The Internet of Things (IoT) is driving the digital revolution. Almost all economic sectors are becoming "Smart" thanks to the analysis of data generated by IoT. This analysis is carried out by advance artificial intelligence (AI) techniques that provide insights never before imagined. The combination of both IoT and AI is giving rise to an emerging trend, called AIoT, which is opening up new paths to bring digitization into the new era. However, there is still a big gap between AI and IoT, which is basically in the computational power required by the former and the lack of computational resources offered by the latter. This is particularly true in rural IoT environments where the lack of connectivity (or low-bandwidth connections) and power supply forces the search for "efficient" alternatives to provide computational resources to IoT infrastructures without increasing power consumption. In this paper, we explore edge computing as a solution for bridging the gaps between AI and IoT in rural environment. We evaluate the training and inference stages of a deeplearning based precision agriculture application for frost prediction in modern Nvidia Jetson AGX Xavier in terms of performance and power consumption. Our experimental results reveal that cloud approaches are still a long way off in terms of performance, but the inclusion of GPUs in edge devices offers new opportunities for those scenarios where connectivity is still a challenge.

BackgroundThe highly active antiretroviral therapy (HAART) has altered the course of HIV infection, transforming it from a fatal illness to a chronic condition, reducing morbidity and mortality. However, this therapy has led to an increased incidence of metabolic problems such as insulin resistance, dyslipidemia, lipodystrophy and impaired glucose metabolism.The objectives of this study are to determine the prevalence of insulin resistance (IR) in a cohort of human immunodeficiency virus (HIV)-infected patients on highly active antiretroviral therapy (HAART) and to investigate the potentially associated factors.MethodsWe conducted a cross-sectional study including 219 adult patients with HIV on HAART. IR was determined through the homeostasis model assessment (HOMA-IR) mathematical model, using fasting plasma glucose (FPG) and insulin. Bivariate and multivariate analyses were performed to assess the association between demographic information, clinical characteristics and laboratory results, and IR.Results75 (34.2 %) [95 % confidence interval (CI) 28.9–40.9] HIV-patients on HAART showed IR. 61 (81 %) of these patients were on HAART for more than one year, which was mainly composed by non-protease inhibitors drugs (88 %). Metabolic syndrome (MS) was found in 59 (26.9 %) subjects. In the multivariate analysis, the factors associated with IR were age ≥ 46 years (Prevalence ratio = 2.767, 95 % CI 1.325 to 5.780) and greater body mass index (BMI) (Prevalence ratio = 1.148, 95 % CI 1.054 to 1.250).ConclusionsThe prevalence of IR was 34.2 %. Factors associated with IR were age and BMI. We did not find any significant association between IR and protease inhibitors (PI), which may be explained by the small number of patients using PI as part of their HAART regimen included in our study.

No abstract

Contribution This study reveals that the programming paradigm is relevant to obtain advanced programming skills. Background Parallel computing has become mandatory for computer science students. The increasing amount of computational resources required by emerging applications need experienced programmers that fully exploit hardware resources. However, the hardware platforms and programming languages to leverage them evolve at a dizzying pace, making very challenging for students the successful learning of the continuously changing high‐performance computing concepts. Research Questions (a) Is the learning curve of the programming language too steep to begin learning parallel programming fundamentals? (b) Are emergent learning methodologies making even more difficult to learn parallel programming in general? Methodology It is analyzed the main challenges for succeeding in parallel programming courses at the undergraduate level in two different learning modalities, namely on‐campus and online. It is analyzed the main tools available within a learning management system, showing their impact on online studies. Findings Our results reveal that the steep learning curve for parallel programming is one of the main barriers to student success, leading to an early drop out of the subject. On‐campus studies mitigate this problem through a close relationship between students and educators. Online studies, however, do not have this tight relationship by its definition.

Among the service models provided by the cloud, the software as a service (SaaS) model has had the greatest growth. This service model is an attractive option for organizations, as they can transfer part or all of their IT functions to a cloud service provider. However, there is still some uncertainty about deciding to carry out a migration of all data to the cloud, mainly due to security concerns. The SaaS model not only inherits the security problems of a traditional application, but there are unique attacks and vulnerabilities for a SaaS architecture. Additionally, some of the attacks in this environment are more devastating due to nature of shared resources in the SaaS model. Some of these attacks and vulnerabilities are not yet well known to software designers and developers. This lack of knowledge has negative consequences as it can expose sensitive data of users and organizations. This paper presents a rigorous systematic review using the SALSA framework to know the threats, attacks and countermeasures to mitigate the security problems that occur in a SaaS environment. As part of the results of this review, a classification of threats, attacks and countermeasures in the SaaS environment is presented.

No abstract

No abstract