LNR-PP: Leaf Node Count and RSSI Based Parent Prediction Scheme to Support QoS in Presence of Mobility in 6LoWPAN

Abstract

6LoWPAN network consists of embedded devices and sensors that senses environmental data and passes it to a root called as a 6LBR (6LoWPAN Border Router) which further transfers it to an Internet server for aggregation and processing. Earlier definition of 6LoWPAN consists of devices which are static with constrained resources such as computing power, memory, battery power and communication bandwidth. Recent additions such as mobile devices which are part of the 6LoWPAN network often cause frequent topology changes and disconnections, leading to degraded performance of the network. Handover delay, data loss rate, number of control messages, energy utilization are some of the performance metrics that has been focused and improved by recent works in the literature. But the proposed approaches work well when the number of mobile nodes and mobility rate are moderate. We identify the issues that arise in the 6LoWPAN RPL based network due to higher number of mobile nodes and higher mobility rate which causes problems like delay in data transmission and degraded parent node performance. In this paper we study, analyse and propose modifications in the popular protocols that support mobility in 6LoWPAN such as EC-MRPL. We propose a best-effort scheme called LNR-PP (Leaf node count and RSSI based Parent Prediction) to identify affected Static Nodes (SN) to whom large number of Mobile Nodes (MN) are connected. We call this scheme as a best effort since the parent prediction depends upon the number of neighbours available within the reach of MN in mobility. LNR-PP performs balanced leaf node (Mobile Node) allocation to SN in an effort to minimize the number of RPL control message packets, delay in forwarding data from the MN and power required to transmit control and data packets through SN. This enables the mobile nodes (MN) to communicate with DODAG-ROOT with guaranteed QoS measures under high mobility scenario. We developed a mathematical model to realize the benefits of the proposed scheme which can be incorporated into EC-MRPL and simulated with sample network parameters in MATLAB. The results obtained through simulation shows a significant improvement in term of number of RPL control message packets such as DIS and DIO, power requirement to transfer control and data packets through SN and delay encountered by SN while forwarding large volume of data from MN.

Introduction

IoT (Internet of Things) has evolved to a greater extent in the recent time. Deployment of applications targeted for IoT have grown up largely with the support of IPV6. 6LoWPAN provides integrity between devices running in (LLN) low power and lossy networks and the Internet. 6LoWPAN provides flexibility, scalability and end to end connectivity making the deployment of IoT easier [1], [2], [3], [4]. Mobility of devices in IoT leverages several applications adding to its growth. At the same time mobility introduces problem such as interruption of connectivity and disconnection of mobile nodes. Loss of data and increased delay in transmission are the problems caused by frequent disconnection of mobile nodes. Existing approaches to handle mobility finds an alternate node (Preferred Node) for the mobile node within a limited time to retain the availability and to avoid disconnection. The limited resources of mobile devices and their mobility throw a challenge to design an efficient protocol to support mobility. Several schemes exist in literature to handle mobility. But they do not consider the energy consumption. Moreover increased handover delay and data loss are the associated problems with the existing approaches. MRPL [5] and EC-MRPL [6] are the well-known protocols to handle mobility in 6LoWPAN.

It has been shown in the literature that EC-MRPL [6] has outperformed RPL and MRPL in terms of energy consumption, hand over time, data delivery rate and number of control messages. As in [6], EC-MRPL assigns a static node, called as Associated Node (AN) for a mobile node to handle mobility. During handover period, the associated node assess the signal strength of the mobile node using RSSI (Received signal strength Indicator) to find node movement. The strength of the received signal depends upon the distance between the mobile node and parent node. When the mobile node is far from the parent node RSS degrades. The value of RSS and setting threshold for RSS is discussed elaborately in Section 3. On exceeding a predefined threshold, initiates a preferred parent (PP) prediction phase to associate the mobile node to a new parent. The mobile node continues to transfer the data with newly attached parent. The proposed work achieves improved node connectivity of mobile nodes with reduced energy consumption. It also improves the handover delay, data loss rate and number of control messages used while supporting mobility.

But the proposed approach may encounter an additional overhead due to more number of mobile nodes connected to a particular parent. Moreover high mobility rate of nodes may invoke several parent prediction phase by an associated node which would slower the data processing and forwarding capacity of the parent to a large extent. This may cause the data transmission across the associated node a bit slower causing an additional delay. Since the parent node buffers the data transmitted by the mobile node during handover period, the buffer of the parent node becomes overwhelmed with the data transferred by mobile nodes. The above identified issues occurs more in scenarios like VANET, where a moving vehicle changes its position and their parent node due to higher speed. Moreover the above identified issues can be observed in networks formed in commercial buildings, hospitals and conference halls with larger number of participating mobile nodes. This is because the frequent movement of participating devices or users within the network often causes change in network topology.

We propose LNR-PP which assess the number of mobile nodes attached to a parent at a particular point of time and balances the leaf node (mobile node) on all parent so that delay in transmitting the data to root can be reduced and the buffer overload can also be reduced during larger number of mobile nodes under higher mobility rate. We provide an analytical model to realize the benefits of the proposed scheme. The paper is organized as follows: Section 2 presents various schemes that exist in literature to avoid congestion, reduce energy consumption and handle mobility in 6LoWPAN. Section 3 briefly presents the existing approach and associated issues. Section 4 discusses in detail the proposed LNR-PP. Section 5 discusses the mathematical modelling and simulation. Section 6 analyses the results of simulation.