11 Popular IoT development platforms

The Internet of things has undergone a rapid transformation since Kevin Ashton first introduced this concept in the 1999 year . With the exponential growth in diversity and quantity of devices connected to the Internet of things in recent years, the Internet of Things has become a mainstream technology that has great potential to promote the way of life in today's society.

There is still a clear line between hardware and software platforms in the technology and engineering of the Internet of things, with most vendors focusing on hardware. Only a handful of vendors provide internet of things software services: for example,Mattermark is the only provider of Internet of things software services in the top four specialty networking startups based on the total investment rankings .

This article provides a comprehensive survey of the existing IoT software platform based on our detailed analysis of IoT vendors. The Internet of things vendor, which was chosen at the end of this article, is based entirely on the criteria of whether these vendors provide software solutions to handle information obtained from IoT devices / sensors. Note: Although we want to be as comprehensive as possible, we may have missed some of the latest improvements to these platforms in this article.

The important features that the IoT software platform wants

Based on several recent surveys, we have selected the most critical features of the IoT software platform: Device management, integration, security, data collection protocols, analysis types, and support visualizations to compare sample capabilities. These features are briefly described in the second half of this article.

Device Management and support integration

Device management is one of the important functions required by the IoT software platform. The IoT platform should maintain a stack of connected devices and track the operational status of these devices, as well as handle configuration, firmware (or other software) update issues, and provide device-level error reporting and processing scenarios. Before the end of each day, device users should be able to get personal device-level statistics.

supporting integration is another important feature required by the IoT software platform. Important operations and data that need to be published from the IoT platform should be accessiblethrough the API, which is commonly used by REST APIs.

Information Security

The information security tools required to operate the IoT software platform are more demanding than common software applications and services. Millions of devices connected to the IoT platform represent the scale of the vulnerabilities we need to address. In general, to avoid eavesdropping, the network connection between the IoT device and the IoT software platform needs to be guaranteed through a powerful encryption mechanism.

However, in modern IoT software platforms, most low-cost, low-power devices cannot support such advanced access control measures. As a result, the IoT software platform itself needs to take alternative measures to address this type of device-level problem. For example, by dividing IoT traffic into private networks, relying on strong security at the cloud application level, requiring regular updates of passwords and supporting verification of updated firmware, as well as signatures to update software, these tools enhance the security level of the IoT software platform.

data Collection Protocol

Another important aspect to note is the type of protocol used for data communication between the various components of the IoT software platform. The IoT platform may need to be scaled up to millions of or even billions of devices (nodes). Lightweight communication protocols should be used to achieve low power consumption and low bandwidth capabilities.

Note: Although we use the agreement as an overview term in this article, the protocols used to collect data fall into the following categories: Applications, payload containers, information delivery, and legacy protocols.

Data Analysis

Data collected from sensors connected to the IoT platform need to be analyzed intelligently to gain meaningful insights.

There are four main types of IoT data analysis: Real-time analysis, batch analysis, predictive analysis, and interactive analysis. Real-time Analytics: Perform online (dynamic) analysis of data streams. Sample operations include window-based integration, filtering, transformations, and so on.

Batch analysis: Operate on the accumulated data set. This allows the batch operation to run for a predetermined period of time, perhaps for several hours or days. Predictive analytics: Focus on forecasting based on a variety of statistical and machine learning techniques. Interactive analysis: Perform multiple exploratory analyses of data flow and batch data. The last one is real-time analysis, which occupies a heavier weight on any software platform.

today's IoT software platform

After careful investigation of the current IoT software platform, we found that each of the features mentioned above has been implemented, but in a different degree. We have listed the relevant platforms below and have done a functional summary comparison:

Table 1 related Platform function Summary comparison (the column labeled " Unknown " indicates that the relevant information cannot be found in the available files )

Obviously, many of the IoT startups listed above may not yet have device management capabilities. In this regard, there is also a need for IoT software platform vendors to provide solutions.

In addition, there is relatively little support for computing and visualization in the analysis of generated IoT data. Most of them support real-time analytics - A must-have feature for any IoT framework. However, only a very small number of IoT software platforms support 3 other types of analytics. The visual interface is mostly represented as a simple model for portals, allowing for the management of the IoT ecosystem, but rarely provides visual data analysis capabilities.

Among the different IoT software platforms, there are several common features, including integration-based REST API, which supports the use of the MQTT protocol to collect data and link encryption using SSL. Although not mentioned in table one, a single Parstream company can achieve a throughput of up to a maximum of one to thousands of rows per second.

This suggests that most IoT software platforms are not designed to take much account of the system performance of IoT deployments, which is critical in real-world scenarios.

features that need to be improved

There are obviously a number of areas that need improvement. In this section, we first provide a list of improved features. With the efforts of the IoT software platform vendors, some of these projects have been implemented and some performance is waiting to be implemented. We then provided a list of these new features that are not yet implemented.

existing features

Data Analysis

Most IoT software platforms now support real-time analytics, but batch analytics and interactive data analysis may be just as important.

at this point, one might argue that including such analysis capabilities in other well-known processing platforms, it is also easy to configure software systems for analyzing scenarios. However, this is not easy. For real-time analytics (Storm,Samza, etc.), for batch analysis (Hadoop,Spark, etc.), for predictive analytics (spark MLLIB, etc.), well-known data processing systems for interactive analysis (Apache Drill, etc.) are not directly used in IoT cases.

Benchmark

The IoT software platform needs to be scalable and includes devices that describe and evaluate system performance. Well-defined performance metrics require: ability to shape and measure the performance of IoT systems, taking into account network characteristics, energy consumption characteristics, system throughput, computational resource consumption, and other operational characteristics.

Edge Analysis

measures need to be taken to reduce the amount of network bandwidth loss between the sensor device and the IoT server. One solution is to use a lightweight communication protocol. Another option is to use edge analysis to reduce the amount of raw data transferred to the IoT server. Edge analysis can be implemented even in simple hardware embedded systems, such as Arduino.

Other questions

It should be noted that there are a number of other issues related to the IoT software platform, such as ethical, ethical, and legal issues that are not covered in this article. Although these issues are important, they are not discussed in this article.

features that need to be added

handling unordered processes

In any internet of things application, it is possible to run into an unordered event, in which the sequence of tuples may be caused by network delay, clock skew, etc. in the event stream emitted by the sensor. Handling unordered IoT events can cause system failures. When dealing with unordered events, you need to make a tradeoff between the accuracy of the results and the latency.

There are four main processing technologies: Cache-based ( buffer-based), punctuation-based (punctuation-based), conjecture-based (speculation-based), and approximation-based ( Approximation-based techniques). In an IoT solution, one or more of these should be used to solve the problem of unordered events.

support for the Internet of things background

The background is mainly composed of individuals, their preferences or past behavior. For example, in a mobile phone case, we can get rich background information because there are many different types of sensors in modern mobile phones. In the Internet of things analysis, these background data should be taken into account.

Conclusion

The rapid development of the IoT model requires a powerful IoT software platform to meet the emerging needs through IoT use cases. In this article, we investigate the capabilities of the most advanced IoT software platforms available, focusing on these areas: device management, integration, security, data collection protocols, analysis types, and visual support. From this study, areas such as device management, IoT data analytics, IoT software system scalability, and performance clearly require the IoT platform Community to devote particular attention.

Article Source: CSDN

11 Popular IoT development platforms