Deploying a Service to consume a published model
As mentioned before, ML models are typically large files and can take several minutes to get delivered to the target Edge nodes asynchronously, depending upon the network bandwidth. In this example, we will use a small text file with easily readable text content to show the delivery of the file to the target service. This simple exercise will help us practice different commands and examine the model delivery process.
In this exercise, we will use a new service-mms example service. This example is a variation of helloMMS to practice MMS foundational concepts. This knowledge will be useful in the complete ML-based object detection application described in Chapter 14.
This standalone service-mms service, in addition to handling MMS functions, also exposes an endpoint that can be queried by another service to retrieve the model file delivered by MMS. The service-mms service also marks or deletes the newly delivered model file so that...
