Earlier today, people throughout the world found out that Skype just wouldn’t work for them. They couldn’t see their contacts online, they couldn’t call them, and their sole comfort was a laconic message from the program – “We’re a bit overloaded right now… Please try again later, or download Skype to use it any time.”
Admittedly, that’s not much of a comfort.
While the UK and Japan bore the brunt of the disconnections, people all over the globe were affected by the outage. Skype is truly a global phenomenon: it brings together more than 300 million users, and has been downloaded 500 million times from Google Play alone. By the year 2014, it had 4.9 million daily active users. For each minute Skype is offline, the firm behind it is losing thousands of dollars. It stands to reason that the engineers behind the system have developed multiple layers of defense against failure. And yet they all failed, resulting in an 11 hours shutting down of services.
What can we learn from this event?
First, that complex systems – a term that covers basically all the transportation, communications and healthcare systems we have today – are bound to fail at some point. Richard I. Cook, Professor of Healthcare Systems Safety, explains that complex systems “contain changing mixtures of failures latent within them”, and hence – that “catastrophe is always just around the corner”. At one point in time, a few latent failures rise up together in an unexpected way, and cause a catastrophe that surprises everyone involved.
Secondly, we have systems that encompass practically the entire world, and yet are controlled hierarchically or communicate with each other in a way that a failure in one zone could cascade to many other areas worldwide. Skype is one such system, and so are the GPS satellites, and the global stock market as well.
Thirdly, we need to prepare future scenarios that take into consideration some extreme occurrences: what happens if Uber’s services, for example, are no longer available because of a system failure? In cities like San Francisco where 65% of traditional taxi business has been replaced by Uber drivers, the sudden crash would strand many travelers, particularly those who rely on their smartphone as a payment method (since Uber is deducting the price of the ride automatically from a person’s account).
Similarly, what happens if the GPS satellites somehow switch off? This possibility is not the stuff of science fiction movies, since most of the GPS satellites are long past their designated lifespan. If somehow the GPS keels over, almost all aspects of our life would be disrupted. Uber taxi drivers won’t know where their clients are, many people will find themselves on the road in their cars desperately trying to find an actual paper map, trucks will miss their shipments, tractor drivers in widespread farms won’t know where they are; cargo and leisure ships will become momentarily lost in the ocean, and so on.
Do these scenarios seem overly dramatic to you? Unbelievable, maybe? Well, it just happened to Skype. And it will happen again to other firms and services, somewhere, sometime in the future.
This forecast is particularly alarming in light of the fact that we are creating automated systems that control our daily lives in a time resolution of mere seconds. For example, we are going to enjoy the use of driverless cars sometime in the next five to ten years. How will such vehicles act when they’re suddenly all disconnected from their servers? Will they all stop in the middle of the street, blocking the roads for everyone? Will they complete one last trip and then try to park safely… along with all the other hundreds of thousands of driverless cars in the street? These are questions that we need to consider ahead of time, in order to develop more robust global systems.
How do we deal with such potential failures in global systems? The usual response in the present is to add layers of defense mechanisms, which are fine at delaying the catastrophe for some time, but can never negate completely the chance of it happening. I would suggest that an additional course of action would be to add redundancy to service providers, so that no single service provider can become a worldwide monopoly in any critical field. This line of thinking is possibly the reason that China, India and Japan, among others, are currently planning to launch and operate a navigation system of their own… just in case something happens to the GPS satellites.
Skype has made a comeback within eleven hours, and the world is buzzing with international video conferences again. We should, however, keep in mind the words of Paul Romer – “A crisis is a terrible thing to waste”. We should learn from Skype’s momentary disappearance to better prepare for the future, and avoid concentrating too much power and control in the hands of a single firm, system or infrastructure.
Roey Tzezana, PhD, is a Futures Studies researcher at the Blavatnik Interdisciplinary Cyber Research Center (ICRC) at Tel Aviv University. He’s currently studying and developing scenarios for the future of crime in the age of the Internet of Things. His full information can be found at www.guidetofuture.com
2 thoughts on “The Fall of Skype: A Lesson for the Future?”
Your solution of “add redundancy to service providers” is no different to adding layers of defense mechanisms, because it doesn’t matter, in essence, if a defense mechanism (such as redundancy) is owned by company X or Y. As you say, it just delays the failure time.
That said, “delaying failure time” can be enough…if you delay failure time for long enough. If a complete system failure occurs once every 1000 years, that may be as good as “no failure”.
THAT said, I do believe humanity’s growing dependence on global, interlocking complex systems is bound to cause havoc at some point, with a gross cost in lives and money. The only way to deal with that is to decouple systems and encourage independence, which is something our current society is loath to do.
LikeLiked by 1 person
What can I say? I agree with pretty much everything you said.
I do believe that a balance needs to be found between decoupled and independent systems (which increase waste of resources), and one global system upon which everything hangs. I just hope we find the right balance.