4 ms·

I work at an IoT company, not for home automation, but GPS vehicle tracking and temperature monitoring of commercial fridges, among others. I agree that requiri

I work at an IoT company, not for home automation, but GPS vehicle tracking and temperature monitoring of commercial fridges, among others. I agree that requiring connectivity for essential home appliances and door locks is downright foolish. My mum saw this news, and asked me "Did this affect your client's IoT system, Peter?" Here's my reply. Cloud computing is just somebody else's computer. https://xkcd.com/908/ https://xkcd.com/908/ Amazon sells time on their computers, and calls the service "AWS". Actually, Amazon have 24 computers•. https://aws.amazon.com/about-aws/global-infrastructure/ https://aws.amazon.com/about-aws/global-infrastructure/ One of their computers is called US-east-1. That is the one that broke. The other Amazon computers (e.g. EU-west-1 in Ireland, AP-northeast-1 in Tokyo) were all fine. Some of our programs run on AWS, but we are using the computer in Sydney (AP-southeast-2). So our services weren't affected. Only the people using US-east-1 were affected. This is the default computer for many people in the US. Even big companies like Google and Facebook sometimes use AWS computers. We also use DigitalOcean's computers for some other clients. That's like AWS but cheaper and slower. I helped copy files from one computer to another. That's called "DevOps". Peter • Technically it's a huge server farm with many computers, but they all have the same basic inputs and outputs: power [0], Internet access, air conditioning. Anything in the same physical location could be affected by earthquakes, fires [1], power cuts, Internet connections going offline, etc. This is why daddy likes to keep an off-site backup, for safety. The government spies on people by seeing what is happening on Amazon's computers. Wikileaks [2] has more detailed information about who is doing the spying in each place (usually a company around the corner). [0] https://www.carolinajournal.com/news-article/amazons-wind-farm-links-nonexistent/ https://www.carolinajournal.com/news-article/amazons-wind-fa... [1] https://www.datacenterknowledge.com/archives/2015/01/09/fire-amazon-data-center-construction-site-ashburn-contained https://www.datacenterknowledge.com/archives/2015/01/09/fire... [2] https://wikileaks.org/amazon-atlas/map/ https://wikileaks.org/amazon-atlas/map/

moduspol 6y ago
> Anything in the same physical location could be affected by earthquakes, fires [1], power cuts, Internet connections going offline, etc. It's worth noting different AZs within the same region are supposedly architected to have their own redundant power and Internet connectivity. AFAIK a single fire shouldn't take out more than one, either. If a region-wide outage is caused by any of those things, it's because something else isn't working as expected. And this region has had region-wide issues before. Still, boiling cloud computing down to "somebody else's computer" is an oversimplification I see too often from cynical techies. It's like calling a restaurant "somebody else's kitchen." Yeah, it is. But it's missing the point to frame it that way.

moduspol 6y ago

> Anything in the same physical location could be affected by earthquakes, fires [1], power cuts, Internet connections going offline, etc. It's worth noting different AZs within the same region are supposedly architected to have their own redundant power and Internet connectivity. AFAIK a single fire shouldn't take out more than one, either. If a region-wide outage is caused by any of those things, it's because something else isn't working as expected. And this region has had region-wide issues before. Still, boiling cloud computing down to "somebody else's computer" is an oversimplification I see too often from cynical techies. It's like calling a restaurant "somebody else's kitchen." Yeah, it is. But it's missing the point to frame it that way.