Best Practices for IIoT Applications

In the previous chapter, we discussed how to build an IIoT application using Predix services. This chapter takes it further and looks at various different best practices to successfully build a multitenant, maintainable, reliable IIoT application. We will look at the following areas in this section:

• Best practices for API development
• The power of the polyglot programming choice
• Eventual consistency for higher performance
• Strategies to handle multiple versions of the application at scale
• The advantages of using established trust between microservices
• UX strategy for application adoption
• Tracing and logging end to end

Best practices for API development

We will start with discussing API development best practices because the API is the cornerstone for good microservices-based architecture, which in turn forms the basis of the IoT cloud tier. APIs in a microservices context are always RESTful interfaces. RESTful APIs use constructs of HTTP at their core and JSON as their payload for request and response. We will look at a few key aspects of developing an awesome API. We can develop RESTful APIs using frameworks such as Spring Boot, Node—Express, or Python—Flask.

The API endpoint should be descriptive

The endpoint of the API should be self-descriptive and should use HTTP methods such as GET, PUT, POST, and DELETE to perform the...

Sorting, filtering, searching, and versioning

In addition to the CRUD operation, we need to handle various other regular operations, such as searching for certain device names. In addition, we can get the list of devices and sorting them into ascending or descending the results. Filtering is another interesting feature we can utilize to filter the responses, so that, by using includes, we can get only the fields we are interested in. Finally, we can have the endpoint support searching capability so that we can request certain types of devices with a name or asset of certain types.

Sorting of devices by name

Sorting is essentially a request for a sorted list of items, for example, a list of devices sorted by name. By default...

The power of polyglot programming

IoT development is, by nature, polyglot, which means it is using different programming languages, frameworks, and so on, for developing IoT applications. It is essential to appreciate the power of polyglot programming and use it wisely to get the maximum mileage. In this section, we will go over some of the benefits and traps of leveraging polyglot programming to improve IoT development.

Using Java programming for microservices development is advisable due to the plethora of tools available for Java, in addition to frameworks such as Spring Boot. Java has good support for many databases, and security tools such as OAuth integration. In addition, Java has many IDEs, built tools, code coverage tools, and so on, and this make it the winner for microservices development. Java stands out as a clear winner for API/microservices development. In addition...

Strategies to handle multiple versions of the @scale application

Typically, to manage applications at scale, we need to make sure we can seamlessly update, deploy, and run multiple versions of the applications in parallel for testing and certification before releasing the application for production use. Automation using CI/CD is mandatory for creating a continuous deployment pipeline. There are certain challenges, such as when to cut over to the newer version, and DB upgrades are still a challenge that need to be overcome. In this section, we will cover the concepts of Blue-Green deployment and best practices of database migration.

Blue-Green deployment

The idea behind Blue-Green deployment is to have two identical environments...

Summary

In this section, we covered some of the key best practices to follow while developing IIoT applications.

In the next chapter, we will cover the future of IIoT.