McDonald's Technology Outage Forces Restaurant Closures

https://www.wsj.com/business/hospitality/mcdonalds-hampered-by-system-outages-in-japan-hong-kong-7e8dd3c7

by crivabene • 2 years ago

19 17 2 years ago

wsj.com

crivabene• 2 years ago

http://archive.today/Hb0zm

joezydeco• 2 years ago

Cloud goes up, cloud goes down. Same story, different day. At least it's not aws-east-1 today.

dylan6044• 2 years ago

All of these things that refuse to work without network connection is just mind boggling. I understand the need/want for the companies to have all of that data centralized, but to not have a graceful pathway for when the network fails is just burning cash. Store data locally so it can be pushed to the cloud when network returns, but don't just stop working.

organsnyder• 2 years ago

The article summary indicates that payment processing was affected. I'm not sure what possibilities there are for offline payment processing, but that's different than just not being able to report sales numbers or something. And opening cash-only is probably not worth the hassle for many locations.

jonhohle• 2 years ago

Maybe it’s their lineage, but I appreciate most native, first party iPhone apps having local first data storage. 4G is increasingly spotty in my area and it’s a crapshoot whether I can reliably load a time sensitive app that requires remote data in some areas. Throwing something in Notes, Photos, or Wallet always works, though. I’m not sure why I still can’t have my full mailbox on my phone, but at least there are several local options.

mortenjorck• 2 years ago

Unfortunately, I think it's more zero-sum than that.

Imagine you're a product manager at McDonald's. For the point-of-sale kiosks, you can build the server component for deployment in the cloud, or you can build it for on-prem and add the overhead of physical deployments across tens of thousands of retail locations, plus maintaining them, plus dealing with franchisees.

You do a cost-benefit analysis of potential downtime from internet outages, and based on internet reliability numbers in the markets you serve (plus an uptime estimate from your devops org), you conclude the occasional outage is worth the trade-off.

sour-taste• 2 years ago

The ice cream machine is actually down today.

johnea• 2 years ago

You have to wonder how much cummulative life expectancy increased by McColonCancer being offline for a day nationwide?

There are some upsides to corps downsides...

ngcc_hk• 2 years ago

Long queue to other fast food is not helping though.

Jakob• 2 years ago

Possibilities to receive cashless payments while the company's network is down:

- printed, static QR code: customer scans it with their payment app and enters the amount

- have an NFC enabled phone that receives payments (most android devices can)

- use the “knuckle busters”: the old school way of imprinting the card number on paper

- NFC enabled device like the one on buses in some cities, synced only at the end of the day

- receive money via one of the marketplace apps: UberEat, Doordash, Grab

- customer browses and pays the company’s online shop in the store instead (this us what Decathlon did once it had a similar outage)

- what else?

I’m quite curious why McDonald’s apparently decided to have no fallback at all. They surely have thought about it. Is the efficiency loss too high? Fraud risk? Training costs?

KRAKRISMOTT• 2 years ago

- NFC enabled device like the one on buses in some cities, synced only at the end of the day

This is also how airlines used to work when it comes to collecting payments

gosub100• 2 years ago

I was thinking about this whenever I hear about airlines going down when their dinosaur-mainframe software breaks. How about each airport PRINTS an hourly update with ETAs for each inbound plane, where they intend it to depart, and the disposition for the crew of each plane. maybe even IDNs for all the baggage on each plane and each outbound passenger. Might take up 20-30 pages per hour at a non-hub airport. hubs could have their own "edge" device that can print on demand if the mainframe dies.

If the db goes down, make a bunch of copies of that paper, disseminate them, and keep running. worst case is planes don't run at 100% capacity. Just do the best you can to keep the next few legs going without the mothership telling you what to do. Shouldn't be that hard. I know one of the biggest constraints is that crew members run out of fly-able hours. So print that in the backup too: X pilot has Y hours left after HH::MM zulu time. Maybe lobby FAA to allow pilots to fly a few hr past their maximum a few times a year only if the scheduler goes down. I just dont accept that these billion-dollar networks should be this susceptible to disruption given the current state of tech.

ChrisArchitect• 2 years ago

[dupe]

More discussion: https://news.ycombinator.com/item?id=39713270

McDonald's Technology Outage Forces Restaurant Closures

Related Stories