EMS: Enhanced Mobility Scheme for Controlled and Lossy Networks

Transcription

1 EMS: Enhanced Mobility Scheme for Controlled and Lossy Networks Riaz A Khan and Ajaz Hussain Mir Department of Electronics and Communication Engineering National Institute of Technology Srinagar, India riazk3@gmail.com Abstract with the emergence of IOT, WSNs have become prevalent for many application areas such as smart city scenario, healthcare, weather monitoring and intelligent transport system etc. Due to their smaller infrastructure and limited energy resources, these networks are considered to be controlled and lossy networks. IP mobility management in such networks playing a crucial role for object tracking or to provide uninterrupted services to mobile users. Existing MIPv6 and its extensions are not suitable for such lossy networks due to the MN s involvement in heavy signaling, required for HO. Network based protocol; PMIPv6 relieves the MN from heavy signaling overhead and being suggested by many researchers for lossy networks to carry out the mobility. But this protocol still suffers from extra signaling and packet loss and there is a scope for further improvements. This paper presents a novel approach based on PMIPv6 for reducing HO latency and packet loss for some critical applications such as Healthcare, battle field, disaster management, ITS etc. where packet loss is not acceptable. The performance analysis and simulation results show that the proposed scheme is better than the standard PMIPv6. Keywords Lossy networks, IP moility, MIPv6 and PMIPv6. I. INTRODUCTION IOT [1]; also known as future internet comprised of interconnected devices, has flooded the world with mobile devices. These interconnected devices are able to sense, process and communicate the information over the internet. Because of limited energy resources of these sensing devices and smaller infrastructure, their network is known as controlled and lossy network. The connectivity of such networks with the general IPv6 IEEE networks was the main issue for the research community. IETF working group for 6LoWPAN [2] designed header compression and fragmentation for IPv6 over IEEE to allow IPv6 to run over the IEEE link layer [3]. In controlled and lossy networks, most of the objects are mobile and there is need for seamless mobility for monitoring the mobile objects (living or non-living). The mobile objects are equipped with sensor nodes enabled with IPv6 to communicate over internet. Introduction of host based mobility protocols such as MIPv6 were proved to be milestone but it had limitations of greater HO delay and packet loss. Also the excessive signaling cost due to again and again registration during HO could consume the wireless link [4]. HMIPv6 and FHMIPv6 were further introduced as extension to MIPv6 [5, 6]. Although these extensions were proved efficient for reducing HO latency but could not address the other issues such as heavy signaling and host involvement for the same. Also these protocols could not impose control over the network and which was not acceptable for lossy WSNs. Host involvement in mobility signaling and limited energy resources of such networks has made host based mobility protocols not fit for these networks. These limitations prompted the development of NETLMM protocol where MN is relieved from heavy signaling. PMIPv6 is the only standardized network based mobility protocol by IETF [7]. PMIPv6 can be used in mobility support for IEEE networks or 6LoWPAN WSN. With the development in state of art, many improvements were shown for PMIPv6 in the recent years. These improvements were mainly focused on reducing HO latency and signaling cost and no attention has been paid for packet loss. In lossy networks, sensors involvement in mobility related signaling and packet loss are the main issues which are not acceptable for some critical applications like healthcare, disaster management, battle field and ITS etc. Heavy signaling burden will consume sensor resources immediately and packet loss will result in loss of critical information. Due to the time criticality of data in such mentioned fields, it is very essential to reduce this critical information loss. To overcome these shortcomings, we proposed an enhanced mobility scheme (EMS); an improvement over PMIPv6. The rest of this paper is organized as follows: Section II describes the controlled and lossy networks. Section III reviews the related work carried out for mobility management in such networks. Existing schemes available in the field are discussed in section IV along with their theoretical analysis. Section V presents the proposed scheme along with its theoretical performance analysis. Simulation of studied network model for proposed scheme is presented in section VI with result discussion. Finally section VII concludes the paper and tells about scope of future work in the field. Table 1 in section II contains the nomenclature used in this paper. II. CONTROLLED AND LOSSY NETWORKS For nearly a decade, with the emergence of IOT, research in WSNs has assumed that the current internet architecture is not suitable for wireless sensor applications. Being composed of resource constrained devices, researchers considered such /15/$31.00 c 2015 IEEE 655

