International Journal of Space-Based and Situated Computing (5 papers in press)
Leader Election and Computation of a Spanning Tree in Dynamic Distributed Networks using Local Computations and Mobile Agents
by Mouna Ktari, Mohamed Mosbah, Ahmed Hadj Kacem
Abstract: Leader election problem is among the important problems in distributed computing. The continued evolution of distributed systems keeps the distributed computing an open area of research. Distributed algorithms are hard to design and much harder to prove. To make designing distributed algorithm easier, we model this latter with a local computations model. Distributed algorithms are formally presented by rewriting rules. Beyond a formal presentation of these algorithms, local computations theory proposes not only a correctness proof by the use of invariants but also a termination proof by the use of the graph mathematical tool-box. Based on both, the local computations model and the mobile agent paradigm, we present in this paper a distributed algorithm that elects a leader and computes a spanning tree in a dynamic graph. Computations in dynamic graphs can be affected by a set of topological events: we address the appearance and the disappearance of places and communication channels. Our goal is to always maintain a tree by a single leader or a forest of sub-trees where each one has its own leader.
Keywords: Dynamic Networks; Distributed Algorithms; Local Computations; Mobile Agents; Leader Election; Spanning Tree.
Acceleration of the K-Means algorithm by removing stable items
by Adriana Mexicano, Ricardo Rodriguez Jorge
Abstract: This work presents an approach for enhancing the K-Means algorithm in the classification phase. The approach consists in a heuristic which at each time that an object remains in the same group, between the current and the previous iteration, it is identified as stable and it is removed from computations in the classification phase in the current and subsequent iterations. This approach helps to reduce the execution time of the standard version. It can be useful in Big data applications. For evaluating computational results both, the standard and the proposal were implemented and executed using three synthetic and seven well-known real instances. After test both versions, it was possible to validate that the proposed approach spend less time than the standard one. The best result was obtained for the Transactions instance when it was grouped into 200 clusters, achieving a time reduction of 90.1% with a reduction in quality of 3.97%, regarding the standard version.
Keywords: K-Means; time reduction; accelerating the classification phase.
Classification Extension based on IoT-Big Data Analytic for Smart Environment Monitoring and Analytic in Real-time System
by Riyadh Ar-Ridha, Sritrusta Sukaridhoto, Dadet Pramadihanto, Nobuo Funabiki
Abstract: Monitoring water conditions in real-time is a critical mission to preserve the water ecosystem in maritime and archipelagic countries, such as Indonesia that is relying on the wealth of water resources. To integrate the water monitoring system into the big data technology for real-time analysis, we have engaged in the ongoing project named SEMAR (Smart Environment Monitoring and Analytic in Real-time system), which provides the IoT-Big Data platform for water monitoring. However, SEMAR does not have an analytical system yet. This paper proposes the analytical system for water quality classification using Pollution Index method, which is an extension of SEMAR. Besides, the communication protocol is updated from REST to MQTT. Furthermore, the real-time user interface is implemented for visualisation. The evaluations confirmed that the data analytic function adopting the linear SVM and Decision Tree algorithms achieves more than 90% for the estimation accuracy with 0.019075 for the MSE.
Keywords: SEMAR; water condition monitoring; real-time analysis; IoT; big data; classification; machine learning.
Proactive vs. Reactive Multipath Routing for Mobility Enabled RPL
by Ines EL KORBI
Abstract: The IPv6 Routing Protocol for Low power and Lossy Networks (RPL) was proposed as a routing standard for wireless sensor networks (WSNs). For instance to cope with sensor nodes' limitations, RPL reduces the overhead generated by the frequent routing updates in a way to maximize the network lifetime and to reduce the routing related tasks within sensors. However, the RPL performance drastically degrade in the case of nodes' mobility. Therefore, we propose in this paper multipath routing schemes to adapt RPL to the mobility scenarios. For instance, with multipath routing, the data is duplicated on alternative paths in the case of primary path failure. The first Mobility based Braided Multipath RPL (MBM-RPL) scheme establishes an alternative path braided with the primary one to replace the loose link (the up to fail link) along the first network route. The second Proactive Multipath Mobility based RPL (PMM-RPL) approach establishes all node disjoint routes between each node in the network and the rink. Data would not be duplicated on secondary paths until a primary path vulnerability detection. Both MBM-RPL and PMM-RPL performance are evaluated in terms of packet loss rate (PLR), average transmission delay (ADT) and the generated network overhead using the Cooja/Contiki simulation environment.
Keywords: WSN; RPL; mobility; multipath; MBM-RPL; PMM-RPL.
Efficient Listen Before Technique for LTE-WiFi Co-existence in Unlicensed Bands
by Visali Mushunuri, Bighnaraj Panigrahi, Hemant Kumar Rath, Anantha Simha
Abstract: Licensed Assisted Access (LAA) has been proposed by 3rd Generation Partnership Project (3GPP) in order to use Long Term Evolution (LTE) in the un-licensed spectrum. As the centralized control of LAA has to co-exist with existing distributed Wireless Fidelity (WiFi), efficient Listen Before Talk (LBT) technique is required to reduce interference and improve spectral efficiency of WiFi as well as LAA users. In this paper, we have proposed an LBT mechanism which attempts to share the medium in a fair manner and improves the WiFi as well as WiFi-LAA combined system performances significantly. We have also proposed an analytical closed form solution for the LBT technique. We have observed network performance of the proposed LBT model using Matlab based numerical results and Network Simulator (NS-3) based simulation results. The results indicated that with proper choice of LAA channel occupancy and backoff counter, significant WiFi as well as overall system gains can be achieved.
Keywords: LTE; WiFi; LAA; LTE unlicensed; Listen Before Talk.