.png)
Summary: Viam simplifies fleet and data management for edge devices by providing the software backend infrastructure to securely connect any hardware with reasonable compute to the cloud across any network. This helps avoid the versioning and data gathering pitfalls that usually come with connectivity lapses in isolated environments.
We regularly talk to companies that are trying to figure out how to manage their smart machine fleets operating in areas of spotty connectivity, as well as the data they generate. Often, these companies feel compelled to build their own platform from scratch to monitor, manage, and interact with their machines.
Without a system like Viam, these companies feel compelled to build the backend to manage fleets of devices themselves, often at great cost and expense, when these expensive resources could be spent on building a better machine or customer solution. They usually have to cobble together tools from multiple vendors to build a patchwork solution that gathers machine data, uploads it to the cloud, stores it, and displays it. In addition, they need to create similarly messy solutions to deploy commands or software updates across asynchronous fleets.
Then other obstacles emerge. Hardware systems created by a multitude of manufacturers aren’t designed to be interoperable, and that often traps data in silos where it can’t be fully exploited - frustrating the ability to view assets through a single-pane-of-glass. The hardware and software these vendors provide were not conceived for the connected era, and may create one or several potential breaking points that prevent system-wide management.
And that’s when your machines are in one place, connected to the same network. The edge adds a whole other layer of potential disruptions.
Smart machines struggle to stay updated at the edge
It’s one thing to manage and update your machines that are all located within a single building, such as at a warehouse. Those machines are generally connected to the same network or there’s a worker onsite to manually update them with a USB drive. This makes version updates a relatively straightforward process, at least when it comes to machines from the same make and model.
However, other organizations need to put their smart machines in far-flung locations for periods of time where connectivity is inconsistent, weak, or expensive. Think about a shipping company’s boats navigating the ocean, a tractor plowing rural farms, or a backhoe digging holes at an isolated construction site. Increasingly, these machines are getting smarter, in that they’re running ML algorithms on cameras and sensors to optimize their performance in the field.
Having many machines spread across different geographies makes the delivery of software updates much more complicated. It becomes a lot more time-consuming and costly to physically send someone to a site to manually update the machine. Other companies may hack together a solution that creates a monotonous process to update machines one by one over-the-air, which quickly turns into an unscalable and unreliable solution as deployments grow. It becomes easy for machines within fleets to start running different software versions, creating more management complexity.
No connectivity leads to a data black hole
Poor connectivity can impact the “smart” part of your smart machine in a myriad of ways.
If the flow of critical data from the edge across a network is disrupted, that data is usually lost, which leads to gaps in reporting and modeling. For example, if an inspection robot loses connectivity for a few minutes, mission-critical data from that time period will be absent, resulting in a costly and inconvenient rerun of the mission. Plenty of teams understand this problem ahead of time and either have to modify hardware requirements, write and maintain in-house middleware, or make costly investments to bring stable networks to smart machine sites.
A fleet of machines spread out across many different locations around the globe generates tons of valuable data, but also creates hurdles when it comes to capturing and uploading that data to the cloud. It is hard enough to make sure your machine can upload data over-the-air; throwing in the risk of unreliable network connectivity is often a recipe for incomplete or tainted data sets.
Viam’s out-of-the-box pipelines designed for smart machines on the edge
Viam provides an out-of-the-box data and fleet management software backend for smart machines so that enterprises don’t have to build and maintain their own.
For areas with low connectivity, Viam provides the software to cache and queue data. Once the machine is connected to a network again, it will upload and sync seamlessly to the cloud, ensuring complete and accurate data for analysis, monitoring, algorithm building, or whatever the data may be used for.
The same concept applies in the opposite direction - from cloud to machine for versioning and software updates. If you are trying to deploy code to your fleet of machines in isolated locations at various points around the world, you can send out an update through Viam that will be carried out asynchronously as to the proper machines as they come back online. This ensures that mission-critical edge devices have all of the modern features they need, even if they undergo long periods of time without a connection.
The Viam platform’s edge connectivity features in action
I have created a short demo video showing how Viam continuously collects data from machines even when they are not connected to the internet, and sync whatever has been collected once a connection is reestablished.
Viam provides a seamless, stable system for fleet and data management no matter where smart machines are running, making it a good option for any enterprise dealing with many edge devices in remote locations.
If you’re a developer and would like to start exploring how Viam can help you manage smart machines on the edge, sign up for free here.
