Testing-as-a-Service for the Internet of Things: A New Level of Automation

Dec 22

Posted by Hemanth Kumar Yamjala

The Internet of Things (IoT) market is expanding rapidly with new devices being installed to automate devices, systems, and networks. According to Statista, the total installed base of IoT devices is projected to touch 13.8 billion units by the end of 2021, and leapfrog to 30.9 billion units by 2025. The IoT ecosystem is likely to change the way people lead their lives by taking smart computing to the physical world. This will make the scope of computing even more decentralized and distributed. The profusion of so many devices connected to the internet generating, processing, and transmitting data streams needs to be stringently tested for quality. This is to ensure the data streams lead to correct decision-making in a real-time environment. Given that distributed and remotely located IoT systems have to process humongous volumes of data, they are open to being exploited by threat actors. The multifarious challenges related to the security and scalability of IoT systems call for implementing anIoT testing framework.

The need for automation in IoT testing

With the increased scale of IoT implementation, aspects such as non-functional requirements and security have become critical. The internet of things usually comprises a blend of components, namely, gateways, sensors, applications, data centers, and networks. This entails the use of automation in the IoT testing approach to manage various types of testing for these components. Needless to say, the traditional software testing methodology cannot suffice to handle a multitude of tests in a distributed environment. These include testing for security, performance, stability, conformance, and integration, among others. To cater to the complexity of internet of things QA testing, service-based testing has become important.

Why opt for service-based testing in IoT?

When the embedded software within the IoT systems goes through various development phases, namely, integration, testing becomes challenging. During such phases, several bugs or vulnerabilities are detected, unlike during testing at the unit level. Moreover, the risks of bugs can emanate from the unsecured hardware resources being used within an in-house test lab. Service-based testing or TaaS, can fit into the requirement for specialized testing vis-à-vis IoT devices. TaaS for the IoT may involve a test platform, a combination of infrastructure and software, or virtualization of the entire test department. The TaaS model for IoT testing may operate in the following way:

A user scenario and environment are developed for testing
A SaaS portal is provided to run tests and retrieve test data
Leverage an on-demand automated test lab
A test library is created with configurable parameters and user scenario descriptions
A metering function is enabled to track used resources and their costs
A test is designed and run in the test environment offered by the cloud vendor
The test performance is monitored and the company’s capability to meet the goals mentioned in the test design is evaluated

Benefits of using Testing-as-a-Service for IoT

The complexity of testing IoT devices due to the multiplicity of components in such devices requires the use of Testing-as-a-Service. The benefits are as follows:

Increased productivity: Automation in testing for IoT devices can create testing loads and service-based tests that are scalable and repeatable. Runtime analysis tools are able to detect bugs, quickly and easily. After fixing the bugs, regression testing can be performed in alignment with the objectives of agile, continuous testing, and iterative development.

Removal of hardware bottlenecks: During test scheduling, the presence of dependent components in the form of higher-level functions or other devices can be a challenge. However, using Testing-as-a-Service, such bottlenecks can be removed through simulation, wherein individual testers can make use of virtual labs.

Robust security testing: A service-based approach to IoT security testing using automation can test every parameter related to security using realistic scenarios. This can ensure the identification and fixing of bugs or vulnerabilities easily. For example, a denial-of-service attack can flood an IoT device,such as a thermostat in an HVAC system. It is only through robust IoT security testing, comprising penetration testing, that the device can be made to withstand such an attack. Moreover, the scalability of a virtual TaaS environment can exceed an in-house test lab comprising real hardware.

Service virtualization: Since any IoT device testing solutions needs to address the testing requirements of a multiplicity of components present in an IoT system, service virtualization can be of help. It can create a virtual lab with repeatable test environments, which would be as effective as the real thing but cost a fraction of the price.

Conclusion

The accuracy and integrity of data streams generated by IoT devices are critical to their functioning and performance. It is only by leveraging testing as-a-service that challenges related to productivity, scalability, and security in testing IoT applications can be addressed.