Christine Jardak

Parallel processing of data from very large-scale wireless sensor networks

In this paper we explore the problems of storing and reasoning about data collected from very large-scale wireless sensor networks (WSNs). Potential worldwide deployment of WSNs for, e.g., environmental monitoring purposes could yield data in amounts of petabytes each year. Distributed database solutions such as BigTable and Hadoop are capable of dealing with storage of such amounts of data. However, it is far from clear whether the associated MapReduce programming model is suitable for processing of sensor data. This is because typical applications MapReduce is used for, currently are relational in nature, whereas for sensing data one is usually interested in spatial structure of data instead. We show that MapReduce can indeed be used to develop such applications, and also describe in detail a general architecture for service platform for storing and processing of data obtained from massive WSNs.

Implementation and Performance Evaluation of nanoIP Protocols: Simplified Versions of TCP, UDP,HTTP and SLP for Wireless Sensor Networks

We present the implementation architecture and performance evaluation of the nanoIP protocol stack. The stack consists of miniaturized versions of UDP, TCP, SLP and HTTP protocols with reduced header sizes and complexity to make the protocols usable in wireless sensor networks. Similarity to the TCP/IP stack facilitates the development of gateways towards IP-based networks and makes the use of the stack easier for developers accustomed to network programming. Our implementation work and experiments show that the footprint of the stack is acceptable even for the most resource constrained sensor nodes.

Design of large-scale agricultural wireless sensor networks: email from the vineyard

We describe the design and implementation of a large-scale Wireless Sensor Network (WSN) for agriculture monitoring. As a part of validation we have deployed a prototype of 64 sensors to monitor a commercial vineyard. The system provides software modules ranging from filtering raw data to a centralised and a distributed data storage applications. The used protocols ensure reliable and robust communication and load-balancing energy consumption. A backend server provides a user-friendly graphical interface offering two main functionalities: logging communication messages of the protocols and providing end-user support for on-demand and periodic data requests. We have made a performance evaluation of a scaled-down WSN. We highlight the efficiency of the system on a large-scale deployment in comparison to the traditional solution based on individual weather station by deploying both systems in parallel and comparing results. We demonstrate that WSNs provide better geographical coverage and an increased spatial resolution compared to traditional solutions.

Distributed Information Storage and Collection for WSNs

Distributed data storage is an important component of wireless sensor networks, which protects the mission critical information from unexpected node failures or malicious destruction of parts of the network. In this paper we present DISC, a protocol for distributed information storage and collection. The two major mechanisms in DISC which make our solution distinct from the related approaches are probabilistic choice of storing nodes and a search engine based on the usage of Bloom filters. In comparison to the deterministic choice of the backup node, the random selection strategy makes it virtually impossible for an attacker to determine and destroy the exact node keeping a particular piece of information. The usage of Bloom filters in the information search engine makes the navigation to a specific data fast and efficient. We show that with DISC the amount of recovered information is more than two times higher than that in deterministic storage schemes.

Email from the vineyard

We describe the design of a large-scale wireless sensor network for agriculture monitoring. We have deployed a prototype of 64 sensors to monitor a commercial vineyard. The system provides different software modules ranging from filtering raw data to a centralized and a distributed data storage applications. The used protocols ensure reliable and robust communication and load-balancing energy consumption. A backend server provides a user-friendly graphical interface offering two main functionalities: logging communication messages of the employed protocols for diagnostic purposes and providing end-user support for both on-demand and periodic data requests. We highlight the efficiency of such a network in comparison to the traditionally used solution based on individual weather station by deploying both systems in parallel and comparing their reading results. The results clearly demonstrate that wireless sensor networks provide better geographic coverage and an increased spatial resolution compared to traditional solutions. They also enable precise measurements of soil conditions near the plants themselves, increasing the quality of the information available to the farmer.

Analyzing the Optimal Use of Bloom Filters in Wireless Sensor Networks Storing Replicas

We study the problem of optimal parameter selection for Bloom filters in content-based routing. In constrained environments, such as sensor network, optimizing lengths of the filters used and the number of hash functions employed can lead to significant reduction of memory overhead and false positives. We demonstrate by analytical calculations and extensive simulations that the commonly used heuristics for choosing these parameters are suboptimal especially in networks storing replicas of the data for redundancy. We show specifically how network topology and use of replication can be taken into account when selecting the parameter values used. Our results show that memory consumption can be reduced by up to 50% compared to usual approaches, and that the occurrence of false positives can be significantly reduced as well.

Flexible Hardware/Software Platform for Tracking Applications

This paper describes the design, implementation and performance evaluation of a wireless sensor network based scalable outdoor vehicular tracking system. The system is highly flexible and configurable both from software and hardware architecture point of views and enables it to adapt to a wide range of vehicle tracking applications. We also present some intermediate results and the rationale behind our design approach. The system was tested for a network of 100 nodes and is scalable to a few thousand node setup. We believe that the vehicle localization and tracking results from our large scale deployment of sensor nodes and system design experience are useful to the community

Flexible hardware/software platform for tracking applications

This paper describes the design, implementation and performance evaluation of a wireless sensor network based scalable outdoor vehicular tracking system. The system is highly flexible and configurable both from software and hardware architecture point of views and enables it to adapt to a wide range of vehicle tracking applications. We also present some intermediate results and the rationale behind our design approach. The system was tested for a network of 100 nodes and is scalable to a few thousand node setup. We believe that the vehicle localization and tracking results from our large scale deployment of sensor nodes and system design experience are useful to the community

Content-based routing using subpart orthogonal codes

We study the use of orthogonal codes as alternative to Bloom filters in content-based routing. In particular we explore the trade-off between centralized solutions using regular orthogonal codes and generating codes locally by hashing. We demonstrate that introduction of subpart orthogonal codes allows distributed operation while leading to shorter key lengths than Bloom filters with the same probability of data localization. We study the performance of sub-part orthogonal codes by both analytical means and in a simulation environment. Results show that the use of sub-part orthogonal codes reduces overhead without sacrificing performance in both tree and grid topologies.