Ivana Podnar Zarko

Web text retrieval with a P2P query-driven index

In this paper, we present a query-driven indexing/retrieval strategy for efficient full text retrieval from large document collections distributed within a structured P2P network. Our indexing strategy is based on two important properties: (1) the generated distributed index stores posting lists for carefully chosen indexing term combinations, and (2) the posting lists containing too many document references are truncated to a bounded number of their top-ranked elements. These two properties guarantee acceptable storage and bandwidth requirements, essentially because the number of indexing term combinations remains scalable and the transmitted posting lists never exceed a constant size. However, as the number of generated term combinations can still become quite large, we also use term statistics extracted from available query logs to index only such combinations that are frequently present in user queries. Thus, by avoiding the generation of superfluous indexing term combinations, we achieve an additional substantial reduction in bandwidth and storage consumption. As a result, the generated distributed index corresponds to a constantly evolving query-driven indexing structure that efficiently follows current information needs of the users. More precisely, our theoretical analysis and experimental results indicate that, at the price of a marginal loss in retrieval quality for rare queries, the generated index size and network traffic remain manageable even for web-size document collections. Furthermore, our experiments show that at the same time the achieved retrieval quality is fully comparable to the one obtained with a state-of-the-art centralized query engine.

Publish/subscribe middleware for energy-efficient mobile crowdsensing

In this paper we focus on mobile crowdsensing applications for community sensing where sensors and mobile devices jointly collect and share data of interest to observe and measure phenomena over a larger geographic area. Such applications, e.g., environmental monitoring or crowdsourced traffic monitoring, involve numerous individuals that on the one hand continuously contribute sensed data to application servers, and on the other hand consume the information of interest to observe a phenomenon typically in their close vicinity. Energy-efficient and context-aware orchestration of the sensing process with data transmission from sensors through mobile devices into the cloud, as well as from the cloud to mobile devices such that information of interest is served to users in real-time, is essential for such applications, primarily due to battery limitations of both mobile devices and wearable sensors. In addition, the latency of data propagation represents their key quality measure from the users perspective. Publish/subscribe middleware offers the mechanisms to deal with those challenges: It enables selective real-time acquisition and filtering of sensor data on mobile devices, efficient continuous processing of large data volumes within the cloud, and near real-time delivery of notifications to mobile devices. This paper presents our implementation of a publish/subscribe middleware system which is tailored to the requirements of mobile and resource-constrained environments with a goal to reduce the overall energy consumption in such environments, and proposes a general architecture for mobile crowdsensing applications. We demonstrate the usability of both the architecture and middleware through our application for air quality monitoring, and discuss the energy footprint of the proposed solution.

Real time analysis of sensor data for the Internet of Things by means of clustering and event processing

Sensor technology and sensor networks have evolved so rapidly that they are now considered a core driver of the Internet of Things (IoT), however data analytics on IoT streams is still in its infancy. This paper introduces an approach to sensor data analytics by using the OpenIoT1 middleware; real time event processing and clustering algorithms have been used for this purpose. The OpenIoT platform has been extended to support stream processing and thus we demonstrate its flexibility in enabling real time on-demand application domain analytics. We use mobile crowd-sensed data, provided in real time from wearable sensors, to analyse and infer air quality conditions. This experimental evaluation has been implemented using the design principles and methods for IoT data interoperability specified by the OpenIoT project. We describe an event and clustering analytics server that acts as an interface for novel analytical IoT services. The approach presented in this paper also demonstrates how sensor data acquired from mobile devices can be integrated within IoT platforms to enable analytics on data streams. It can be regarded as a valuable tool to understand complex phenomena, e.g., air pollution dynamics and its impact on human health.

Urban crowd sensing demonstrator: Sense the Zagreb Air

We demonstrate an urban crowd sensing application for monitoring air quality by use of specially-designed wearable sensors and mobile phones. The application is built upon the OpenIoT platform1 with the goal to support context-aware and energy-efficient acquisition and filtering of sensor data in mobile environments while ensuring adequate sensing coverage. We demonstrate how sensors and mobile devices jointly collect and share data of interest to measure air quality. In particular, we outline the main features of our wearable air quality sensors, present the data acquisition process as well as the user view of the system, which, in contrast to similar applications, provides a personalized real-time notification mechanism to mobile application users. The solution was used in an air quality measurement campaign “Sense the Zagreb Air” performed in the City of Zagreb, Croatia, in early July 2014 with 20 participants.

IoT data management methods and optimisation algorithms for mobile publish/subscribe services in cloud environments

Recent advances in the Internet of Things (IoT) domain for deploying IoT data systems within the cloud have generated Internet-connected silos of sensor technology which make the collection and processing of sensor-generated information more complex. A viable solution to this problem is the use of local sub-servers acting as collectors hubs between mobile sensing devices and the cloud. Publish/subscribe mechanisms have the capacity to provide full control over the data acquisition and filtering process in mobile IoT environments as well as the methods for implementing optimisations related to energy-efficient data harvesting. The paper formulates design principles for IoT data management methods and optimisation algorithms by means of publish/subscribe middleware and linked data which span over mobile networks and cloud infrastructures to produce a coherent IoT ecosystem. Mobile sensing thus becomes a prominent feature of the OpenIoT platform1, the flaghship Open Source IoT middleware for services based on the automated cloud-based formulation of societies of Internet-connected objects.

