MADRID: Why does the internet keep breaking?

MADRID: Why does the internet keep breaking?

MADRID: But, if he did, it would take
him approximately 145 days, without sleep, to wade through the deluge of
comments left for him after he apologised for the meltdown of services last
week.

“Sorry
for the disruption today” the Facebook founder and chief executive posted,
following almost six hours of Facebook, WhatsApp and Instagram being offline.

Facebook blamed a routine maintenance job for the disruption – its
engineers had issued a command that unintentionally disconnected Facebook data
centres from the wider internet.

Around
827,000 people responded to Mr Zuckerberg’s apology.

The
messages ranged from the amused: “It was terrible, I had to talk to my
family,” commented one Italian user, to the confused: “I took my
phone into the repair shop thinking it was broken,” wrote someone from
Namibia.

And, of
course, the very upset and angry: “You cannot have everything shut down at
the same time. The impact is unprecedented,” one Nigerian businessman
posted. Another from India asked for compensation for the disruption to their
business.

What is
clear now, if it wasn’t obvious already, is just how reliant billions of people
have become on these services – not just for fun but also for essential
communication and trading.

What is
also clear is that this is far from being a one-off situation: experts suggest
widespread outages are becoming more frequent and more disruptive.

“One
of the things that we’ve seen in the last several years is an increased
reliance on a small number of networks and companies to deliver large portions
of Internet content,” says Luke Deryckx, Chief Technical Officer at Down
Detector.

“When
one of those, or more than one, has a problem, it affects not just them, but
hundreds of thousands of other services,” he says. Facebook, for instance,
is now used to sign-in to a range of different services and devices, such as
smart televisions.

“And
so, you know, we have these sort of internet ‘snow days’ that happen now,”
Mr Deryckx says. “Something goes down [and] we all sort of look at each
other like ‘well, what are we going to do?'”

Mr
Deryckx and his team at Down Detector monitor web services and websites for
disruption. He says that widespread outages affecting major services are
becoming more frequent and more serious.

“When
Facebook has a problem, it creates such a big impact for the internet but also
the economy, and, you know… society. Millions, or potentially hundreds of millions,
of people are just sort of sitting around waiting for a small team in
California to fix something. It’s an interesting phenomena that has grown in
the last couple of years.”

Inevitably,
at some stage, during a large outage of services, people worry that the
disruption is the result of some sort of cyber-attack.

But
experts suggest, more often than not, it’s down to a more mundane case of human
error, compounded, they say, by the way the internet is held together with a
complex set of outdated and fiddly systems.

During
the Facebook outage, experts joked on social media platform Twitter that some
of the usual suspects, or reasons for outage problems are “older than the
Spice Girls” and “designed on the back of a napkin”.

Internet
scientist Professor Bill Buchanan agrees with this characterisation: “The
internet isn’t the large-scale distributed network that DARPA (the Defense
Advanced Research Projects Agency), the original architects of the internet,
tried to create, which could withstand a nuclear-strike on any part of it.

“The
protocols it uses are basically just the ones that were drafted when we
connected to mainframe computers from dumb terminals. A single glitch in its
core infrastructure can bring the whole thing crashing to the floor.”

Professor
Buchanan says improvements can be made to make the internet more resilient, but
that many of the fundamentals of the net are here to stay for better or worse.

“In
general, the systems work and you can’t just switch certain protocols of the
internet ‘off’ for a day, to try to remake them,” he says.

Instead
of trying to rebuild the systems and structure of the internet, Professor
Buchanan thinks we need to improve the way we use it to store and share data,
or risk more mass outages in the future.

He argues
that the internet has become too centralised, i.e. where too much data comes
from a single source. That trend needs to be reversed with systems that have
multiple nodes, he explains, so that no one failure can stop a service from
working.

There is
a silver lining here. Although significant internet outages affect users’ lives
and businesses they can also, ultimately, help to improve the resilience of the
internet and the web services plugged into it.

For
example, Forbes estimates
that Facebook lost $66m
(£48.5m), during the six-hour outage, from the suspension, or exodus, of advertisers
on the site. That sort of loss is likely to focus the minds of senior
executives on preventing it happening again.

“They
lost a huge amount of money in that day, not just in their stock price but in
their operational revenues,” according to Mr Deryckx.

“And
if you look at outages caused by content delivery networks like Fastly and
Cloudflare, they also lost a huge number of customers to the competition. So, I
think these operators are doing everything they can to keep things
online.”

Leave a Comment