Future internet services and applications

In this special issue of Concurrency and Computation: Practice and Experience (CCPE), we focus on three complementary aspects that have to be considered while setting up future Internet services: (i) their modeling, provisioning, and management; (ii) data protection; and (iii) data collection, storage, and analysis. The special issue includes extended versions, containing at least 50% new material of selected accepted papers from the Future Internet Services and Applications(FISA) track of the 29th IEEE WETICE Conference. It also includes four new submissions dealing with FISA-related issues. In total, we received 11 submissions from six different countries. After two rounds of reviews, we accepted seven articles: four extended submissions from FISA'2020 and three new submissions. During the review process, all submissions have been carefully reviewed by at least three reviewers. In the paper Multi-perspective business process discovery from messaging systems: State-of-the art,1 the authors discuss the opportunities provided by messaging systems and logs for automated business process discovery. A business process (BP) is known as a sequence of activities designed to create something of value to deliver to customers. Business always needs structured documentation that requires experienced people to achieve. Hence, emails exchanged between workers involved in a BP are considered as much as a package that includes a detailed record of all aspects of the BP. In addition, logs represent a valuable resource offering information that will help in the analysis, optimization, and management of a BP. The authors conducted a literature review and identified multiple approaches based on messaging systems; however, they conclude that additional efforts are still required in this context. In the paper Semantic thingsourcing for the Internet-of-Things,2 the authors discuss the concept of crowdsourcing that represents a practice of co-creation by offering the general public (a priori anonymous people) the possibility of participating in the process of creating a project or a brand fulfilling the requirements of given consumers. In the last few years, a new concept, called thingsourcing, has emerged and has caught the attention of researchers. Indeed, thingsourcing gives great importance to collective behavior and how to exploit it. This work focuses on the description and discovery of things semantically based on a proposed ontology and takes into consideration specific scripts that aim at managing the interaction behavior of each thing with its peers. This work is supported by a case study about the dairy supply chain. In the paper SBM: a smart budget manager in banking using machine learning, NLP, and NLU,3 the authors focus on machine learning algorithms with natural language understanding (NLU), which have been introduced in the banking sector. Indeed, finance companies should adopt novel techniques and methods to serve and keep their customers, especially through the intense flow of data obtained today. Banks and fintechs are in fierce competition to transform customers' data that incorporate their transactions into pertinent and significant knowledge for decision making. This article proposes a set of techniques to fill gaps encountered by the banking system, namely, how to classify banking transactions automatically for smart budget management. A system implementing these techniques is available in a Tunisian bank. This system relies on the fusion of natural language understanding and incremental machine learning algorithms. In the paper QoE-aware traffic monitoring based on user behavior in video streaming services,4 the authors studied a web QoE (Quality of Experience) monitoring approach and applied a test methodology based on real-time QoE estimations depicted by end-users. Indeed, during the previous years, online videos have been very successful thanks to various global media connections. In addition, the understanding of QoE of the different receivers represents a powerful tool for over-the-top (OTT) providers to satisfy multiple end-user requirements. The authors identified various factors which affect the QoE and which were overlooked: human, contextual, and systematic, for example. As a solution, the authors propose a QoE prediction model that will be trained on data gathered from the monitoring of Web applications. This model helps in building an optimization solution for the video transmission chain in the context of software-defined network (multi-access edge computing). Different supervised machine learning algorithms have been used in this work, such as decision trees, k-nearest neighbors, and support vector machines. In the paper Configuration approach for personalized travel mashup,5 the authors proposed a configuration-based approach for personalized service-oriented travel mashup, developed as a recommender system. Currently, recent technology is based on a multitude of databases and services, allowing human beings to have rapid fluidity in accessing the functionalities necessary for their daily lives. Indeed, the different platforms serving users are not suitable for everyone because they favor expressive power over intuitiveness. This is why users must have deep perceptions to be up to date with the use of these systems. Mashup solutions would meet these users' requirements in the easiest way. This article presents a mashup methodology based on a user-understandable formal language. This language includes visual elements called domain-specific visual language, allowing people to build mashup-based applications. The proposed mashup methodology relies on a mashup schema, which is a link between users and services allowing the users to put in the background all the problems resulting from the integration of services. Finally, a proposed configuration module produces a personalized trip plan by performing a travel mashup query. In the paper Orchestrated sandboxed containers, unikernels and VMs for isolation enhanced multitenant workloads and serverless computing in cloud,6 the authors focus on container virtualization. Container virtualization is a method of resource partitioning at the level of the operating system that virtualizes the environment of execution such as CPU, RAM, and file system, not the machine. Among the major advantages of this method, containers are easier to migrate, and faster to restore. Running environments isolated from each other in containers that require a high level of authorization and share the same kernel makes containerization more vulnerable to security problems. In this work, the authors studied and examined the OCI (open container initiative) runtimes and the default Kubernetes runtime runC, by deploying containers for each of them on a Kubernetes cluster and additionally on a Docker node. Similarly, they deployed unikernels and security-oriented lightweight Linux-based VMs created and managed by the FireCracker monitor on a Kubernetes cluster. In addition, they propose a solution to automate the deployment and replication of containers on Kubernetes clusters. In the paper Capitalizing the database cost models process through a service-based pipeline,7 the authors discuss database design that has undergone a major revolution thanks to the diversity of subjects imposed by users in the IT sector. It aims to improve performance metrics, namely, execution time, accuracy of data, energy consumption, and so on, in order to design an adequate database cost model. The effort made by the database designers is confronted with problems related to the different information sources. Designers did not deny that the major problem in understanding and reproducing cost models lies in the absence of a step-based design of the formal schema. To succeed in carrying out such a project, this work studies the synchronization of the processing of collected information. The proposed solution combines natural language processing and machine learning techniques. The authors demonstrate the feasibility and benefits of their implementation through a case study. The authors would like to thank CCPE Editor, Professor David W. Walker, for his strong support. A special thank-you goes to all the authors for their valuable contributions to this special issue. Special thanks go to all the reviewers for the time and effort in providing their reviews that helped in enhancing the quality of the papers.