Towards the cross-domain interoperability of IoT platforms

The Internet of Things is evolving around a plethora of vertical platforms, each specifically suited to a given scenario and often adopting proprietary communications, device and resource control protocols. The emerging need for cross-domain IoT applications and services highlights the necessity of interoperability across IoT platforms for a unified and secure sharing of and access to sensing/actuating resources. This position paper describes the current state of the IoT landscape, the opportunities that appear towards its sustainable evolution as well as the challenges that need to be addressed. In this context, the vision and objectives of the H2020 symbIoTe project are also presented; symbIoTe aims at the interoperability of IoT platforms by offering a flexible interoperability framework that will allow i) vertical IoT platforms to cooperate, ii) collaborative IoT platforms to form IoT-platform federations for exchange of resources and iii) independent developers to create innovative and cross-domain applications.

Comparison of the CUPUS middleware and MQTT protocol for smart city services

Publish/subscribe messaging pattern is often used as a communication mechanism in data-oriented applications and is becoming wide-spread, especially due to the expansion of the Internet of Things (IoT) services and applications. In addition to MQTT, which is one of the commonly used publish/subscribe protocols in the context of IoT, there are a number of other message queuing solutions, either open or proprietary. We have designed a CloUd-based PUblish/Subscribe (CUPUS) middleware solution within the framework of the FP7 project OpenIoT1 that has developed an open-source cloud platform for the IoT. CUPUS is one of the core OpenIoT components which enables flexible integration of wearable sensors and mobile devices as data sources within the OpenIoT platform. In this paper we compare MQTT and CUPUS in the context of smart city application scenarios. Smart city services pose different key-requirements on IoT publish/subscribe solutions and thus we propose a taxonomy to identify vital features of IoT publish/subscribe middleware. The comparison shows that CUPUS is more appropriate for mobile environments with frequent context changes, while it can filter out unrequired data on devices prior to being reported to backend cloud servers. The MQTT protocol proves to be suitable for Wireless Sensor Networks (WSNs) and heterogeneous environments due to its small code footprint, low bandwidth usage and standardized interfaces. Finally we evaluate the two solutions in terms of message footprint in a real-world scenario, latency and delivery semantics.

Query-driven indexing for peer-to-peer text retrieval

We describe a query-driven indexing framework for scalable text retrieval over structured P2P networks. To cope with the bandwidth consumption problem that has been identified as the major obstacle for full-text retrieval in P2P networks, we truncate posting lists associated with indexing features to a constant size storing only top-k ranked document references. To compensate for the loss of information caused by the truncation, we extend the set of indexing features with carefully chosen term sets. Indexing term sets are selected based on the query statistics extracted from query logs, thus we index only such combinations that are a) frequently present in user queries and b) non-redundant w.r.t the rest of the index. The distributed index is compact and efficient as it constantly evolves adapting to the current query popularity distribution. Moreover, it is possible to control the tradeoff between the storage/bandwidth requirements and the quality of query answering by tuning the indexing parameters. Our theoretical analysis and experimental results indicate that we can indeed achieve scalable P2P text retrieval for very large document collections and deliver good retrieval performance.

Edge Computing Architecture for Mobile Crowdsensing

Mobile crowdsensing (MCS) is a human-driven Internet of Things service empowering citizens to observe the phenomena of individual, community, or even societal value by sharing sensor data about their environment while on the move. Typical MCS service implementations utilize cloud-based centralized architectures, which consume a lot of computational resources and generate significant network traffic, both in mobile networks and toward cloud-based MCS services. Mobile edge computing (MEC) is a natural choice to distribute MCS solutions by moving computation to network edge, since an MEC-based architecture enables significant performance improvements due to the partitioning of problem space based on location, where real-time data processing and aggregation is performed close to data sources. This in turn reduces the associated traffic in mobile core and will facilitate MCS deployments of massive scale. This paper proposes an edge computing architecture adequate for massive scale MCS services by placing key MCS features within the reference MEC architecture. In addition to improved performance, the proposed architecture decreases privacy threats and permits citizens to control the flow of contributed sensor data. It is adequate for both data analytics and real-time MCS scenarios, in line with the 5G vision to integrate a huge number of devices and enable innovative applications requiring low network latency. Our analysis of service overhead introduced by distributed architecture and service reconfiguration at network edge performed on real user traces shows that this overhead is controllable and small compared with the aforementioned benefits. When enhanced by interoperability concepts, the proposed architecture creates an environment for the establishment of an MCS marketplace for bartering and trading of both raw sensor data and aggregated/processed information.

Inferring presence status on smartphones: The big data perspective

In the context of communication services, presence is defined as the willingness and ability of a user to communicate across a set of devices with other users, and thus an up-to-date user presence status represents an essential prerequisite for real-time communications. Smartphones are a rich source of presence-related contex information, however; this information is currently not applied by the prevailing over-the-top communication systems to implicitly change user presence status in accordance with his/her context and typical daily behavior. Smartphone battery limitations and the abundance of context data generated from built-in sensors and mobile applications are the major factors limiting the adoption of rich presence solutions in state-of-the-art communication solutions. This paper presents an approach to learning and inferring user presence status on smartphones using the available context data with a goal to enable non intrusive and energy-efficient maintenance of presence status without user intervention. We apply the Mobile Data Challenge (MDC) data set collected during the Lausanne Data Collection Campaign from October 2009 until March 2011 in our evaluations.