Contemporary solutions for cloud-supported, edge-data analytics mostly apply analytics techniques in a rigid bottom-up approach, regardless of the data's origin. Typically, data are generated at the edge of the infrastructure and transmitted to the cloud, where traditional data analytics techniques are applied. Currently, developers are forced to resort to ad hoc solutions specifically tailored for the available infrastructure (for example, edge devices) when designing, developing, and operating the data analytics applications. Here, a novel approach implements cloud-supported, real-time data analytics in edge-computing applications. The authors introduce their serverless edge-data analytics platform and application model and discuss their main design requirements and challenges, based on real-life healthcare use case scenarios.

When the requirements and processing appetites have grown enough in the era of personal computing, users began to think about cloud servers. This initiated the growth of cloud computing and set up the era of Everything-as-a-Service. When the IoT idea spread, lots of sensors and smaller devices were introduced for the realization of the ubiquitous and pervasive computing. The simplest way to collect and process all this data was to connect the devices to the cloud. This required high-speed networks and interconnections. However, no matter how fast they could go, engineers understood that theres no need to transfer all the data through links to the central data server, and a more efficient organization would be to distribute smaller servers in front of the central data server and closer to the user location. This began the evolution of pushing data collection and processing back to end-user devices, as a new computing approach called edge computing, which appears in different forms such as fog computing, cloudlets, and mobile edge computing.

Diabetes is one of today’s greatest global problems, and it is only becoming bigger. Constant measuring of blood glucose level is a prerequisite for monitoring glucose blood level and establishing diabetes treatment procedures. The usual way of glucose level measuring is by an invasive procedure that requires finger pricking with the lancet and might become painful and obeying, especially if this becomes a daily routine. In this study, we analyze noninvasive glucose measurement approaches and present several classification dimensions according to different criteria: size, invasiveness, analyzed media, sensing properties, applied method, activation type, response delay, measurement duration, and access to results. We set the focus on using machine learning and neural network methods and correlation with heart rate variability and electrocardiogram, as a new research and development trend.

The vision about having information at our fingertips anytime and anyplace was only a dream in the mid 1990's. Today, Web access, mobile applications and ubiquitous mail is a reality experienced by millions of users worldwide. The main obstacle for today's mobile devices is the lack of resources, since today's applications are becoming more and resource extensive. This requires increasing the processing power of the device, increasing the memory, a need for greater durability of the battery and so on. All of these tendencies would increase the processing speed of applications and would improve the way they function. Therefore, the need for a more powerful device with a better processing speed becomes inevitable. This is a valid reason for introducing cloudlets, which represent intermediate layers located between the cloud and each mobile device. In this way, instead of accessing a distant cloud, the users connect to the nearest cloudlet through a wireless network. The purpose of this article is to explain the importance of using Cloudlets in our everyday life, by presenting the main challenges and possibilities of further research. We present a short introduction about Cloudlets; the cloudlets architecture and the way they operate. Moreover, we discuss about what their features are, how they work with thin and tick clients, and how they operate as offload elements. In the end, we state the challenges and open research question related to the topic

There are various ways a computer can detect QRS complexes, all of which depend on two main factors - the algorithm, and the data. In practice, doctors can pinpoint patient's R-peaks in an electrocardiogram by analyzing the complex morphology of the ECG signal; computers accomplish the same by utilizing different QRS detection algorithms. In this paper, we address the data problem to select the sample rate and resolution that obtain the highest accuracy. For this purpose, we conduct experiments to find the impact of the sampling frequency and resolution on the quality of the QRS detection, by several open-source QRS detection algorithms with multiple variations of data representation. The final goal is to recommend the lowest sampling frequency and smallest sampling resolution (bit depth), that will have sufficient data representation to enable high QRS detection accuracy.