Data Center Operators and IoT’s Perfect Storm
According to Steve Bass, Chevron's North America data center manager, data center operators should prepare for the perfect storm that will bring a sea change that will drastically change the metrics of data center operations.
The storm consists of the advent of IPv6, the coming of 5G, and everybody's favorite 2018 buzzword phrase, machine learning.
In a Data Center World presentation on Tuesday, Internet of Things Impact on Data Center Taxonomy, Bass explained that before IPv6 came on the scene, IoT didn't have much of a change of taking hold due to a shortage of IP addresses. With IPv6 providing 340 undecillion addresses (that's 340 trillion, trillion, trillion in human readable language), that means the shortage has been broken seemingly forever, opening the door for each and every device that wants its own unique IP identifier to get one.
That paves the way for a heck of a lot of machines to ping data centers with an “I'm OK” message every couple of minutes, which could be ignored if the messages didn't contain the occasional “I've got a problem” message, which can't be ignored.
5G, which will vastly increase the speed that mobile devices can talk to the mother ship (or the server's motherboard, take your pick), opens another door for IoT devices that need to say a lot in a hurry.
The problem is that will change the balance at data centers, which will see a quantum leap in the ratio of incoming to outgoing traffic.
Bass said that this is where machine learning comes in. As already observed, most of what these IoT devices will be saying is stuff like “here I am,” which absolutely doesn't need to be routed all of the way back to the data center. A little machine learning will enable that sort of traffic to remain at the edge, and pass through only necessary traffic.
“The evolution of the technology is going to push you into an IoT world,” Bass told the audience of mostly data center operators and vendors.