[thepossum]

Many years ago in an earlier professional existence near the early days of the world wide web I was lucky to be exposed to a lot of the intricaies of DNS

recently I had been getting concerned about the concentration of DNS to a few major corporations

Can anyone explain how this risk was not known by others

I could not possibly imagine what could go wrong with all our services including DNS being based on a few major platforms - and how many other previously trusted entities outsource or host all their services etc

Strangely others are able to charge big consultancy fees and all I have is schadenfreude - and scahdenfreude in my case is quite sad and empathetic so not worth much

PS small shoutout to those who taught me the basics of the internet all those years ago - I cannot make head nor tail of DNS files these days but I am sure the principles are the same. I get Chat GPT to help me interpret such things these days, and I am lucky enough to have a very professional web service provider who never seem to cause me any trouble. And my requirements are so simple I do not need to use all those wonderful services available now. I did get very close to speccing myself up an Amazon server one day. And I may have been asleep when it all happened and I have no idea if I was affected at all. I wonder what Dan Kaminsky would have said. Sorry for rubbing it in but I heard half the internet was affected by a small DNS issue

[pilowsky]

Remember the days when we could fix our own cars?

[barmar]

DNS is not concentrated in a few corporations. It's a distributed system where many companies are involved. The most sensitive parts are the root and top-level domains, and they're each spread over about a dozen companies.

Are you talking about yesterday's AWS outage? While DNS was involved in that, it was just Amazon's DNS, not the DNS system in general.

The problem yesterday is that most Internet services make use of cloud services, and there are only a few large-scale providers of this: Amazon, Microsoft, and Google. A part of Amazon's cloud service went down due to a problem with their DNS configuration.

When operating a web site that has to handle thousands or millions of concurrent customers, it's simply not feasible to DIY.

FYI, BBO operates using a mix of self-managed servers and AWS. We've been gradually migrating to AWS because it allows easy scalability. For instance, many of our robots run on AWS so we can deal with varying load better.

[pescetom]

 barmar, on 2025-October-21, 10:07, said:

FYI, BBO operates using a mix of self-managed servers and AWS. We've been gradually migrating to AWS because it allows easy scalability. For instance, many of our robots run on AWS so we can deal with varying load better.

Could AWS maybe even handle the load of me running a trick taking simulation with Dealer every now and then?

[Gerardo]

Why the Web was Down - Explained by a Retired Microsoft Engineer