Internet of Things (IoT) | 24/7 Monitoring

Curb Management Pioneer Automotus Predicts and Prevents Edge Device Failure with Trek10 and AWS

Trek10 builds a robust monitoring solution that would help Automotus quickly catch and manage potential failures in their edge processing devices.

Overview

Founded in 2017 and headquartered in Los Angeles, California, Automotus is a mobility-focused video analytics company that empowers cities with the tools to manage their increasingly congested curbs. Using Automotus’ technology, cities can automatically analyze, manage, monetize, and enforce all forms of curbside activity, including passenger vehicle parking, ride hailing pickups and drop offs, and commercial vehicle loading and unloading. Their technology is powered by video analytics that run on cameras and edge processing devices strategically placed along city streets. These cameras capture metadata that is then reported back up to their cloud application where it is stored, processed, analyzed, and delivered to city administrators.

The Challenge

Automotus was in the market for a robust monitoring solution that would help their team quickly catch and manage potential failures in their edge processing devices. These failures can be caused by a number of factors, including power outages, memory leaks, over-filled disks, and network outages. Edge processing devices in particular are more vulnerable to these kinds of failures because of their exposure to the elements.

Before engaging Trek10, Automotus had tried several off-the-shelf monitoring solutions. However, the offerings available were not designed for the limited resource constraints of IoT products. As a result, Automotus designed a suboptimal solution that notified them if a device was no longer reachable through the same system that provides remote access to their edge devices. To anticipate looming failures, Automotus’ engineers would then perform manual routine checks to anticipate and protect against any potential failures - but this kind of piecemeal system left significant room for error.

Manual monitoring is a manageable task when working with five to ten edge devices in the field. However, once the device fleet exceeds ten, the complexity, distribution, and size make manual monitoring infeasible. At the time of the engagement, Automotus’ small engineering team was dedicated entirely to building out the core functionality of their platform. Realizing that his team lacked the bandwidth to establish domain expertise and build a fully featured monitoring solution, Automotus co-founder and CTO Harris Lummis began looking outward.

Automotus needed a solution that provided insight into how to prevent future failures, as well as what initially caused them. With this visibility, Automotus would be able to understand the status of their systems at all times, including the ability to predict and prevent failures from occurring.

Solution Overview

Trek10 successfully delivered Automotus’ monitoring solution within their tight deadline and at the optimal time –right before Automotus deployed an additional 20 to 30 devices. The monitoring solution Trek10 built allowed Automotus to understand their end-to-end IoT ecosystem –everything from the edge device itself to the services used for device connectivity and data ingestion down to the databases, container applications, and queues on the cloud. Trek10 used AWS IoT Device Management, AWS IoT Device Defender, Amazon CloudWatch, and AWS Lambda. This was all delivered in one solution that leverages Datadog’s brand new IoT Agent to have a single pane view of the IoT ecosystem. Lummis noted that, “Trek10’s team did a phenomenal job of listening to our high-level requirements, establishing clarity, then designing and implementing a solution that was easy to use and highly extensible.”

The solution included three crucial elements:

  • Monitor all activity: The ability to monitor the activity (i.e. memory, disk space, CPU utilization) on AWS.

  • Detect possible failures: The ability to detect if data falls outside a specific range or if there is an anomaly in one of the data points coming in. After a detection event, an alert is generated through Datadog.

  • Perform audits: The first two aspects, monitor and detect happen in near real time, while the audit aspect is a cyclical polling process to make sure that everything in the system is compliant with security standards. Trek10 implemented an auditor service that checks Automotus’ IoT device configurations, credentials, and policies within the IoT device management platform. An alert will be generated if a violation is found –that message will show up as an event in Datadog.

    Now, with Trek10’s robust monitoring solution active, Automotus gets a notification upon reaching a threshold before and upon a failure. They can analyze what causes the failures and determine the type of maintenance necessary. The solution gives Automotus the visibility to address the failures before they happen and provides the information necessary to perform maintenance.

Why Trek10 and AWS?

Once Automotus settled on executing the project with third-party assistance, their team researched potential partners. It was important that their partner hold deep IoT expertise and be well versed in Amazon Web Services (AWS), as Automotus valued the AWS pace of innovation and the maturity of their IoT and serverless offerings.

In their research, Trek10 rose to the top. Not only were they an AWS Premier Consulting Partner with an AWS IoT Competency, but they had a track record of success. This gave Automotus the confidence that Trek10 had the expertise they were searching for.

Additionally, Automotus’ timeline and budget were tight, and they were looking for a partner who was ready to deploy a solution before they reached mass production. Automotus reached out to Trek10 when they were in a pilot-prototype phase but were edging closer to large production deployments each day.

As an AWS Premier Consulting Partner, Trek10 is well versed in the AWS ecosystem and played an integral role in securing project funding from AWS, which was critical to Automotus as an early-stage company. This assistance from Trek10 proved to be a major driver in completing the project before Automotus went into mass production.

"My first impression of Trek10 was their persistence - both in terms of getting us interested in the first place and figuring out a solution where the project budget actually made sense. They made it dead simple to get the ball rolling."

-Harris Lummis, Co-founder & CTO, Automotus

Benefits and Results

Within a week of rolling out the solution, Automotus was able to leverage it to anticipate, identify, and resolve several issues that may have otherwise compounded into more serious problems. By simply reducing the burden, this solution will help Automotus effectively operate at scale. When asked about how the project went, Lummis stated, “about as smoothly as you could hope for. I had the chance to work closely with a talented and charming group of people, and we ended up with a great product that was delivered on time.”

The true customer value is improving the uptime of the devices. When outages do occur, Automotus is now fully equipped to identify what happened and make improvements to their system to prevent similar instances from happening down the road.

The IoT audit aspect of the solution confirms their devices’ conformity to security best practices. This enables the Automotus team to dedicate more mental energy towards delivering value for their customers, and customers can rest easy knowing that their data is well protected.