Data Merge: A data aggregation technique for Wireless Sensor Networks Dimitris Tsitsipis *† , Sofia-Maria Dima † , Angeliki Kritikakou † , Christos Panagiotou † , Stavros Koubias † Abstract The Wireless Sensor Networks (WSNs) have limited power and communication capabilities, combined with the requirement for long network lifetime. To increase it, methods to reduce energy consumption are highly re- quired. To achieve this goal, we study a data aggregation technique without size reduction, i.e. data merge. It is a generic technique, since it is also usable in applications with heterogeneous data and requirements for high accu- racy. This study presents the impact of the data merge technique on WSNs applications executed under various realistic data flow scenarios, traffic loads and wait time intervals. Our results show significant reductions in both packet loss and radio energy consumption. 1. Introduction A wide range of real sensor-based applications is used in WSNs, e.g. environmental monitoring, health monitor- ing and military applications [1]. The term WSN refers to networks of autonomous, resource constraint, distributed sensors, which interact by utilizing wireless communica- tion. The nature of WSNs and their applications require long network lifetime. The information is transmitted to the destination nodes through a multi-hop ad-hoc topol- ogy, introducing overhead. Hence, methods to reduce the traffic load and to average the energy consumption, expanding the lifespan of the WSN, are highly required. This is achieved by employing efficient routing schemes, or by limiting the data transmissions, since the radio mod- ule is the most energy consuming part of a node[2]. This paper mainly addresses the latter case. The main approach to reduce transmissions is data ag- gregation techniques. Data aggregation is gathering and routing information and processing data [3]. The two ap- proaches are data aggregation with and without data size reduction (data merge). The most commonly studied tech- nique is data aggregation with size reduction [4]. These techniques propose aggregation functions to combine re- ceived data, e.g. lossy, lossless, duplicate sensitive and du- plicate insensitive [3]. Strategies to adaptively employ dif- ferent link-local error control mechanisms to increase the overall aggregated information are proposed in [5]. The optimization of data aggregation to decrease the transmit- * Corresponding author: mitsarionas@ece.upatras.gr † Dept. Electrical & Computer Engineering, Univ. of Patras, Greece ted messages in WSNs through the ”network coding” is proposed in [6]. A qualitative comparison of these tech- niques is presented in [7]. However, the data aggregation techniques with size reduction require sensing of homogeneous data, e.g. at nearby clusters [8]. Otherwise, data aggregation often re- quires very complex data fusion techniques, e.g. a WSN which monitors temperature and humidity [9], or they re- quire the reduction of data accuracy to be acceptable at the final destination node. The data merge techniques, i.e. data aggregation without size reduction, merge packets in a single merged packet with a longer data field. They re- duce the overall transmission overhead by transmitting a longer packet. It has less header transmissions without compression of high correlated information. These tech- niques are more general as they are applied also in cases with transmission of heterogeneous data or high data ac- curacy requirements. When data fusion assumptions hold, data merge can be applied on top of the results of the data aggregation with reduction to provide further improve- ment. Few approaches dedicated to data merge technique exist in the current literature. A general definition of the data merge is found in [9]. Other studies also refer to it, but without presenting any evaluation results [3], [8], [10]. A merged packet routing algorithm is presented in [11], where the packet merging nodes are also used as the next hop in the routing path. The technique is used for 8 nodes, under an all-to-all communication. The main contribution of this paper is to implement and thoroughly study the data merge technique and its impact on the applications executed over WSNs under different realistic data flow scenarios and traffic loads. Our goal is to offer in-depth insight to the behavior of the technique and to guide in its efficient use. To support this contribu- tion, the data merge technique is analyzed and applied in several realistic data flow scenarios, under different traffic loads and wait time intervals. To the best of our knowl- edge, the data merge technique is not thoroughly studied in the current literature under several realistic data flow scenarios in WSNs. In this paper, we explore its impact by evaluating several performance measures through sim- ulation results. 2. Data merge technique The data merge technique is applied when overlapping paths exist in the route of the packets towards their desti- nation nodes. Data from different sensor nodes are com- Authorized licensed use limited to: University of Patras. Downloaded on January 17,2023 at 10:09:40 UTC from IEEE Xplore. Restrictions apply.