Amazon cloud outage fuels call for Europe to limit reliance on US tech

https://www.politico.eu/article/aws-amazon-web-services-outage-europe-limit-reliance-us-tech/

13 comments

  1. It’s kind of funny to remember that AWS was built on the internet, which came from Arpanet – where Arpanet was *designed to be decentralized so as to be resistant to nuclear attack.*

    Yet here we are a few decades later, where corporate monopolies have utterly destroyed that decentralization.

  3. Wow, seems like the wrong message was heard.  The message should have been “don’t rely on 1 provider”.  Where that provider exists is kinda an orthogonal issue.  

  5. it’s less about this being specifically US tech and more about clients and their big fat service providers putting all the eggs in one basket.

    Don’t rely entirely on one service provider for all your needs, it is kinda silly to sit and expect these hugely complex global operations to not have the occasional issue.

  6. Sir, the latest outage has revealed that our entire system actually has 3 large red blinking weak spots. If any intrepid adventurer were to fire bolt arrows or throw a boomerang at them…

  7. That’s now how Amazon works lol. Both the Chinese, US and UK governments use AWS with local servers and even physically isolated networks

  9. And that’s as far as they’ll get. Calling for alternatives.

  11. I work as a developer for a company that has proper disaster recovery.

    It’s worth it for us because we are a payment service provider that moves over $4B USD per day. I couldn’t imagine being multi region or even worse multi provider being worth it for 90+% of companies.

    We distribute between three availability zones per region in two regions.

    It is insanely complex and also expensive to not own develop but also run and maintain this setup.

    Our AWS bill alone is over $100M/year and it would probably be about $30M/year cheaper to not have multiple regions (redundant data and over-scaling of services to handle failovers is expensive)

    We spent tens of millions in one-time development cost when we implemented this. We spent millions per year just maintaining and routinely testing our disaster recovery.

    Every new project we implement or feature we add is a lot more complex.

    I honestly can’t imagine the majority of companies being able to implement multi region correctly or in a way that makes sense from a cost perspective.

  12. He’s not completely wrong – it is unhealthy for the internet to be too concentrated in one company – but let’s be real: this is more about wanting an internet giant in Europe than anything else. 

    AWS is pretty damn reliable, but not perfectly reliable. Moreover there’s nothing to indicate that European cloud providers would be any more reliable. 

    Literally right now [OVH cloud is having degraded performance with their North American backbone.](https://status.us.ovhcloud.com)

    [Hetzner’s S3 equivalent was failing a week ago.](https://status.hetzner.com/incident/5782c136-a5c7-4849-beaa-0f50f3b9f192)

    It happens. 

  13. The only true measure here is to stop devices to require constantly being online. Seriously. It does not matter if the server is in the US or in the EU from a technical standpoint. If it is down and your device goes haywire or becomes a hazard to humans, you are doing it wrong.

    Limit the need for internet. Do things locally. Sync once a day if you must. Turning a smart feature on or off in my home should not require a round trip over the Net.

Comments are closed.