts he did alone, work well enough to form a substrate for the highly participatory parts that extend them.

If a 1X coder did the same, I suspect the same model would not work.

This is all true and good as long as you keep in mind that open source is not a business model. You can make money with oss, but sometimes we have idealistic views of that and the hard truth is that it's very very hard to be sustainable financially.

And with free money drying up I think we will start to see a lot of startups realized that painfully and move to more restrictive licenses.

License bait and switch is a promising business model though, lure in the suckers with a liberal license, just make sure that contributors sign a CLA. Then you can hit them with a change to an onerous license after you have gained traction and a commanding market share.

I can think of plenty of companies that are or have tried this. However I can’t think of many that have done so with notable future success. It mostly seems to be a defensive/reactive move by players having their work more effectively monetised by others than a pre-planned strategy. Boil-the-frog SaaS price increases on the other hand seems to be the default strategy. I am curious to see how well this will work as everyone gets squeezed.

Well, so many words and also such strong words, e.g. "Performance is the key focus. A design goal for Aeron is to be the highest throughput with the lowest and most predictable latency of any messaging system."

I find that hard to believe after taking a glance at their C++ code. Too much std::function for my taste. Maybe it's reliable, but it's probably not lowest-latency. And I'm puzzled that I have to run a Java daemon to interface with my C++ project.

did we need any more proof of this?

Reply on news.ycombinator.com