Allow me to begin my ponderings by picking up a point made by Karissa Bell on Mashable related to the major disruption of Amazon’s cloud services that took down a wide range of online services:
Amazon’s lengthy AWS outage Tuesday was a stark reminder of just how much farther we have to go to realize the seamless Jetsons-esque future gadget-makers so desperately want us to buy into. The hours-long disruption took down much of the internet, crippling day-to-day activities for many who rely on AWS-backed services, like Slack and Trello, to do their jobs.
She further extends the problem definition to the Internet of Things:
[…] for smart home enthusiasts, the outage also exposed the fact that our lights and doorbells and other gadgets have (yet another) major weakness. By relying on AWS and other cloud services for core functionality, the devices that are supposed to make our lives easier can all too easily be rendered useless.
The “disconnect” is a core UX question
This triggers an interesting thought on user experience and usability. While centralised (cloud) services are widely celebrated for the seamless experience and almost magical interoperability of devices, services and applications, their failure is commonly not considered in the design process (and I am not talking about the design of error pages, but about the deep foundation of technology concepts).
Granted, any website or server - centralised or not - may go down at times for a variety of reasons. And if it happens to be a big provider that is affected, so is the amount of sites and services failing. Even though we’d love to think of digital tools as uninterrupted 365/24/7 commodities, the reality is that, like everything on this world, tech infrastructure can have its hiccups.
But from a design perspective, this is not the core issue. Much more interesting is how devices and services designed to interoperate behave when one of the links fails. This leads my argumentation to the “Offline First” movement (more on their blog and on A List Apart) an initiative concerned with the question of how to design web services when turning the widespread baseline assumption of constant connectivity into one of disconnection and failed links:
We live in a disconnected & battery powered world, but our technology and best practices are a leftover from the always connected & steadily powered past.
What a powerful mission statement!
The reason I am fascinated by this premise of the “Offline First” ideology is rather obvious, as it is closely related to my long-standing interest in “design for non-users” (as opposed to always designing for idealtypical “users”). “Offline First” does the same thing: challenge the established core assumption of “being connected” and turn it into a paradigm that highlights the exact opposite.
Naturally, coming from the web community, the movement primarily highlights questions of web services:
Offline capability is a key characteristic of modern Progressive Web Applications. Offline first thinking must learn from and further what we’ve seen work with Responsive and Mobile First thinking.
Nonetheless, the same questions are valid on a broader scale - any technological development should see consideration of the offline state (and, I shall add, the absent peer) in the design process. Only if we base design on realistic assumptions from real life (disconnects, empty batteries, server outages…) will the outcome be of true value for humans.
Bonus question: Is endless centralisation really the solution?
Yet, it cannot be denied that incidents like the AWS outage also challenge the idea of endless centralisation as such. As one example among many, Bell’s article illustrates how the AWS-hosted IFTTT service, used by many to automate all kind of processes, temporarily bricking a lot of smart devices. And closes:
Having so many interconnected devices that rely on the same third-party providers for critical infrastructure puts all those gadgets at risk when one link in the chain stops functioning. […] Just how many times will people put up with server outages half a continent away knocking their lights and doorbell offline before they decide that they don’t really need the tech to begin with?
We must not assume technology to always work as specified
It is about time to move the utopia of always-on, always-connected, centralised-in-one-place into the realm of technology sci-fi, and work hard to develop more human-centred, realistic and most importantly disruption-proof solutions that do not fail to deliver on their value proposition once a single server flips somewhere on this globe or a user is temporarily disconnected.