Untitled

IEEE Globecom 2010 Workshop on Ubiquitous Computing and Networks Inter-Mobility Support in Controlled 6LoWPAN Email: {zinonos,vasosv}@cs.ucy.ac.cy Abstract—The research and industrial community started to protocol. Those benefits include the easy interconnection with think of more complex application scenarios for wireless sensor other IP networks, the use of existing internet infrastructure, networks, where the use of mobile sensor nodes is essential. The the application of the well-known IP-based technologies to support of mobile sensor nodes in such applications requires theexistence of a suitable mobility management protocol. However, sensor networks, and the reuse of existing power monitoring existing mobility protocols, like MIPv6, can not be directly and diagnostic tools. The challenge in supporting IP protocols applied on mobile sensor nodes, since they are inefficient in terms in WSN is to overcome the limitations faced by sensor of energy, communication, and computation cost, and fail to meet networks, like lower power consumption, low duty cycles, and the stringent operational requirements of a mobile sensor node.
In this paper we propose a new mobility management protocol for6LoWPAN which uses the technology of Proxy Agents and aims to The most significant effort is performed by IETF 6LoWPAN enhance the handoff time by predicting or rapidly responding to working group, which aims to develop the support of IPv6 over a handover event. The proposed protocol lessens the involvement the standard IEEE802.15.4, in order to import the well known of the mobile node in mobility-related message exchange.
capabilities of IPv6, such as Neighbor Discovery (ND) andMobile IP (MIPv6) into low power devices [2]. A low-power wireless personal area network (LoWPAN) is a simple low Wireless Sensor Networks (WSNs) consist of small au- cost communication network that allows wireless connectivity tonomous devices that sense and collect information from in applications with limited power and relaxed throughput their immediate environment. The collected information is requirements. A LoWPAN typically includes devices that work transported hop-by-hop through the network to a control together to connect the physical environment to real-world station where it is analyzed and acted upon. In spite of the diverse applications, sensor networks pose a number of The remaining of this paper is organized as follows: the unique technical challenges due to their ad hoc deployment, next section presents the related work. Section III presents our unattended operation, and dynamic environment changes.
proposed solution to support mobility. In section IV we present Most sensor applications require the deployment to be the packet format of the mobility messages and in Section V infrastructure-less, without any human intervention. It is re- we evaluate the mobility solution. Concluding remarks and sponsibility of the sensor network to be adaptable to any items for future work are given in Section VI.
physical changes like the addition of extra nodes or the failure of some of them. In addition, there is only a finite source of energy, which must be optimally used for processing and com- Mobility can be in general approached from three points munication. An interesting fact is that communication domi- of view: the first one, and also the most common solution, nates processing in energy consumption. In current research is to deal with the handoff procedure locally in the Network studies and testbed deployments of sensor networks, individual Layer of the mobile entity (MIPv6, HMIPv6, MIFAv6). The sensor nodes are usually assumed to be static. However, the second approach uses information from Layer 2 to speed up research community has started thinking of more complex the handoff process (FMIPv6) and the third solution is based WSNs application scenarios, where interoperability between on a non-evasive method known as network-based mobility different WSNs and communication with other networks is (PMIPv6). Mobile IPv6 (MIPv6) [3] is the global mobility necessary. To support these applications different requirements protocol that aims to maintain IP connectivity for the mobile are needed, such as the use of mobile sensor nodes, which pose node (MN). In MIPv6, when a MN moves from one network to some unique challenges in aspects like resource management, the other, it forms a care-of address (CoA) based on the prefix topology control, routing protocols, security, quality of service, of the foreign link. Thereafter, the node sends the Binding Recently, there have been attempts also to integrate Internet Enhancements of MIPv6 are HMIPv6 [4] and FMIPv6 [5] protocol (IP) with the WSN [1]. The motivation behind this protocols. In HMIPv6 new entity called Mobile Anchor Point attempt is to exploit the benefits offered by the use of the IP (MAP) is introduced which acts as a local HA of the MNs.
