Jarkko Kuusijärvi

Situation-based and self-adaptive applications for the smart environment

Situation-based and self-adaptive applications are the key enablers of smart environments and ecosystems. In those environments, developers and users focus on innovating and making added-value applications, instead of solving the problems of interoperability and complexity of heterogeneous systems. This paper contributes by introducing an innovative adaptation framework for the situation-based and self-adaptive applications of smart environments. The framework embodies a novel architecture, generic ontologies for context, security, and performance management, and dynamic models for performing runtime reasoning and adaptation. The framework is intended for an application developer who is i creating application scenarios, and ii transforming the scenarios into annotated sequence diagrams with the help of the static models of the framework, the ontologies, and the rules defined in them. Thereafter, the application developer iii transforms the annotated application behavior description into the selected rule language, SPARQL. The approach is exemplified through the creation of the GuideMe application, which exploits context, security, and performance information to adapt the service according to the quality requirements and the context of the user, as well as the smart environment, without bothering the end-user.

A semi-automatic end-user programming approach for smart space application development

Abstract This article describes a semi-automatic end-user programming approach that: (i) assists in the creation of easy-to-apply Semantic End-User Application Programming Interfaces(S-APIs) for the APIs of legacy software components; and (ii) enables the usage of S-APIs in command-oriented and goal-oriented end-user application programming. Furthermore, a reference implementation is presented for the approach that provides visual programming tools and an agent-based execution environment for smart space applications. The use of the approach is exemplified and tested in a case study in which S-APIs are created for a home automation system and for a personal assistant application, and then utilized in end-user programming performed in desktop and mobile environments.

Piecemeal Development of Intelligent Applications for Smart Spaces

Software development is facing new challenges as a result of evolution toward integration and collaboration-based service engineering, which embody high degrees of dynamism both at design time and run-time. Short times-to-market require cost reduction by maximizing software reuse. Openness for new innovations presumes a flexible development platform and fast software engineering practices. User satisfaction assumes situation-based applications of high quality. The main contribution of this paper is the piecemeal service engineering (PSE) approach developed for and tested in application development for smart spaces. The intent of PSE is to maximize the reuse of existing knowledge of business and design practices and existing technical assets in the development of new smart-space applications. Business knowledge is mostly informal and domain-dependent, but architectural knowledge is generic, at least semiformal, and represented in principles, ontologies, patterns, and rules that together form a reusable architectural knowledge base for fast smart-space application development. The PSE facilitates the incremental development of intelligent applications by supporting abstraction, aggregation, and adaptability in smart-space development.

Supporting situation-awareness in smart spaces

This paper reports results from ongoing research, relating to the development of the context ontology described with web ontology language (OWL). Our aim is to reach relevant context ontology for smart spaces that are, by their very nature, heterogeneous, pervasive and ubiquitous. We illustrate the usage of the context ontology in a case where lights are switched on according to a calculated wake-up time and the preferences of an individual. With the introduced Context Ontology for Smart Spaces, CO4SS, we managed to perform reasoning actions based on the users context.

Framework for End-User Programming of Cross-Smart Space Applications

Cross-smart space applications are specific types of software services that enable users to share information, monitor the physical and logical surroundings and control it in a way that is meaningful for the users situation. For developing cross-smart space applications, this paper makes two main contributions: it introduces (i) a component design and scripting method for end-user programming of cross-smart space applications and (ii) a backend framework of components that interwork to support the brunt of the RDFScript translation, and the use and execution of ontology models. Before end-user programming activities, the software professionals must develop easy-to-apply Driver components for the APIs of existing software systems. Thereafter, end-users are able to create applications from the commands of the Driver components with the help of the provided toolset. The paper also introduces the reference implementation of the framework, tools for the Driver component development and end-user programming of cross-smart space applications and the first evaluation results on their application.

Context-awareness micro-architecture for smart spaces

Reaching context-awareness is still an open issue in pervasive systems like smart spaces (SS). With the help of context a smart space application can react to the current situation or be proactive and take into account coming circumstances. This paper describes the ongoing research on how contextawareness - an ability to identify and react - is enabled for the heterogeneous smart space applications. This work presents innovative context-awareness micro-architecture to design software in a way that it is able to react to changes based on information gathered via a context monitoring agent and inferred by a context reasoning agent. The usage of the context-awareness micro-architecture is validated with two scenarios which are instantiated to a personal smart space and a smart home.

A tool for security metrics modeling and visualization

Measuring the security level of an information system to acquire reliable perception of its state requires the use of various different security metrics that can provide extensive security evidence of the system. Visualization can then be used to facilitate the management of the security metrics and measurements and to enhance understanding on their relationships. This paper introduces a tool for modeling and monitoring the security state of a system and focuses on the visualization aspects of the tool. The security metrics of a system are organized hierarchically in the tool, so that more general and conceptual security metrics on the higher levels are connected to detailed, low-level measurements. The tool helps bring meaningfulness to the security metrics and helps the user be more aware of the security state of the system during runtime use of the tool. By having organized security evidence from high-level objectives to low-level measurements the user is able to act on the security incidents more proficiently.

Mitigating IoT security threats with a trusted Network element

Securing the growing amount of IoT devices is a challenge for both the end-users bringing IoT devices into their homes, as well as the corporates and industries exposing these devices into the Internet as part of their service or operations. The exposure of these devices, often poorly configured and secured, offers malicious actors an easy access to the private information of their users, or potential to utilize the devices in further activities, e.g., attacks on other devices via Distributed Denial of Service. This paper discusses the current security challenges of IoT devices and proposes a solution to secure these devices via a trusted Network Edge Device. NED offloads the security countermeasures of the individual devices into the trusted network elements. The major benefit of this approach is that the system can protect the IoT devices with user-defined policies, which can be applied to all devices regardless of the constraints of computing resources in the IoT tags. Additional benefit is the possibility to manage the countermeasures of multiple IoT devices/gateways at once, via a shared interface, thus largely avoiding the per-device maintenance operations.

Visualizing structure and quality properties of Smart Spaces

Smart Spaces embedded into our surroundings are dynamic in nature; variable types and amount of services and applications are provided by a set of interoperable devices and systems that are joining and leaving a space every now and then. Thus, the structure, functionality and quality properties of the space are changing at every turn. In order to understand, what devices/systems there are in the space, what applications and services the space provides and what is the quality level of these services, application developers and smart space users need to have an assistant that helps them in figuring out these things. On this account, we use a visualization tool IQVis for illustrating how the run-time properties of a smart space are changing in time. Application developers can use the tool for seeing how an application behaves in a real situation. A user of a smart space can use the quality information provided by the IQVis tool, e.g., on the screen of his/her mobile phone, for adapting his/her profile/actions. The use of IQVis will be exemplified by a smart greenhouse scenario, a case study implemented in laboratory settings.

Architecture for end-user programming of cross-smart space applications

This paper presents an architecture and execution components that assists the end-user programming of smart space applications that integrate the information/computing capabilities of different software products via smart spaces. Before end-user programming activities, the software professionals must focus on the more difficult tasks and develop easy-to-apply Driver components for the APIs of existing software systems. Thereafter, end-users are able to create applications from the commands of the Driver components with the help of the provided toolset. The paper introduces a reference implementation for the architecture and tools for the Driver component development and end-user programming of smart space applications.