Zebra improves big data analytics using MongoDB Atlas and Google Cloud

Smart devices are everywhere, from our pockets to our factories to our cars. The NFL even uses them to track players on the field. But the Internet of Things (IoT) means nothing if it’s not easy to capture, analyze, and action data from devices in real time.

Zebra Technologies, a large tech company that focuses primarily on B2B hardware and solutions, launched Savanna to create a seamless connection between customer devices for advanced automation, improving the user IoT experience and providing them with a competitive edge. “Customers have hundreds of thousands of devices to manage,” says Jeffery Stovall, manager of software engineering at Zebra Technologies. “Savanna detects existing issues and uses predictive intelligence to prevent others from happening.”

The platform is built on highly-scalable, highly-available architecture using Google Kubernetes Engine and other Google Cloud Platform (GC) technologies. It has an operational data store built on MongoDB Atlas and a Google Cloud BigQuery analytics warehouse.

Let’s take a look at what each technology brings to the table, and why MongoDB and Google Cloud’s partnership enables Zebra Technologies to provide the best experience for their end customers.

Zebra connects to many types of devices, so it needs a flexible document model that can also handle huge volumes of traffic with burst periods. “We’re working with all different kinds of devices and applications. Our data model is very fluid, so a document model makes a lot of sense for us,” shares Stovall. “Additionally, MongoDB Atlas provides us with the ability to scale both horizontally and vertically. Autoscaling around peaks saves on costs, it’s easy for developers to use, and as a managed service, they can focus on high-value tasks knowing that the platform is up to date and working with high performance.”

“As our platform grows,” continues Stovall, “Atlas allows us to easily meet our update and scaling requirements, which frees up developers and SREs to produce more valuable functionality. Atlas makes it really easy. You don’t have to worry about the configuration on your own side…you can update your database really quickly with no downtime.”

MongoDB Atlas Search supports Regex (regular expression) searches that used to be run in Elasticsearch. “Our previous approach wasn’t the best — it risked data getting out of sync. It was awesome to switch to Atlas Search and run searches at a lower cost without increasing CPU usage,” Stovall explains.

Zebra uses analytics and machine learning to get smarter about device performance. Historical data helps to spot trends and identify issues that might occur in other devices. These can then be proactively recalled for repair, avoiding a situation where they suddenly stop working.

While MongoDB handles real-time data with millisecond response times, it would take longer to query historical data. This is where BigQuery comes into play. Curated data sets are replicated for complex analytics and machine learning.

“MongoDB and Google Cloud enrich operational data and create a better customer experience. They’re a great example of complementary technologies working better together to support our business,” concludes Stovall.