2 networks as lossy networks. Also such networks are having smaller infrastructure and each node (whether static or mobile) is well aware of it, so sometimes these are also referred as controlled networks. Many researchers in the field argued that Internet protocols were impractical for the resource constrained devices. Robustness and scalability required for network processing is not possible in such networks. As a result, research efforts in the field introduced evaluated new protocols without regard to an existing architecture allowing experiencing sensor applications. In the recent couple of years, the new developing technologies such as IOT, green computing and smart grid etc. present a need to deploy networks that allow multivendor interoperability, and utilize low-cost communication devices. Sensor networks are considered as best fit for such applications. To address this need and make such lossy networks compatible with the present architecture, some groups under IETF were established [13, 14]. A. 6LoWPAN: The 6LoWPAN working Group was created to standardize necessary adaptations of IPv6 for networks that use IEEE physical (PHY) layer, and has defined how to carry IPv6 datagrams over IEEE links and perform necessary configuration functions to form and maintain an IPv6 subnet [15]. The 6LoWPAN Working Group has focused on two work items: How to carry IPv6 datagrams in frames. How to perform necessary IPv6 neighbor discovery functions (e.g., address resolution, duplicate address detection) in a network with overlapping broadcast domains. 1) Fragmentation: IPv6 specifies that the link must support 1280 bytes as MTU. Therefore for transmission of IPv6 packets over IEEE , 6LoWPAN requires link-layer fragmentation and reassembly mechanism. Fragmentation only provides the ability to encode a datagram using multiple link frames. No recovery of lost fragments is provided, as it is anticipated that link-layer acknowledgments will provide sufficient delivery success rates. The included datagram size in all fragments helps a node to allocate an appropriate reassembly buffer even if fragments are delivered out of order. 2) Header Compression: Generally the header compression techniques are flow based for point to point links. In sensor networks, the flow path can change due to the time varying characteristics of these constrained devices. Therefore a new compression format independent of flow technique was developed [17]. B. Neighbor Discovery (ND) and Address Auto configuration: Like in IPv6 ND (RFC 4861), 6LoWPAN ND also uses RA and RS messages to allow nodes to discover neighboring APs. RS messages are sent using multicast, but this is the only use of multicast within 6LoWPAN. When on-link determination and address resolution proceeds, 6LoWPAN nodes do not require to cache or store packets per neighbor state. A packet is either sent directly to the destination (if it is link-local) or otherwise sent to a router. Address auto detection relieves from the problems of address unreachability and DAD check. 6LoWPAN ND nodes must treat the neighbor table as a registry as IPv6 ND takes the neighbor table as cache. Routers can optionally perform duplicate address detection by forwarding address registration requests to a central server to verify whether or not the address is already in use [13]. Table 1: Nomenclature used in this paper Notation Description IOT Internet of things WSN Wireless Sensor Network MIPv6 Mobile Internet Protocol version 6 HMIPv6 Hierarchical Mobile IPv6 FHMIPv6 Fast Handovers for MIPv6 PMIPv6 Proxy Mobile IPv6 MN Mobile Node MSN Mobile Sensor Node HO Handover IETF Internet Engineering Task Force 6LoWPAN IPv6 over Low Power Personal Area Network RFID Radio Frequency Identifier NETLMM Network Based Localized Mobility Management ITS Intelligent Transport System MTU Maximum Transmission Unit RA Router Advertisement RS Router Solicitation AP Access Point DAD Duplicate Address Detection LMA Local Mobility Anchor/Agent MAG Mobile Access Gateway PBU Proxy Binding Update BCE Binding Cache Entry HNP Home Network Prefix PLR Packet Loss Ratio PDR Packet Delivery Ratio WLD Wireless Link Delay T L2 Time for which layer 2 link goes down T (HO-init+HI) Time for HO initiate T HAck Time for HO acknowledgement T RS Time to send router solicitation request message T MN-ID Time to obtain MNs