A “smart bed” in Europe down because of the Amazon Web Services (AWS) outage?
Surely not. A few up-down buttons for something you literally sit on would not need to call a web service 5000km away to operate? Well, yes, in fact.
This particular example was on the BBC to show how pervasive it is for even the simplest of applications to use (let alone be hosted by) cloud microservices as if they were simple pieces of local code under your control.
I recently comment on this on the Institution of Engineering and Technology (IET) press release.
Dr Maurizio Pilu, a member of the Digital Policy Centre at the IET said:
“Amazon has just released a report indicating that the root of the issue seemed to be in a core database Application Programming Interface located in a specific data centre.
“This system plays a crucial role in resolving where exactly data and services on the Amazon cloud are located. While the technical failure is understood, I am sure that like in all case like these the underlying cause will be thoroughly investigated.
“Typically, Amazon outages last 1 – 2 hours, and with the root cause identified, services should fully resume shortly. What makes this incident particularly significant is the widespread impact across services, regardless of whether they rely on American or EU infrastructure.
“The published list of affected sites is extensive. Events like this serve as a stark reminder of how concentrated and potentially vulnerable our global information infrastructure has become – something worth reflecting on.”
More resilient designs are possible, but the allure and simplicity of cloud-based hosting and services is strong. BTW that smart bed company has switched to bluetooth now.
Yet with 3-4 companies now hosting or being used by a huge part of the whole civilian digital ecosystem of the West, and now also the AI infrastructure, the list of questions and that risk register is growing longer and longer.