There was an AWS outage a few days ago; You were probably impacted in one way or another.
If any of my appliances would ever e-mail me, I think I'd die of shock.
Every thing from Disney+ to McDonalds was affected, as US-East-1 was US-East-None for a few hours. Even Amazon delivery drivers and warehouse workers couldn't complete their breakneck, no bathroom, tasks.
My Alexa couldn't reach AWS. "Guess I better attach it to the new hotspot..." I thought, before moving on with my day and not doing that. Turns out, my hotspot wasn't the issue.
This is not the first cloud provider outage to happen, not even this year. So, how can technicians prevent this from happening?
Outages will happen, and this may have broken whatever 5 9's streak AWS strives for as a whole, but I'm 100% sure AWS networking technicians are working to implement high-level ways to recover quickly from this in the future. Someone uncovered that error, and should be commended. Who knows how long it would have been lying in wait?
That doesn't mean you can't be prepared. The streaming giant Netflix uses AWS to host its thousands of shows and movies - and they still have onsite systems and network engineers in Los Gatos, California.
What about multi-cloud? That a thought, but this Twitter thread showed me the cons to that idea.
I think nothing but the absolute largest companies would benefit from multi-cloud; The question is, can it be justified cost-wise.
What would work for your company -
Losing x$ during a 4 hour outage
Spending x$ a year to maintain two cloud relationships?
It depends partially on how often you expect your primary cloud provider to fail - Something that's happening more and more.
VentureBeat has a great article of what happened, with thoughts from many industry professionals.
Comments
Post a Comment