X : Did you hear AWS was down?
Me : Didn't notice, been busy with other things. Was it a region?
X : US-East-1 had problems.
Me : Ok. Last long?
X : A couple of hours?
Me : And I assume we discovered that there are companies out there still not building across multiple regions?
Me : Didn't notice, been busy with other things. Was it a region?
X : US-East-1 had problems.
Me : Ok. Last long?
X : A couple of hours?
Me : And I assume we discovered that there are companies out there still not building across multiple regions?
X : I don't know.
Me : Probably. This was good practice almost a decade ago. It wouldn't surprise me.
X : Delta had problems.
Me : Delta Air Lines? Like the outage in their own data centres about five years ago which grounded flights for three days?
X : I didn't know that.
Me : Probably. This was good practice almost a decade ago. It wouldn't surprise me.
X : Delta had problems.
Me : Delta Air Lines? Like the outage in their own data centres about five years ago which grounded flights for three days?
X : I didn't know that.
Me : Data centre outages which lasted days if not weeks used to happen quite often, especially in the old days of home grown data centres. A region of AWS being down for a couple of hours is a pain but ... well, multi-region was good practice long ago. When did the outage happen?
X : Were you not online?
Me : Yes but I didn't notice. I certainly was using services that consumed AWS but my focus was not on keeping up with news.
Me : Yes but I didn't notice. I certainly was using services that consumed AWS but my focus was not on keeping up with news.
Hmmm, AWS console ...
Ok, this is spot on by @lizthegrey -
https://twitter.com/lizthegrey/status/1468993507511599104
Ok, this is spot on by @lizthegrey -
https://twitter.com/lizthegrey/status/1468993687044575239
Eventually these sorts of outages (which have already become far less common than the past) will be extremely rare as multi region serverless (across all services) becomes the norm. You won't even notice the occassional outage.
X : Do you not think there are risks?
Me : Systemic failure? Of course, that's why I talked about the need for a competitive market based upon open source and open standards back at OSCON in 2007. But there was a problem.
X : Which was?
Me : Systemic failure? Of course, that's why I talked about the need for a competitive market based upon open source and open standards back at OSCON in 2007. But there was a problem.
X : Which was?
Me : A whole army of misguided that didn't understand strategy created collective prisoner dilemmas (OpenStack) or failed to react (giving AWS a 7 year head start) or turned up late and focused on the wrong part of the stack (Kubernetes).
Me : We're now in the long game ... Alibaba seems to be playing that by co-opting AWS as the standard because it is the standard. Others (after 12 years of being obvious) have adopted S3 etc as the standard. So, we're getting there slowly but need to move up the stack.
X : So, systemic failure is ...
Me : ... a risk we will have to accept, for now. Same as a Carrington event. We will however be able to resolve it over time.
Me : ... a risk we will have to accept, for now. Same as a Carrington event. We will however be able to resolve it over time.
X : Does Alibaba copy AWS?
Me : It's just a long list of coincidence - medium.com/@cloudstatepl/… - Alibaba isn't daft, neither is China Gov.
Me : It's just a long list of coincidence - medium.com/@cloudstatepl/… - Alibaba isn't daft, neither is China Gov.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