identifiers and profile T PBU Time to send PBU T HNP Time to allocate HNP to MN T PBA Time to send proxy binding acknowledgement T RA Time to send router advertisement T ADD Time taken for address configuration T NMAG-MN Time for MN receives first packet from NMAG Data traffic rate (packets per time) P L Data packet length III. RELATED WORK Mobility management schemes are there since 2000 for IPv4. With the introduction of IPv6 and IOT, these schemes were also implemented for IPv6 and WSNs for building IOT. Force et al. [8] stated MIPv6 as basic mobility management protocol, which is developed to provide uninterrupted connectivity to MNs while handover from one wireless network to another. MIPv6 for lossy networks is not widely deployed because it requires a mobility stack to be installed on MN which results in tremendous burden on MN and ultimately may consume its resources. Moreover HO registration in MIPv6 incurs additional signaling overhead on MN, which is not acceptable International Conference on Green Computing and Internet of Things (ICGCIoT)

3 in case of resource constrained devices in lossy WSNs. To overcome this problem S. Gundavelli et al. [7] introduced PMIPv6 as a network-based localized mobility management protocol in PMIPv6 was aimed at relieving the MN from heavy signaling burden during HO registration process. The signaling process was supposed to be carried out by the network elements on behalf of MN. Researchers found PMIPv6 suitable protocol for mobility management in controlled and lossy WSNs. [9, 10] used PMIPv6 as a solution to mobility of body sensors. Network components (MAGS) were stated responsible for signaling messages and no need of duplicate address detection (DAD) check is required. The control information is exchanged only between LMA and MAG to reduce signaling cost and handoff latency of each body SN. A scheme based on dispatch type of 6LoWPAN was introduced in [11]. It reduces packet loss and HO latency significantly but only found suitable for intra domain issues. To reduce the packet loss, Inter Mario in [12] was proposed. Inter Mario was based on the strategy of make before break and sensor node (acts as AP) is responsible for pre-configuring the HO. Pre-configuration of HO helps in preventing packet loss to great extent but there is further scope of reduction. [17] Compared the proposed SH-WSN6 with the traditional MIH- PMIPv6, FPMIPv6 and PMIPv6. Although improvements were shown in the results but it lags in preventing the packet loss in critical applications. In [18], routing and mobility solutions were presented for 6LoWPAN mesh networks but the challenge of requirements and resources for such networks was stated in the article. Recently in [19], a survey of mobility in WSN and a proposed solution based on NoP [20, 21] was presented. NoP overlay on WSN and relives the SNs from signaling and processing tasks. NoP found beneficial in terms of HO time and node energy consumption but use of additional structure which causes more expenditure is major downside of proposed system. Also unnecessary HOs can occur due to the NoP s complex structure. From the related work available on the different resources of literature, we have concluded that mobility in lossy networks is still an issue. Although tremendous research in the field is being carried out but it is mainly focused on HO latency issues in such networks. Packet loss in critical areas such as Healthcare, where sensor nodes are deployed to track patient s mobility. Health related data is captured by body sensors and transmitted over IPv6 network to monitoring station where an available expert (doctor) can give immediate advice by interpreting the received data. Any loss in such time critical data can put patient s life in danger. Therefore there should be a proper and efficient mobility management scheme which can prevent or reduce packet loss and signaling overhead for lossy and time critical network applications such as battle field, disaster management, environmental monitoring and intelligent transport system etc. IV. EXISTING MOBILITY SCHEMES As stated above, MIPv6 is the basic mobility scheme for providing uninterrupted and seamless network services to mobile users. Extensions of MIPv6 were developed over time to make improvements in terms of HO latency and signaling overhead. MIPv6 and its extensions involve the mobile hosts or MN in signaling process required to carry out mobility and therefore need IP stack modifications [5, 6]. Due to this reason, researchers in the field have declared that these host based protocols are not suitable to carry out the mobility in controlled and lossy networks. The involvement of MN in signaling messages for registration purpose may consume its resources and hence these protocols are not suggested for