Amazon S3 down!

This is a great example why centralization is bad! “Half” the internet is down currently. Time for safenet!

https://news.ycombinator.com/item?id=13755673

10 Likes

Looks like a big domino game:-) services collapse one after the other !
That is a really big mess

2 Likes

So, who’s still planning to run a vault on AWS? :laughing:

15 Likes

Truth though, they’ve been pretty darn stable. Stuff happens and when something is this huge, it’s very visible. I don’t think it’s time to write off Amazon; they will fix it, learn from it, and keep doing business. No doubt, many SAFE nodes will run on their VMs, too. Especially during the initial phase, when mining is super efficient, because almost no coins are taken yet.

4 Likes

Apparently centred on US-EAST-1 https://status.aws.amazon.com/

Domino’s is absolutely right @nice, you only start to see how coupled everything is when you read comments like: ‘I can’t change my mouse sensitivity because @razer @razersynapse servers are down :roll_eyes: cc @internetofshit’ and ‘Joys of the @internetofshit - AWS goes down. So does my TV remote, my light controller, even my front gate. Yay for 2017.’ These are from this article in the register.

Hopefully we’ll see more detail over the next couple of days about what caused the issue.

8 Likes

sure, this reminds me an old thread here, about autonomous machines.
The S3 episode, specially those surreal IoT stories, should remind us that, besides decentralizing the network with Safe, we need to keep physical access to things and machines.

And tools for when machines go haywire.

6 Likes

Do you think people love hosting with a company that’s got their own back, but not their customers?

1 Like

Test a vault, yes.
Run it in production, nah…

1 Like

Not exactly surprising, but Netflix had no issues. They run everything off AWS, but there are ways to use even this cloud thing right. Chances are, Amazon’s store used similar methods to fail over.

Sorry, but lol. Did your home internet turn into something more reliable than AWS networking just because they screwed up yesterday? I didn’t think so, either.

Also, isn’t a main idea behind SAFE that it can recover? Your vault may lose some points, but life will go on the moment your network is back up. If anything, vaults are perfect for AWS, because you can just set them up and then forget about them.

1 Like

Totally correct, pal.
Everybody has a different situation. Last week I used AWS because my wannabe data centre was unavailable. By the time vaults are in production, my wannabe data centre will also be in production, hopefully BT OpenReach will have at last proved themselves to be mildly competent and my racks will be worth filling.
So as I said - fine for testing but production, nah…

And I’ll keep the profit - and gladly pay tax on it. Will you be happy to contribute to the already bulging pockets of the Amazon shareholder parasites?
Who knows, given your gushing infomercial above, perhaps you are one of the aforementioned tax-dodging parasites?

1 Like

Well yea safenetforum.org is clearly the place to round up some extra user base for them :joy_cat:

As for the moral aspects of owning shares in a tech company, or the moral aspects of labeling somebody based on the single attribute of their owning a certain stock, let me not worry about those too much here.


But back to the actual topic. Amazon just had a major outage, but when one’s dealing with a large and complex system, these are simply inevitable. SAFE, with its tiny, identical, loosely coupled components will be infinitely more resilient, at least if the design won’t implicitly introduce positive feedback loops.

By the way, Amazon provided resources that companies (that actually cared) were able to utilize to avoid the effects of even an outage this big. Netflix, anyone?

So yea, I do get that when things go south for a big company in a spectacular way, everybody screams it’s unacceptable and they are doomed yada yada, but let’s face reality. On another note, SpaceX is doing fine, too.

Who’s read the reason they gave for the outage?

https://aws.amazon.com/message/41926/

It’s amazing the thing doesn’t go down more often! All caused by a single typo. Blimey.

Another thing that could never happen to SAFEnetwork :slight_smile:

1 Like