Submitted by Airlines, banks, stock exchanges and trading platforms suffered brief website outages early Thursday after a key piece of internet infrastructure failed, sparking the second major interruption of the past 10 days.
Virgin Australia said in a statement on Thursday that it had resolved an IT outage caused by a failure at Akamai Technologies, a global content delivery network.
The second major internet outage in a few weeks. Not a good look.
Outages are nothing new obviously, technicians have always made mistakes, but it’s interesting to see the impact that years of consolidation in our industry has done. Although faults are fairly rare, now when they happen it can take many more companies down simultaneously.
I remember when google classroom went down during the pandemic many different schools all independent decided to close for the day. The cloud outage earlier this week that brought down southwest airlines and reportedly caused thousand of flights to be delayed and canceled leaving people literally stranded and having to go back home or book a hotel until they could reschedule.
Cloud services are much cheaper than having inhouse IT staff to manage critical infrastructure 24/7, which is the primary reason for companies to decide to outsource. There’s no denying it makes financial sense on paper. But it also leaves the infrastructure in the hands of non-staff with far less familiarity of your critical applications, which can exacerbate emergency scenarios.
The reason cloud providers are cheap is because they optimize their staffing levels for a typical day where 99% of their customers do not need any support for the day. But try calling your cloud provider when they are experiencing their own emergency and 100% of users need support at the same time. Meanwhile your customers are calling you for help, but you can’t help them because you are waiting on another company to bring you back online. Fun times, haha.
Cloud services are much cheaper because they deliver standard products, so you don’t have any special snowflake individual critical infrastructure, where you as a small fry might have issues getting vendor support when shit hits the fan.
Their main benefit is exactly that you need less IT staff because you have less to manage and hence less stuff that might break because you manage it, making you able to focus your resources instead on something that give you a strategic benefit.
This is true even if we totally ignore all the cloud native parts and just look at them as an alternative to traditional hosting.
Troels,
Well, believe it or not most of the cloud services I’ve seen during my career have been for specialized processes and not standardized applications. I’d bet you that amazon does not understand half of the data & code running on amazon cloud services. And that the cloud providers who took down southwest and frontier airlines probably do not have much expertise in the airline’s systems running on their infrastructure.
Obviously less IT staff is the main selling point, but there are cons, like becoming more dependent on an outside cloud operators who are less knowledgeable to fix your stuff.
You are talking about SaaS, something like office 365 where a highly standardized application is used by millions of customers. But that’s an extremely different scenario than when a business like an airline moves it’s operations into “the cloud”.
I remember when the internet went up and down like a yo-yo. If it wasn’t a modem disconnecting it was a server crashing or a file transfer connection slowly going to sleep. Mind you most people had never used the internet to a significant degree back then.
Not only am I glad I don’t code now but IT support is someone else’s problem too!
I do miss dialling the bank directly from my green screen Amstrad PCW.