sensor mobility in lossy networks. Therefore, network based mobility management schemes were introduced where MN is relieved from heavy signaling overhead and makes lossy networks more reliable by increasing sensor node s life. NEMO (network mobility) was the basic network based mobility protocol [22]. In NEMO, whole network along with mobile router is made mobile to carry out the mobility instead of one single MN or more MNs. Due to change in the point of attachment of the whole network to the access link; it was not standardized by IETF. The only standardized network based mobility protocol by IETF is PMIPv6 [7]. A. PMIPv6: PMIPv6 allows MNs to switch AP without any signaling cost. Two new entities LMA and MAG are introduced. LMA is responsible for maintaining reachability state of MN and MAG performs signaling on behalf of MN. Due to this reason PMIPv6 is suggested by many researchers for lossy WSNs. Different researchers used PMIPv6 in different perspectives and applications. Fig. 1 shows the basic PMIPv6 domain where LMA is a topological anchor point and MAGs perform signaling on behalf of MNs [7]. Fig. 1: PMIPv6 basic domain When MN handovers from PMAG to NMAG, the signaling messages exchanged for MN s registration with NMAG are shown in the fig. 2. NMAG will be responsible for detecting and registering MN s movement on its access link and sends PBU message to containing MN-Identifier to LMA. LMA updates its BCE and assigns HNP to MN. A proxy binding acknowledgement (PBAck) message containing HNP is sent to NMAG which updates its binding cache and sends a RA message to MN. MN configures its IP address by using the received HNP and a connection between MN and LMA is established. Now CN can communicate with MN via LMA and NMAG [7, 10] International Conference on Green Computing and Internet of Things (ICGCIoT) 657

4 In PMIPv6, although MNs are not involved in mobility related signaling, yet they suffer from longer HO latencies and packet loss (see equations (1), (2) and (3)). This longer delay and packet loss is not acceptable for some time critical applications in lossy networks. Our proposed EMS considers these issues and makes enhancements to improve QoS. V. PROPOSED SCHEME To achieve the goal of reducing HO latency, signaling cost and packet loss, we proposed a PMIPv6 based enhanced mobility scheme (EMS). It works on the basis of predictive HO mechanism (i.e. make before break). Fig. 2: PMIPv6 Signaling flow diagram Fig. 2 illustrates the working of PMIPv6 and its signaling flow. This signaling flow causes delay in the registration process and total delay composed is called HO latency. Due the signaling overhead and greater HO latency, there is some packet loss. Signaling cost is another parameter which is responsible for performance. B. Theoretical Analysis: There are many parameters such as HO latency, PLR, Signaling cost, throughput, Jitter, Packet delivery ratio PDR etc. which are used for performance analysis. Here, in this article we considered HO latency, PLR and throughput only because these are the import parameters in mobility management scenarios. HO latency: The total delay time from the moment when MN stops receiving packets from previous AR to the moment when it starts receiving packets from new AR during HO. We denote total HO latency by T PMIPv6, It is calculated by the addition of different delays according to the following equations (for notations: refer table 1): T PMIPv6 = T L2 + T RS + T MN-ID + T PBU + T HNP + T PBA + T RA + T ADD + T NMAG-MN (1) Average HO latency (T AVG-PMIPv6 ) can be calculated by dividing the T PMIPv6 by number of MNs handover; here in this article we are considering only one MN, so it will be same as T PMIPv6. T AVG-PMIPv6 = (T L2 + T RS + T MN_ID + T PBU + T HNP + T PBA + T RA + T ADD + T NMAG-MN ) / (no. of MNs) (2) PLR: Total number of data packets lost divided by total number of transmitted data packets. We denote PLR by PMIPv6_PLR in PMIPv6 domain and it is calculated in terms of delays during which packets are lost (for notations: refer table 1). PMIPv6_PLR =. P L. (T L2 + T RS + T MN_ID + T PBU + T HNP + T PBA + T RA ) (3) Fig. 3: Signaling Flow of Proposed EMS EMS makes some improvements in signaling by making Rtr. Adv. and multicast messages very few, required for registration process. Also these lossy WSNs are controlled networks and mobile wireless sensor nodes need not to register again and again while roam in the same PMIPv6 domain. EMS also avoids multi-hop communication by deploying more APs, although it will increase the expenditure but provides less complexity and single hop communication with seamless mobility of sensor nodes. A buffering mechanism to prevent packet loss is embedded with EMS which proved very efficient for controlling packet loss. While MN moves