Tommi - Fotolia
IT departments are facing an uphill struggle to manage the masses of data being generated by modern IT systems, and the situation is set to get worse as more internet of things (IoT) devices become embedded into everyday products creating new opportunities for customer engagement.
At the recent Moogsoft AIOps Symposium in London, Rüdiger Schmidt, a manager at Daimler’s diagnostics and connected car division, discussed the challenges the automotive manufacturer faces as IoT becomes embedded into its vehicles and roadside furniture such as traffic lights.
He said Daimler has needed to embrace new software to remain competitive and meet new customer expectations. “We are building cars, but simply building cars is not enough any more. We have new competitors who are working on aggressive software models, snaking in between customer and producer.”
To address this threat, Daimler has established Digital One, an initiative to build a secure IT foundation to support digital marketing, the digital value chain, sales and after-sales experiences.
As Daimler builds more software-powered services into its vehicles, IT is now exposed directly to Daimler’s customers. “Five to 10 years ago, we saw customers once a year when they went to the workshop,” said Schmidt. “Now it is completely different. Consumers want services in real time that are always available.”
Among the areas being developed is roadside assistance, which sends data to Daimler when the driver presses an SOS button on the car’s dashboard. Schmidt said the experience the driver has with Daimler can be greatly enhanced if the operator in the customer assistance centre has the right information available.
“Health data from cars and events are processed in the back-end system and additional information is added to predict a solution to the problem and trigger follow-up processes.”
Increase in data
Schmidt expects telematics data generated by road vehicles will increase exponentially. “Cars, buses and lorries are not a smartphone. They are on the road for 20 years.”
Each day, 80GB of data is being collected from each car, and this is set to grow to 250GB per day. Analysing this amount of data to provide meaningful insights poses a big challenge for the car industry.
“This is only the beginning. In the next three to five years, mobility platforms will be integrated with IoT objects like gas stations and traffic lights,” he said.
“Mercedes is looking at a new range of A-class cars for private car sharing that can be unlocked via a smartphone app. If a customer uses his own smartphone and it is not working, our brand value is affected. We don’t have any experience supporting these new ecosystems and I don't have a view of how we will support this in the future,” said Schmidt.
Learning from logs
It is not just the automotive industry that faces the challenge of supporting an increasingly complex IT environment.
According to Moogsoft, which specialises in the emerging market for AI-powered IT operations tools (AIOps), the average enterprise will accumulate 44TB of data a day, which is set to increase 40-fold.
However, speaking at the AIOps Symposium, Moogsoft CEO Phil Tee warned that even though there is a lot of data being generated by system logs, “the cost of server downtime is increasing, and in IT operations, outcomes are not improving”.
Moogsoft believes that for IT departments to improve how they tackle server downtime, IT needs AI. “Use data science to power service improvement,” he said.
AIOps holds the promise to automate many of the manual tasks needed to keep the lights on in IT by delivering analytics across silos.
Forrester’s Vendor landscape: Cognitive operations paper states: “AIOps primarily focuses on applying machine learning algorithms to create self-learning – and potentially self-healing – applications and infrastructure. A key to analytics, especially predictive analytics, is knowing what insights you’re after.”
In effect, data logs produced by software systems can be used in a machine learning algorithm to understand the health of that system. As the AI learns, it is able to identify patterns that suggest potential problem areas that need to be fixed for the system to run optimally. Over time, most experts predict these AIOps tools will become fully autonomous IT admins, and will be able to fix IT problems without human intervention.
According to Gartner research director Vivek Bhalla, by 2022, almost a third of IT organisations that fail to adopt AI will not be operationally viable. He said Gartner’s research has found that half of CIOs in a survey of 3,000 are already looking at AI and IT operations and are among the early adopters.
Bhalla said AIOps tools will be used to assist and augment human IT operators by offering visibility across an IT system, noise reduction on the masses of data generated by system logs and predictive warning. “Legacy tools have thus far failed to deliver on these,” he said.
At Daimler, the various IT services run across a heterogeneous infrastructure, covering many generations of systems including mainframes. Handling an IT problem behind what may seem like a simple request, such as when the driver presses the SOS button on one of the new private hire A Class Mercedes, is non trivial. “Who is responsible? No one support team can cover all of this,” said Schmidt.
The company has been using Moogsoft’s AIOps tool in one particular area of the business. Schmidt said the IT teams used to spend 80% of the time detecting the root cause of an IT problem. Once the root cause of a problem had been identified, generally the fix could be applied in a matter of minutes.
“This was killing our availability rates,” he said, adding that AIOps could help, as it can identify the root cause of the problem quicker.