CV_JulyAug_2017_linked

6 / 60

Sure, it sounds crazy. After all, cloud computing is arguably just really getting
going in terms of acceptance and adoption. Heck, we haven’t even reached the
actual end of legacy phone service, despite VoIP’s promise to displace it for nearly
20 years. Even so, Peter Levine of Andreessen Horowitz makes the case that we
must undergo, if we’re not already undergoing, a return to the edge and distributed
computing in a type of “back to the future” scenario.
For starters, everything that’s popular in technology always gets replaced, says
Levine. “That’s the opportunity and beauty of the business.” So when considering
where “the puck will be in the future,” Levine paints the now-familiar picture of a
world where billions of devices and sensors make up the Internet of Things. And
these devices, for the first time, are collecting massive amounts of “real-word”
data and information.
Keep in mind, we’re not talking about just text, voice and database output here,
but rather all types of sophisticated endpoint devices collecting massive amounts of
information on video streams, images, depth perception, temperature, acceleration,
location and gravity, among other data points. In many cases, the information will
need to be processed and acted upon in real-time – including the use of machine
learning to extract the relevance of that data in real-time.
“With the latency of the network and the amount of information for many of these
systems, there isn’t the time for that information to go back to the central cloud to get
processed in the same way that a Google search gets processed in the cloud right
now,” Levine argues. “And this shift is going to obviate cloud computing as we know it.”
Consider a self-driving vehicle that senses a stop sign, says Levine. If data has
to go back to a data center to decide it’s a stop sign, “the car would blow through the
stop sign before that data came back from the cloud saying, ‘Hey you need to stop.’”
In many cases, machine learning will be needed to decipher the nuances of the
real-world information in real time, and the algorithms and applications employing
machine learning will run at the endpoint, not in the cloud, Levine predicts.
In Levine’s future, self-driving cars, which already carry impressive amounts
of processing power, are effectively “a data center on wheels,” he says, “and a drone
is a data center with wings; and a robot is a data center with arms and leg.” Pull
together these cars, drones, robots and billions of sensors and devices, and you
have a massive distributed computing system at the edge of the network.
Not that the cloud goes away. It becomes “the last point of information storage,”
says Levine, where important information still will get stored, as well as where learn-
ing occurs before it is pushed out to the edge, where the most important decisions
are made. Ultimately, it’s a matter of “agility over power,” says Levine, where the
cloud remains the most powerful, but the edge is faster, smarter and more agile.
No, were not going to proclaim: “The cloud is dead.” But Levine’s vision certainly
provides lots to think about moving forward.
The end of cloud computing? Martin Vilaboy
Editor-in-Chief
Tara Seals
Contributing Editor
Percy Zamora
Art Director
Berge Kaprelian
Group Publisher
Rene Galan
Associate Publisher
Anthony Graffeo
Associate Publisher
Miki Takeuchi
Marketing & Digital Media
Beka Business
Media
Berge Kaprelian
President and CEO
Neil Ende
General Counsel
Corporate Headquarters
14700 N Frank Lloyd Wright Boulevard, #157
Scottsdale, AZ 85260
Voice: 480.503.0770
Fax: 480.503.0990
© 2017 Beka Business Media, All rights reserved.
Reproduction in whole or in any form or medium
without express written permission of
Beka Publishing is prohibited. ChannelVision
and the ChannelVision logo are trademarks
of Beka Publishing
LETTER
6 ChannelVision | July - August, 2017