from one access domain to another access domain, data packets are buffered at either of the MAGs and delivered after the completion of HO via the established tunnel between the two MAGs. Fig. 3 shows the signaling flow required for registration process. Since EMS is particularly proposed for lossy WSNs, therefor MN is denoted by MSN. A. The registration Process: PMAG receives anticipated HO request (HO info) from AP or MSN. HO-info contains the information about neighboring APs or MAGs. PMAG easily derives the address for NMAG. PMAG initiates the HO and informs its L2 link (HO init) and also sends handoff initiate (HI) message (containing MSN identifiers and timestamp) to NMAG. Then NMAG sends a PBU message to LMA by using the timestamp in HI message. LMA installs the new binding International Conference on Green Computing and Internet of Things (ICGCIoT)

5 information in its cache and returns a PBAck message to the NMAG. Then LMA forwards all traffic intended for MSN to NMAG. Upon receiving the PBA message, NMAG sends handover acknowledgement (HAck) message to the PMAG and initiate buffering of packets for MSN until it connects to the NMAG. After connection establishment with NMAG, all the buffered packets are tunneled to MSN. The PMAG after getting the HAck message, sends HO-complete message to instruct its MSN to get disconnected from PMAG and connect to NMAG if not already connected. MSN uses address auto configuration property of IPv6 and sends NDP message request to NMAG. Soon after this MSN starts receiving traffic from LMA via NMAG. B. Theoretical Analysis: 1) HO latency: Theoretical analysis for HO latency and Packet loss for EMS is given in the following equations. We denote the HO latency by T EMS and is calculated as (for notations, refer table 1): T EMS = T L2 + T (HO-init+HI) + T PBU + T PBA + T HAck (4) When we compare equation (4) with equation (1), we observed that the proposed scheme reduces the HO latency with reduced signaling overhead. 2) Packet loss: If HO happens timely, then there is probability that no packet loss will incur as all arriving packets at PMAG will be delivered to MSN. The packets arriving at NMAG will be buffered during HO and delivered to MSN after HO completion. If the HO prediction does not happen timely, packets that are already sent by the PMAG to its AP before the PMAG gets notified of disconnection of MSN are lost. This packet loss is mainly during delay time between PMAG and its AP. In this case if HO prediction was slightly delayed then the packet loss could be little lesser because some more packets would have delivered to MSN before its disconnection. Therefore handover packet loss because of untimely HO prediction is dependent on different delay factors. For analysis, we denote PLR with EMS_PLR for the proposed scheme and calculated as: Fig. 4: Studied Simulation Scenario for proposed Scheme MS or CN can be in the same domain or at far place over the internet. Here the delay link between MS/CN and LMA is 10ms and packet size 1000bytes with data rate of 250kbps. For simulation, first the network model is built, then implementation of computer program to study the behavior of proposed scheme and standard PMIPv6. A. Simulation Results: Results for different performance metrics such as HO latency and Packet loss ratio were taken against varying speed of MSN/MN and WLD. MSN speed is varied from 1m/s (walking speed) to 10 m/s. The initial delay value for WLD is kept 10ms. Fig. 5(a) and fig. 5(b) show the graphs for HO latency versus MSN speed and WLD respectively. EMS_PLR =. P L. (T U-Pred + T MAG-AP + T PMAG-NMAG ) (5) When we compare equation (5) with equation (3), we find lesser PLR for proposed scheme in comparison to standard PMIPv6. VI. SIMULATION AND RESULTS The proposed scheme is simulated to validate the theoretical analysis. We used NS2 simulator on LINUX (fedora) platform. The studied simulation scenario is shown in fig. 4. The MAGs in same LMA1 (or LMA2) domain are kept 100m apart which shows micro mobility. The distance between MAGs of different LMA domains is kept 150m which shoes macro mobility. The coverage area of each AP under MAGs is set to 50m of radius. The links use the Droptail (FIFO) queues and one MSN shown with stochastic movement. Fig. 5(a): HO latency vs MSN speed Fig. 5(b): HO latency vs Wireless Link Delay 2015 International Conference on Green Computing and Internet of Things (ICGCIoT) 659

