Metadata Flashcard-01

The Value of Device Metadata

Posted by Scott Cranton on June 30, 2020

Customer Relationship Management (CRM) systems like Saleforce.com model the lifecycle and relationships for your sales and partner ecosystems. Device Management systems like EdgeIQ model the lifecycle and relationship for connected products and partner ecosystems. Just like CRM, Device Management systems need to model how customers, manufacturers, and partners interact with devices from creation through usage and eventual decommissioning. In a connected product world, partners can be resellers and field service providers who play a role in the sales and servicing of devices during their lifetime. This article is going to focus on one aspect of Device Management, the need for user-configurable Metadata associated with Devices. Add Metadata to Devices is analogous to CRMs allowing users to add and update data fields (Metadata) to Contact and Opportunity records.

Before we dig in, let me clarify what I mean by Connected Product, which is in part an evaluation of the Internet of Things (IoT). We're still talking about things (devices) connected to the Internet. The market has also evolved with systems focusing on Industrial IOT (IIOT) targeting manufacturing automation and Connected Product targeting new business models such as subscription models. Subscription models where the manufacturer manages the device AND its servicing; you pay for when you use it versus capital upfront. A large scale Connect Product is a smart building like [Hudson Commons] in New York. You rent the space AND deep access to all the building services like HVAC, lighting, security, visitor access, and even the elevators. By deep access, you can use your mobile phone to control all the systems involved in your rented space. Now that's a Smart Building.

On the software side, there's been a lot of evolution. Most think of IoT platforms as all-in-one suites. They include capabilities to manage and monitor devices, store data, analyze it, and display that data with various widgets. Sounds great, until you want to swap out any of those components because it doesn't quite meet your needs, then not so great. And these days, data is queen, and devices generate a lot of data. By that, I mean, most companies have realized that using lots of data to create useful AI/ML models is very helpful and valuable. So why would you want your IoT platform to own your relevant device data? In the early day of the CRM market, all-in-one Contact Managers like ACT! made it easy to get started but hard to scale. Modern IoT platforms are evolving into newer best of breed capabilities that make it easier to scale and mix-and-match as needed.

Coming back to Device Management and Metadata. As some software providers break out Device Management capabilities into its standalone offering, you're starting to see refinements like editable Metadata associated with devices. Metadata means extra data fields that are associated with a single Device entity. For example, I have a Raspberry Pi Model 4 sitting on my desk, so I might call that “Scott’s Raspberry Pi” as a Device configuration. I could associate a Metadata key-value pair with that Device, say key “Owner” and Value “Scott”. That way I can easily search and filter all Devices who’s “Owner” equals “Scott”.

Let me provide a use case to give this some more context.

For this example, let's use a smart tag to be used for shipping medical supplies. This smart tag is our device to be managed. The smart tag is connected over cellular and includes a thermometer and an accelerometer. The smart tag allows the user to validate that the drug shipment stayed within the desired temperature range. The tag also confirms it did not get dropped or impacted in some way that would damage the medical supplies. Great! So drug manufacture buys many of these smart tags and tracks them all in their Device Management account. When they are ready to ship some supplies, they will use some of the smart tags and attach them to the shipping palettes. Let's say they use 6 of their 100 tags for a given shipment. This is where it starts to get complicated. How does the user know which tags are for shipment A, for shipment B, and so on? You guessed it, using device Metadata can help make this all manageable.

When we create a new Device entry in our EdgeIQ device management system, we can add additional fields. For our current example, we can add a `metadata.shipment` field and set it to the shipment name say `Shipment A`. That could make it easier to sort and filter devices by the shipment name. Filtering smart tag's by shipment name makes it easier to view health status and apply policy-based alerts like "if the temperature for shipment A exceeds 50 degrees than send an alert to the driver and dispatcher".

A note of caution. Think carefully about where you keep your data and try not to duplicate data. In our example, it would NOT be appropriate also to associate the billing address of the shipment with the individual tags. You'd be repeating a lot of information - a billing address field for each device - which takes us space and dramatically increases the odds of data going stale. Also, you're unlikely to reference order billing addresses in your device management system. Billing address data is more appropriate within your CRM system. Now, if your sales and support people tend to track and reference things by order number, then associating an order number with the devices could be helpful. That device order number field would be an intuitive way to sort, filter, and act on groups of devices based on your existing processes.

I find it is those finer details that highlight a well thought out system. The delight happens when you say, "I wish I could add this one extra piece of data to my device" to make your day to day life easier. And then you realize you can.