With such crucial, machine learning-reliant data, a resilient infrastructure can be the difference between an Amazon Go customer that has no idea a migration took place, as it should be, or an Amazon Go customer that reads all about it in the news the next day because the migration caused major glitches at the store and even more speculation about Amazon’s concept.
Automatically Balanced Resources Most companies today run constant failover tests to ensure their disaster recovery strategy works as intended for when they find themselves in a real disaster situation.
However, a common challenge IT organizations grapple with is having enough storage to continually run these failover tests; tests that involve duplicating and moving data over and over again.
With machine-learning-based prediction tools, businesses will be able to see expected storage growth rates and be automatically provided with relevant options.
Click this button to see ‘what if scenarios’ involving moving data to different points, click this button to see how much it would cost to move the data to different cloud platforms and/or choose one to move to, click this button to see how much storage would be saved by removing redundant data.
Or, taking things a step further, if an organization has machine learning that’s advanced enough it can act on pre-determined preferences and automatically balance storage resources based on an algorithm that factors in available space, cost, previous manual decisions and more.
Hands-Free Disaster Recovery In a perfect world, when a business encounters a disaster, its disaster recovery system does its job automatically, quickly and better than any human could do manually.
We’re not there yet, but it’s coming.
While automatic disaster recovery sounds appealing, the machine learning still needs to get advanced enough to understand when, for example, a network is only down for five minutes, and a complete resource-intensive recovery is most certainly not necessary and not worth it.
That said, with the right algorithms and variables, and enough data to help machines determine when it’s right to wait five minutes, for example, before performing a failover, a hands-free disaster recovery system is a possibility.
The ultimate goal with this kind of advancement is less downtime for businesses and customers.
The idea is that machines, once the kinks are worked out with enough algorithm-based trial and error, will always perform better, faster and smarter than humans.
This can equate to disaster recovery processes that happen so efficiently that even an unexpected tornado hitting a 200,000 square foot datacenter means absolutely nothing to the businesses the datacenter serves.
Security Integrations As machine learning advances, the line between security and recovery blends more and more.
Machine learning can help security and recovery systems work together to make better IT decisions.
Let’s take ransomware for example.
Today, before ransomware makes itself known, the virus encrypts a large quantity of data to have the leverage to hold a person or an organization’s data ransom.
A security system equipped with advanced machine learning might be able to detect the encryption taking place as it happens and then automatically connect with a recovery system to both stop the attack as it’s happening and failover to unencrypted data.
This is one of the most exciting and promising ways in which AI and IT resilience can come together.
This scenario will be a reality in the near future, and eventually, with the rate at which cyberthreats are growing, no business will be able to survive or compete without this kind of protective technology in place.
These are just a handful of potential AI-based scenarios in the world of backup and recovery.
There are certainly others, many of which we likely won’t even discover for years to come – such is life in this new, evolving, data-driven world.
But as in all facets of IT, what we know of backup and recovery today will look completely different in only a few short years.
It’s true that we’re only at the beginning of what’s possible, but the speed at which innovation is advancing increases more and more each day, and the implications for how an organization can better protect and preserve its data can equate to a future where the use of the word ‘downtime’ becomes virtually obsolete.
Sign up for the free insideBIGDATA newsletter.