sparql2sql

---

The Internet of Things is a large and constantly growing source of time-series sensor data. Storing Internet of Things data as Linked Data in triple stores and using SPARQL to query them increases interoperability through the use and referencing of common identifiers and ontologies. Furthermore, the self-describing aspect of Linked Data is a boon for integration. However, storing sensor data in triple stores is often inefficient in terms of storage and query performance which is further exacerbated by the resource constraints of Lightweight Computers - compact & mobile machines that can be deployed in close proximity to sensors.

We propose an approach of storing time-series data, produced by sensors from the Internet of Things, more succinctly in relational tables that use a simple, bottom-up mapping approach to integrate and expose the data for SPARQL querying. We show that in comparison to storing the data in triple stores, our approach provides gains in both storage efficiency and query processing performance from over 3 times to over three orders of magnitude on Lightweight Computers. A range of analytical queries from a Linked Sensor Data benchmark, SRBench, on meteorological sensor data and analytical queries focusing on space-time-aggregations on Smart Home sensor data are tested.

The contribution of this research is to enable applications that utilise analytical queries to be built upon a performant framework that does not sacrifice interoperability.

Siow, E., Tiropanis, T., Hall, W.: Efficient SPARQL Queries for Analytics on Lightweight Computers for the Internet of Things (2015)

The PiSmartHome is an example of a working visualisation system that uses SPARQL queries translated using the sparql2sql engine for execution on Lightweight Computer hardware.

• Source: github.com/eugenesiow/sparql2sql
• Queries: github.com/eugenesiow/sparql2sql/wiki
• PiSmartHome: Source

Close

sparql2stream

---

The distributed and heterogeneous nature of the Internet of Things (IoT) drives the need for technologies that aggregate & integrate the data produced by physical entities like sensors. Linked Data, which semantically interconnects structured data, can be used on lightweight and mobile computers, deployed close to the source, to integrate data from sensors and devices with the added benefit of fine-grained control over data privacy. However, we show the performance limitations of a variety of systems on Lightweight Computers in inserting and querying Linked Data. We then design and demonstrate a solution that utilises stream processing, with an efficient row-based representation of time-series IoT data, and a translation engine to allow querying as SPARQL. We show this improves insertion and query performance by more than an order of magnitude, providing a case for Lightweight Computers with Linked Data for IoT data integration.

Siow, E., Tiropanis, T., Hall, W.: Linked Data on Lightweight Computers for the Internet of Things (2015)

See the "IoT Freeboard" demonstration in the Projects section of the page to see the how a dashboard for a Smart Home can run with a powerful stream processing engine and websockets.

• Source: github.com/eugenesiow/sparql2stream
• Queries: github.com/eugenesiow/sparql2stream/wiki

Close

IoT Freeboard

---

IoT Freeboard is a highly cutomisable dashboard (supporting many widgets and datasources) for the IoT that is based on Freeboard but uses the powerful sparql2stream engine to pick out complex events with streaming SPARQL queries. The dashboard is updated via push events from the streams and uses websockets for 2 way communication.

This demo shows the IoT Freeboard in action replaying the published Smart* dataset at a 100x increased rate (simulating high velocity). Streams include those from the environmental sensors, motion sensors and meter sensors. Several streaming queries are registered and the events are pushed to an events stream, fed into a websocket and added as the 'Events' datasource in the dashboard. In the demo the customisation options of the dashboard are also shown.

Barker, S,, Mishra, A., Irwin, D. et al.: Smart*: An Open Data Set and Tools for Enabling Research in Sustainable Homes, Workshop on Data Mining Applications in Sustainability (2012).

• Source: github.com/eugenesiow/iotwo

Close

A Survey on IoT Analytics

---

The Internet of Things (IoT) envisions a world-wide, interconnected network of smart physical entities. These physical entities generate a large amount of data in operation and as the IoT gains momentum in terms of deployment, the combined scale of those data seems destined to continue to grow. Increasingly, applications on the IoT involve analytics. Data analytics is the process of deriving knowledge from data, generating value like actionable insights from them. This article reviews work in the IoT and analytics from the perspective of their utility in creating efficient, effective and innovative applications and services for a wide spectrum of domains. We review the broad vision for the IoT as it is shaped in various communities, examine the application of data analytics by communities and provide a categorisation of analytic approaches in different types of settings in the IoT domain. This categorisation provides us with insights on the appropriateness of enabling infrastructures for IoT analytics. Finally, we examine the research challenges and opportunities going forward from the perspective of different stakeholders in the IoT.

Siow, E., Tiropanis, T., Hall, W.: Analytics for the Internet of Things: A Survey (2015)

Close

Lightweight Computers in Disaster Management

---

Siow, E.: Rapid Response Linked Data, International Workshop on Social Machines for Diasaster Management (2015)

Close

IoT Observatories

---

An IoT Observatory System consists of an IoT Observatory that can be on a cluster or a computer/server, Observatory Nodes that can be hosted on computers or lightweight computers and sensors & devices that collect and harvest data from the environment.

The Web Observatory Project is developing a global, distributed infrastructure that allows the use and exchange of datasets, analytical apps and visualisations. The communities that it is fostering, through the network of Web Observatories that has emerged, are able to observe the web and explore its processes while developing an understanding of its impact on varying sectors of human activity. The foundation of the web of observatories that form the Web Observatory Project, are web observatory instances that exist as collections of datasets and analytical tools protected by access controls. The principles espoused by the Web Observatory as a distributed infrastructure are a possible solution for managing datasets and apps in the IoT, however, differences in IoT data and applications necessitate we differentiate IoT Observatories from Web Observatories.

'Things' in the IoT, produce data that is markedly different and unique from Web data. Whether you consider the perspectives of machine-to-machine (M2M) systems or smart objects that interact with humans and other objects, time-series sensor data are a common output from such devices. Unlike web data which ranges from documents and metadata to semantic and social media data, it can be argued that IoT data is characterised more by its volume and velocity than its variety (it is usually well-structured).

From various surveys that show the potential range of unique application domains that IoT applications can contribute to, there is a fundamental shift from human-human communication through the web to multi-partite communications that include machines in the picture, differentiating it from the Web when it comes to building applications.

The unique data, applications and challenges of the IoT discussed above provide motivation for building personal IoT datastores that are interoperable but also efficient at storing time-series sensor data and processing analytical queries. These datastores can then become nodes within distributed infrastructure called IoT Observatories that draw inspiration from Web Observatories and seek to be a means of engaging IoT communities to co-create, manage and share IoT resources (datasets, apps and infrastructure) while respecting privacy and access control.

Close