A model-driven framework for data-driven applications in serverless cloud computing.

In a serverless cloud computing environment, the cloud provider dynamically manages the allocation of resources whereas the developers purely focus on their applications. The data-driven applications in serverless cloud computing mainly address the web as well as other distributed scenarios, and therefore, it is essential to offer a consistent user experience across different connection types. In order to address the issues of data-driven application in a real-time distributed environment, the use of GraphQL (Graph Query Language) is getting more and more popularity in state-of-the-art cloud computing approaches. However, the existing solutions target the low level implementation of GraphQL, for the development of a complex data-driven application, which may lead to several errors and involve a significant amount of development efforts due to various users' requirements in real-time. Therefore, it is critical to simplify the development process of data-driven applications in a serverless cloud computing environment. Consequently, this research introduces UMLPDA (Unified Modeling Language Profile for Data-driven Applications), which adopts the concepts of UML-based Model-driven Architectures to model the frontend as well as the backend requirements for data-driven applications developed at a higher abstraction level. Particularly, a modeling approach is proposed to resolve the development complexities such as data communication and synchronization. Subsequently, a complete open source transformation engine is developed using a Model-to-Text approach to automatically generate the frontend as well as backend low level implementations of Angular2 and GraphQL respectively. The validation of proposed work is performed with three different case studies, deployed on Amazon Web Services platform. The results show that the proposed framework enables to develop the data-driven applications with simplicity.

Clinical Utility Of Cavea2t2 Score For Assessing The Survival Of Brachiocephalic Arteriovenous Fistula.

BACKGROUND: There are not many error proof clinical scores to assess the native dialysis access. CAVeA2T2 score is a recent tool in use. Objective of the study is to assess the clinical utility of CAVeA2T2 scoring system in predicting the survival rate of brachiocephalic arteriovenous fistula (BC-AVF). METHODS: All consecutive patients fulfilling the inclusion criteria for BC-AVF from January 2016 to January 2018 were included. According to their CAVeA2T2 score they were divided into two groups (Group A: < 2 and Group B: ≥2). Cumulative primary and secondary patency survival of BC-AVF for both groups were measured. RESULTS: A total of 112 BC-AVFs were analysed. Mean age was 42±SD 14 years (M: F =5:1). Mean CAVeA2T2 score was 1.45±1.8. In terms of primary patency, there was no statistically significant difference between two groups (p=0.074, p = 0.229 and p=0.357 at 6 weeks, 6 months and 12 months respectively). However, the difference was significant in terms of secondary patency (p=0.002, p=0.036 and p=0.032 at 6 weeks, 6 months and 12 months respectively). On comparing the cumulative survival between two groups; a significantly low primary patency rate survival (Log Rank x2 = 12.9, p-value = 0.001) and secondary patency rate survival (Log Rank x2 = 7.6, p-value = 0.001) of BC-AVF was found in Group B. CONCLUSION: We found CAVeA2T2 score an easily applicable and useful tool to assess the patency and survival of BC-AVF. Patients have a poor patency and significantly low survival rate when their CAVeA2T2 score was ≥2.