Las Vegas is using a new machine learning platform to troubleshoot and predict IT system failures, getting the networks back online more quickly.
An AI-powered system to monitor and trouble-shoot the increasingly complex and dynamic technology networks is helping officials to know where system failures may occur, and how to fix them quickly. Think of it as an IT help desk on steroids.
Las Vegas uses the FixStream artificial intelligence for IT operations (AIOps) platform to constantly monitor its many IT and smart city networks, across two dozen departments, looking for anomalies, outages or other malfunctions. The software is able to quickly trouble-shoot the system, zeroing in on the issue, allowing city officials to quickly get in front of a problem.
“Without this type of technology, the ability to keep the up-time that we need to keep, we don’t see that being possible,” said Michael Sherwood, chief information officer for Las Vegas. “Because our systems are so complex now, and the demand for restoration of service when something goes wrong is so high, without using machine learning, without using advanced artificial intelligence, there’s really no way that we would be able to keep up.”
“Being down a day, two days, three days, that’s just unacceptable today,” said Sherwood. “When people want to pay their water bill, they want to pay their water bill.”
Las Vegas has acquired a reputation for being one of the most connected cities in the country, setting aside a large swath of the city’s downtown as an Innovation District — a veritable testbed for exploring autonomous vehicle technologies, as well as a bevy of sensors set up to gather data related to traffic, pedestrians or bicycle use.
“Where it’s more vital is coming into our IoT, our industrial network, or smart cities efforts, where you have services now where we’re using analytics, camera data, lots of data resources in multiple locations,” said Sherwood.
Those systems gather and process data at alarming speeds and complexities, requiring IT officials to have an equally robust protocol to understand and analyze them when something goes wrong.
“IT is becoming very, very complex now because we’re deploying very modern and agile technologies,” explained Enzo Signore, chief marketing officer for the San Jose, Calif.-based company. “The problem with this is they become very invisible. The environment is becoming very hard to understand and manage. And today it takes, typically, many hours to understand the root cause of a problem ... You’re dealing with millions of data points that need to be correlated.”
When smart city functions controlling areas like traffic control or autonomous vehicle communications fail, these setbacks can create a trickle-down flow of inconveniences and other concerns, said Sherwood.
In the case of functions like traffic lights, crosswalk timing, and general mobility through an area, what once might have been just an IT problem can quickly balloon into a significant inconvenience for the traveling public, he explained.
And like so many IT systems working behind the scenes of governments, where the networks are becoming increasingly intelligent, it helps if IT trouble-shooting systems gain some intelligent predictive capabilities as well.
For Sherwood and the city's smart initiatives, having a system that is analyzing data flows for anomalies while learning from what it finds is vitally important.
The trouble-shooting system works by surveying all devices and systems across the IT environment, said Signore. As it stands, Signore said the technology can predict an outage or issue between 80 and 100 percent of the time.
“We collect information about every single device and the relationship between the devices,” he explained. “The system will then apply machine learning, artificial intelligence algorithms to the tech patterns across the entire site, so you can actually predict when the next outage will occur.”