IoT Data and Context Discovery
Guest Editors’ Introduction • Arkady Zaslavsky and Prem Jayaraman • September 2016
Translations by Osvaldo Perez and Tiejun Huang
Listen to the Guest Editors' Introduction
Audio by Martin Omana, Timothy K. Shih, and Steve Woods
One of the most valuable aspects of the emerging Internet of Things (IoT) is the data it produces. Businesses use that data to support their decisions, and — as IoT grows — they need better tools for relevant and timely discovery. These days, discovery systems can find the right data without even knowing its structure, semantics, sensor description, or location. These systems can also deduce context information such as annotations and metadata.
The term “IoT data and context discovery” refers to both activities that are specific to data providers (certain prepublication curation tasks, for example) and those that are specific to end publishers or brokers (such as integrating datasets to support data linking and context-driven search). The discovery process comprises two successive loops:
- The foraging loop identifies, assesses, and validates data sources at the point of data acquisition, as well as extracts and formats the relevant data into a consumable form.
- The sense-making loop processes, analyzes, and exploits the extracted data to generate relevant context, aiming to provide answers and insights.
The advent of IoT has fuelled a paradigm shift in data and context discovery. Datasets that were once confined to single applications are now discoverable and available for reuse and repurposing in multiple applications. This new paradigm provides vendors with incentive-based approaches to opening their IoT data repositories while still upholding their security and privacy policies. Despite this progress, the diversity of capabilities and standards among devices poses significant challenges. Computing Now’s September 2016 issue includes seven articles that examine opportunities and challenges in IoT data and context discovery.
A Common Consortium
Currently, different devices store data in separate “silos.” For example, Fitbit devices collect personal health data, and EarlySense devices monitor patient vital signs. Both companies produce zettabytes of data, but each keeps its data on its own servers. To fulfill the ambitious dream of a truly interconnected IoT, data would have to be stored in widely distributed, heterogeneous databases to ensure global availability. Retrieving the data would require a common, machine-readable data-representation framework.
Figure 1 depicts the guest editors’ vision of a common consortium of IoT service providers. In Figure 1a, the current vendor-specific approach creates IoT data silos by tightly coupling applications with specific sensors. In Figure 1b, a discovery-enabled system loosely couples applications and sensors to allow for interoperability and IoT data reuse and repurposing.
Figure 1. Vision of IoT discovery. a) Vendor-specific IoT approaches currently create data silos. b) Our vision of a discovery-enabled system would leverage loose coupling between applications and sensors to enable interoperability, reuse, and repurposing of IoT data and context.
A discovery engine, supported by a context engine and an integration engine, would discover underlying IoT data sources that multiple vendors host and manage. The system would require common, well-described interfaces that utilize Internet standards such as the semantic web. The evolution of IoT and big data requires support for new data- and sensor-discovery techniques to overcome issues that prevent seamless data access and reuse.
In this Issue
This month’s Computing Now theme begins with “Physical-Cyber-Social Computing: Looking Back, Looking Forward,” by Payam Barnaghi and his colleagues. The authors introduce physical-cyber-social computing, explaining how systems interpret users’ social structures through IoT data and provide (near) real-time actionable information and services.
Dimitrios Georgakopoulos and his colleagues provide a vision of a future IoT system architecture driven by service discovery and real-time service integration in “Discovery-Driven Service Oriented IoT Architecture.” This vision includes on-demand discovery, IoT device integration, cloud storage, and computing resources.
In “Semantic Description, Discovery and Integration for the Internet of Things,” Sejin Chun and his colleagues provide a semantic, model-based IoT directory system to help manage metadata and relationships among devices. The proposed model enables shared conceptualization for efficient interaction between the devices and the online directory service.
Security is an important concern in IoT. In “Context-Sensitive Policy Based Security in Internet of Things,” Prajit Kumar Das and his colleagues propose a framework that lets IoT devices capture, represent, and enforce information-sharing policies. The authors use semantic web concepts to enable consistent policy representation and present use cases to demonstrate their design.
“Matching Over Linked Data Streams in the Internet of Things,” by Yongrui Qin, Quan Z. Sheng, and Edward Curry, explores techniques for efficient dissemination of IoT data to consumers. The authors use linked open data to represent IoT data and its relationships, and then disseminate matched data based on system-registered queries. The paper presents a use case to validate the system’s applicability, speed, and efficiency.
In “Resource Discovery in Internet of Things: Current Trends and Future Standardization Aspects,” Soumya Kanti Datta, Rui Pedro Ferreira Da Costa, and Christian Bonnet discuss the current technology landscape for discovery in IoT, including advantages and limitations. They propose a centralized registry for storing resource configurations and parameters, as well as a search engine that ranks available resources and provides URIs for direct access to resources.
The final article, “The Web of Things: Challenges and Opportunities,” presents concerns arising from the increasing use of virtual representations for physical or abstract things that are accessible via web technologies. Author Dave Raggett argues that achieving a new phase of exponential growth will require open markets, open standards, and the vision to imagine the potential for this expanding network.
Rodolfo Milito of Cisco offers insights into the intersection of distributed computing and IoT.
Manfred Hauswirth of Fraunhofer FOKUS discusses challenges and opportunities in IoT data and context discovery.
This month’s Computing Now issue includes two video interviews with prominent industry experts on the importance and challenges of data and context discovery in IoT middleware, services, and applications. The first features Rodolfo Milito from Cisco in San Jose, California, who is reputed to be the father of fog computing. The second provides insight from Manfred Hauswirth, the director of Fraunhofer Institute for Open Communication Systems (FOKUS) in Berlin.
The articles in this Computing Now issue address various challenges that IoT data and context discovery raises, as well as propose solutions to make IoT practical, feasible, deployable, and usable. Of course, more research is necessary before IoT data and context discovery becomes a common feature of IoT applications, systems, and services. We encourage interested readers to dive into the research and join the large community of IoT champions, developers, architects, and users.
Arkady Zaslavsky is senior principal research scientist at Data61 of the Commonwealth Scientific and Industrial Research Organisation (CSIRO), Australia. He is also an adjunct professor at the University of New South Wales, La Trobe University, and ITMO University. Zaslavsky has a PhD in computer science from the Institute of Control Sciences, USSR Academy of Sciences. His technical interests include the Internet of Things; pervasive, ubiquitous, and mobile computing; context-awareness; semantic data management; and mobile analytics. He is a member of the Computing Now editorial board. Contact him at firstname.lastname@example.org.
Prem Prakash Jayaraman is a research fellow at the Swinburne University of Technology. His research interests include the Internet of Things, cloud computing, big-data analytics, and mobile computing. Jayaraman has a PhD in computer science from Monash University. Contact him at email@example.com or www.premjayaraman.com.