6 We observed that the proposed EMS shows little reduction in HO latency in comparison to standard PMIPv6. The increase in HO latency in both the graphs is mainly due to (i) macro mobility handover from NMAG_1 to PMAG_2 and (ii) increased WLD from 10 ms to 100ms. Fig. 6(a) and fig. 6(b) are plotted for PLR vs HO latency and WLD respectively. Fig. 6(a): Packet Loss Ratio vs MSN Speed Fig. 6(b): Packet Loss Ratio vs Wireless Link Delay Again we observed that the proposed scheme suffers less packet loss when compared to PMIPv6. In fig. 6(a), initially PLR increases due to the lost connection while MSN experiences macro mobility from NMAG_1 to PMAG_2 and finally it comes down with the increased MSN speed as MSN reaches in new (LMA2) domain. In fig. 6(b) PLR increases with increase in WLD and then it starts decreasing when HO is completed. This increase is due to macro mobility among two LMA domains. In both the cases, proposed scheme performs better when compared to standard PMIPv6. VII. CONCLUSION In this paper, we proposed a scheme to handle mobility in lossy and controlled WSNs. Due to the host involvement in mobility, host based protocols are not suitable for mobility in such domains. Already standardized NETLMM PMIPv6 suffers from packet loss and extra signaling and further needs some improvements. Proposed scheme performs better with reduced HO latency, signaling cost and PLR when compared to PMIPv6. Also in IOT, sensor applications such as healthcare monitoring, environment monitoring, disaster management and intelligent transport system are getting more and more importance with mobility as essential requirement. Mobility in such domains should be carried out by network side to protect sensor node s resources for network reliability. The proposed scheme performs better than PMIPv6 for such networks. In future an attempt would be made for test bed implementation of proposed scheme with consideration of compatibility issues of IEEE networks with lossy IEEE networks. Also a study of security issues in mobility management in such domains would be attempted. REFERENCES [1] D. Miorandi, S. Sicari, F. De Pellegrini and I. Chlamtac, Internet of things: Vision, applications and research challenges, J. Ad Hoc Networks, 2012, PP [2] G. Mulligan, The 6LoWPAN architecture, Proceedings of 4 th Workshop on Embedded Networked Sensors, ACM, [3] G. Montenegro, N. Kushalnagar, J. Hui and D. Culler, Transmission of IPv6 Packets over IEEE Networks, Internet Proposed Standard RFC [4] A. Ahmada and D. Sasidharan, Handover efficiency improvement in Proxy Mobile IPv6(PMIPv6) networks, in: Elsevier Procedia Computer Science, ( 2015 ) [5] Soliman H, Castelluccia C, El Malki K and Bellier L., Hierarchical mobile IPv6 mobility management (HMIPv6), RFC 4140, August [6] Koodli R., Fast handovers for mobile IPv6, RFC 4068; July [7] Sri Gundavelli, Kent Leung, Vijay Devarapalli and Kuntal Chowdhury, Proxy Mobile IPv6. IETF Request for Comments 5213, August [8] I. force, Mobility support in IPv6, RFC 6275, [9] M. Shin, T. Camilo, J. Silva and D. Kaspar, Internet draft-6lowpan mobility, Internet-Draft, Nov [10] Proxy Mobile IPv6 [Online]. Available at: , Aug [11] Bag, G., Raza, M.T., Kim, K.-H. and Yoo, S.-W., LowMob: Intra PAN mobility support schemes for 6LoWPAN, Sensors, Vol. 9, no.7, Jul. 2009, PP [12] Minkeun Ha, Daeyoung Kim, Seong Hoon Kim and Sungmin Hong, Inter-MARIO: A fast and Seamless Mobility protocol to support Inter_PAN Handover in 6LoWPAN, In Proc. Global Telecommunications Conference (GLOBECOM), Dec. 2010, PP [13] J. Ko, A. Terzis, S. D. Haggerty, David E. Culler, Jonathan W. Hui and P. Levis, Connecting Low-Power and Lossy Networks to the Internet, IEEE Communications Magazine, 2011, PP [14] P. Levis et al., The Emergence of A Networking Primitive in Wireless Sensor Networks, Commun. ACM, vol. 51, no. 7, 2008, pp [15] IEEE : Wireless Medium Access Control (MAC) and Physical Layer (PHY) Specfications for Low-Rate Wireless Personal Area Networks (LR-WPANs), available at pdf. [16] J. Hui and P. Thubert, Compression Format for IPv6 Datagrams in 6LoWPAN Networks, Internet draft, [17] Juha Petajajarvi and Heikki Karvonen, Soft Handover for Mobile Wireless Sensor Networks Based on 6LoWPAN, In Proc. Int. Conf. Distributed Computing in Sensor Systems and Workshops (DCOSS), June 2011, PP [18] Oliveira, L. M. L., de Sousa, A. F. and Rodrigues, Routing and Mobility approaches in IPv6 over LoWPAN mesh networks, Int. journal of communication system, Vol. 24, no. 11, Feb. 2011, PP [19] Ricardo Silva, Jorge SA Silva and Fernando Boavida, Mobility in wireless sensor networks-survey and Proposal, J. Computer Communication, (article in press), [20] R. Silva, J. Sa Silva, F. Boavida, A proxy-based mobility solution for critical WSN applications, in: IEEE International Conference on Communications Workshops (ICC), 2010, PP [21] R. Silva, J.S. Silva, F. Boavida, A proposal for proxy-based mobility in WSNs, Comp. Commun. 35 (10) (2012), PP [22] V. Devarapalli, R. Wakikawa, A. Petrescu and P. Thubert, Network Mobility (NEMO) Basic Support Protocol, RFC 3963, January International Conference on Green Computing and Internet of Things (ICGCIoT)