BREAKINGVIEWS-Amazon outage triggers internet alarm

Shutterstock photo

(The author is a Reuters Breakingviews columnist.  The opinions
expressed are his own.)
    By Robert CyranNEW YORK, March 1 (Reuters Breakingviews) - Amazon's
<AMZN.O> web-services outage on Tuesday is cause for internet
alarm. Websites, apps, connected cameras and the like went dark
after a breakdown of the $400 billion online giant's
cloud-storage service. Amazon's dominance, Silicon Valley's
self-confidence and users' reluctance to pay extra for backup
all add up to rising risk for the internet of things.
    The internet was initially conceived as a massively
redundant communications network that could survive a nuclear
war. Yet as it matures, commercial realities and human nature
are making the system more fragile. Companies find it
convenient, simple and cheap to let Amazon, Google parent
Alphabet <GOOGL.O> and Microsoft <MSFT.O> run the data centers
and applications they need. Furthermore, the cloud business has
aspects of a natural monopoly - the bigger a network, the more
efficient it is both operationally and financially.
    So it came as a bit of a shock to everyone from hip
programmers at Slack to bureaucrats at the Securities and
Exchange Commission when Amazon's simple-storage service, known
as S3, had problems at its East Coast operations. That made it
impossible for many apps to pull up data, news publications to
publish pictures and internet-enabled ovens to turn off for
several hours. was among the sites affected. Even Amazon
suffered an ironic glitch. The online dashboard it uses to
indicate how well its giant Amazon Web Services unit is working
flashed up green - implying all was well - because problems at
S3 prevented it changing color.
    The blame isn't Amazon's alone. Customers could have reduced
their exposure by doubling up on data storage even within AWS,
since its West Coast systems weren't troubled. Better yet, they
could have also signed contracts with Google and Microsoft. The
problem is, it's easier and cheaper for optimistic developers
and time-starved users to rely on a single service provider and
its promises of near-perfect reliability.
    Sensors and web connectivity are being built into more and
more devices. Thin margins, inexperience and complacency mean
security and redundancy are often lacking. It's comical when
someone can't turn off a remote light in their beach cottage;
it's less funny when autonomous cars or life-saving medical
equipment go haywire. Tuesday's outage may concentrate minds,
but the danger will continue to grow.
    On Twitter

    - Multiple websites and apps failed on Feb. 28 after
Amazon's simple-storage service, known as S3, had difficulty
sending and receiving client data for several hours. The
problems centered on Amazon Web Services' operations on the East
Coast of the United States.
    - Websites that rely on AWS for storage and encountered
problems included Slack, Quora and the Securities and Exchange
Commission. Many news sites, including Breakingviews, could not
publish pictures and experienced other disruptions. Multiple
everyday objects with internet connections such as lightbulbs,
stoves and cameras - also known as the internet of things - also
had problems.
    - For previous columns by the author, Reuters customers can
click on [CYRAN/]

Disruption in Amazon's cloud service ripples through internet
 (Editing by Richard Beales and Martin Langfield)
 (([email protected];)(Reuters Messaging:
[email protected]))