adam121 - Fotolia
A little over a year ago, Drew Miranda and his team of systems engineers at Ochsner Health System in Louisiana started collecting log data from hospital remote desktop systems as a pilot project. They started pulling data from desktops tied to a single server with the aim of analyzing it to potentially do preventative maintenance and fixes. But Miranda quickly realized he was on to something big in collecting and analyzing log data.
"It was almost like a eureka moment," said Miranda, systems engineer at Ochsner. "When I started to realize how much value we were getting out of it, it was something compelling I could take to my bosses."
Miranda said he used to look in on streaming log data periodically, but it was impromptu. He said this approach is normal in his industry. "One of the common things I hear from friends is how few people look at logs." The slow pace of industrial analytics adoption is primarily due to the sheer volume and velocity of the data, as well as a lack of clear value proposition.
But all that's starting to change. Today, the technical hurdles to industrial analytics, also called the industrial Internet of Things, or IoT, are lower than in the past. Organizations are building use cases that prove the value of monitoring log data as well as data from other industrial systems that was simply discarded in the past.
Getting proactive with industrial analytics
Drew Mirandasystems engineer, Ochsner Health System
At Ochsner, Miranda and his team implemented a log management system from Houston-based Graylog Inc. that ingests and organizes log data. The team is monitoring data from about 17,000 desktops within the network of hospitals and physician offices. The data is used to track the root cause of common problems, enabling preventive maintenance, and to constantly monitor the status of desktop systems. This allows them to fix recurring errors that prevent applications from opening and make sure that all desktops are running the latest version of their operating systems and all updates are installed.
"We know before our users are reporting problems," Miranda said. "A lot of times in a service-driven IT department, if you don't get a call, we assume there isn't a problem. But now we can fix things before the users even know."
Since Miranda and his team have only been monitoring log data on a wide scale for a short time, they have not measured a precise return on investment. But Miranda said he has been able to work more proactively with desktop services vendors and improve uptime.
Sensing a new business opportunity with industrial IoT
Industrial analytics isn't just about log data. Companies are also realizing a benefit from monitoring and analyzing sensor data. For example, San Francisco-based startup Compology has built its entire business around sensor data. The company places optic sensors in commercial Dumpsters that monitor the volume of waste. The company uses software from Seattle-based Dato Inc. to ingest and analyze the sensor data, which allows Compology's customers to efficiently route collection vehicles based on which Dumpsters are most in need of emptying. This approach reduces the number of trucks needed for waste collection by as much as 40% compared to more routine routing, said Jason Gates, one of Compology's founders.
Gates said the main benefit of collecting and analyzing sensor data is that, compared to other data sources, it is more actionable. It is almost always tied directly to some business process that can be improved by closer monitoring. In his view, this benefit is going to have the greatest impact in industries like waste management that are traditionally data poor.
"What I feel strongly about is that it's not necessarily access to data and numbers, but it's taking that information and making it actionable," he said.
Experts agree. Dale Neef, consultant and author of the new book, Digital Exhaust, said he thinks industrial analytics will soon deliver greater business value than what we typically think of as big data. He said compared to collecting data from customers, such as social media posts, Web search history and demographics, the payoff from log and sensor data is more immediate.
"There is an enormous explosion going," he said. "[Businesses] are looking at sensors through the supply chain, which allows you to do things like computer-based maintenance. I think the industrial Internet is going to change things enormously."
See how one company made sensor data even more actionable
Manufacturers go in big on the industrial IoT
The industrial IoT has big potential for healthcare companies