In recent years, platform as a service (PaaS) incumbents have rapidly built up their internet of things (IoT) capabilities and started to become very competitive in the space. Just in the last year, IBM, Microsoft, and Amazon have each released very sophisticated IoT platforms as well as a complementary number of strategic alliances with IoT manufacturers and system integrators that have brought credibility to their IoT offerings. Different from previous trends like mobile, PaaS incumbents have entered the IoT space early and with very robust offerings.
While the current portfolio of IoT services in the PaaS infrastructure can be considered a solid first iteration, they are still in very early stages in order to power mainstream IoT solutions in either the enterprise or consumer market. From this perspective, there is a well-defined and very interesting set of capabilities that the industry will welcome as part of the roadmap of IoT PaaS platforms.
Amazon, Microsoft, IBM and the current IoT PaaS landscape
Amazon, Microsoft, and IBM have been the main PaaS incumbents to venture into the IoT space with offerings like AWS IoT, Azure IoT Suite, and Watson IoT Platform respectively. These releases have placed AWS, Azure, and Bluemix a step ahead of PaaS competitors like Google Cloud or Salesforce in the IoT market. Despite the rapidly evolving nature of the IoT landscape, the current generation of IoT PaaS is all focused on enabling the most common building blocks of IoT solutions such as device management, support for IoT protocols, stream analytics, or data storage.
While there are other IoT services included in PaaS platforms, the previous categories offer a good taxonomy of the current feature set of IoT PaaS platforms. While the current feature set can certainly be considered as a solid foundation for IoT solutions, there are key capabilities of IoT scenarios that are still missing in the present set of IoT PaaS solutions. These capabilities could make for a very intriguing addition to the roadmap of technologies — technologies like AWS IoT, Azure IoT Suite, or Watson IoT Platform.
Thinking ahead: An immediate roadmap for IoT PaaS solutions
IoT solution templates and rapid application development
A significant percentage of IoT solutions can be considered an assembly of well-defined patterns. For instance, an IoT process can collect data from sensors from a well known device manufacturer, process them using stream analytic jobs, store the results using IoT tracking services, and execute some business logic involving proprietary APIs. Many of these building blocks require considerable engineering efforts and need to be repeated time and time again across different IoT solutions.
A catalog of reusable IoT assets, and the tools and processes to incorporate those assets into IoT solutions, can be a quick but incredibly valuable addition to IoT PaaS solutions. This model is not exactly new to the PaaS incumbents, AWS created a very similar mechanism for assembling the building blocks of mobile solutions using the AWS Mobile Hub Console. Similarly, IBM’s Node-RED IoT orchestration engine includes a catalog of functions and templates that can be used across different Node-RED processes.
IoT data flow and event orchestration
Event orchestration and choreographies are an essential element of IoT solutions. However, neither AWS IoT nor Azure IoT Suite, nor Watson IoT Platform include orchestration solutions as part of their suites. IBM and Microsoft have certainly made some inroads in the IoT data flow space with technologies like Node-RED and Azure Logic Apps respectively, but neither one is currently integrated in their IoT platforms.
Providing visual models for authoring data flows and orchestration of IoT events should be a simple addition to the roadmap of IoT PaaS solutions. This value proposition is even more trivial if we consider that the industry has already produced open source technologies like Apache NiFi and IBM’s Node-RED that are optimized for IoT data flow scenarios and can be easily incorporated into IoT PaaS suites.
Testing, event replay, and simulations
Testing remains one of the most difficult challenges of IoT solutions. Recreating runtime conditions that involve the processing of hundreds of thousands or even millions of events require a considerable engineering and infrastructure effort. To address this challenge, IoT PaaS solutions should provide test case modeling, event replay, and simulations as part of their infrastructure.
From the functional standpoint, an event replaying and simulation service should allow the recording or runtime IoT events grouped into specific test cases as well as the replaying of those event streams to recreate runtime conditions. Incorporating these capabilities as native IoT services should increase the testability and operational readiness of IoT PaaS solutions.
IoT domain-specific analytics
Analytics is an essential element of IoT platforms and one that has favored the relevance of the PaaS incumbents in the IoT market. Companies like IBM and Microsoft have done a remarkable job integrating their IoT capabilities with their advanced analytic technologies enabling extremely sophisticated analytic experiences in IoT solutions. Despite the initial progress, the IoT PaaS stacks can still benefit from including a richer set of IoT specific capabilities.
Similar to its mobile and cloud predecessors, there are specific types of analytic models that are common across different IoT solutions. Ranging from areas like traffic optimization to device performance monitoring, there are hundreds of domain-specific IoT analytic capabilities that can be enabled in PaaS stacks like Azure, AWS, or Bluemix.
Embedded and hardware solutions
Since the inception of their IoT offerings, Amazon, Microsoft, and IBM have established relevant strategic alliances with some of the top IoT manufacturers in the market. Most of these alliances have resulted in the creation of samples or SDKs that integrate specific IoT toolkits with IoT PaaS backend services.
The next logical step in this evolution is to create a number of hardware products or IoT sensors or chips already integrated with IoT PaaS platforms. Providing integration at the hardware or sensor level will not only expand the adoption of IoT PaaS services across different IoT products but will result in an incredibly strong validation of the space.
The IoT PaaS space has experienced an accelerated growth in the last few months but there is still a lot of work to be done to make those stacks part of mainstream IoT solutions. The capabilities listed in this article could be a step in that direction.
This article is published as part of the IDG Contributor Network. Want to Join?