In this way there is a separation between global mobility and increases. A message format of location update (LU) message local mobility. FMIPv6 aims to reduce the handoff delays for and data packets is proposed. Finally, in [15], Bag et al.
mobile connections by anticipating their IP layer mobility and presented a network-based intra mobility scheme for mobile discovering the new router prefix before being disconnected 6LoWPAN nodes in which the mobility of 6LoWPAN nodes from the current router. Another form of mobility is NEMO [6] is handled at the network-side. LoWMob ensures multi-hop whereby the whole network moves together and its mobility is communication between gateways and mobile nodes with the handled by the mobile routers. The above IP mobility protocols help of the static nodes within a 6LoWPAN. In order to reduce are referred as ’host-based’ mobility management protocols the signaling overhead of static nodes for supporting mobile because the MN is involved in the processing of mobility nodes, LoWMob proposes a mobility support packet format management. In some other cases the MN is not involved at the adaptation layer of 6LoWPAN. Also they present a in mobility management, thus the solutions are referred as distributed version of LoWMob, named as DLoWMob, which ’network-based’. An example of a network-based protocol is employs Mobility Support Points (MSPs) to distribute the Proxy Mobile IPv6 (PMIPv6) [7] where a new entity, the traffic concentration at the gateways and to optimize the multi- mobile access gateway (MAG), is responsible for performing hop routing path between source and destination nodes in a the mobility management on behalf of the MN. The IETF 6LoWPAN. Moreover, they have also discussed the security 6LoWPAN working group has already released two related considerations for their proposed mobility schemes. However, RFCs. RFC 4919 [8] provides an overview, defines the as- this solution supports only the intra mobility scenario and sumptions, identifies the problems and establishes goals to requires extra hardware(antennas) in order to obtain the Angle improve IPv6 over LowPANs. The second RFC, RFC 4944 [9], of Arrival (AoA) measurements. In addition, each SN must defines the transmission of IPV6 Packets over IEEE 802.15.4 be equipped with a radio-triggered hardware component that and the LoWPAN adaptation layer and frame format, a layer activates sensors from/to the sleep state, by sending a special responsible for fragmentation and reassembly, providing the connectivity between the MAC frames and the IP packets.
Due to the well-known constrains of WSNs, we believe that The latest draft about mobility in 6LoWPAN [10] identifies a network-based mobility is theoretically the most desired the mobility scenarios, the main challenges and security issues solution to be applied to sensor networks. Solutions like the for this types of networks. In addition a compression scheme MIPv6 cannot be efficiently applied to mobile sensor networks about applying MIPv6 in 6LoWPAN is presented in IETF since are usually resource restricted and their involvement in draft [11]. Kim et al. [12] proposed a compressed packet mobility management process consumes battery and incurs header format to support the mobility of 6LoWPAN and a Lightweight NEMO protocol to minimize the signalingoverhead between 6LoWPAN mobile routers and 6LoWPAN gateways by using a compressed mobility header. To pro- In order to support a network-based solution we introduce vide mobility for 6LoWPAN nodes, they adopt the Network an entity called 6LoWPAN proxy agent (PA). The 6LoWPAN Mobility (NEMO) protocol. To support mobility headers in PA belongs to the Full Function Devices category. We consider 6LoWPAN packets, a new dispatch header pattern is defined, that the PA, as an FFD, is a ”powerful” device and will LOWPAN MH, to add a compressed IPv6 mobility header typically have more resources and may be mains powered.
to a dispatch. Kim et al. [13] also proposed an interoperable The use of 6LoWPAN proxy entities is supportive in mobility architecture between NEMO and 6LoWPAN. To accomplish scenarios in order to reduce the number of signaling between the inter operability, they have enhanced the routing protocol: the MN and the Home Agent. The idea is to force the An extended LOAD routing scheme for mobile routers to sup- 6LoWPAN PA to be responsible for MNs mobility signaling port mobility in 6LoWPAN sensor nodes. Enhanced routing and operation with the HA and just inform the MN, whenever performs default gateway discovery and mobile network prefix necessary, to handoff to a new serving proxy (within or outside discovery operations for packet forwarding, path optimization the same PAN). Any other data that is sent to, or received from, and backup route maintenance. The drawback of last two the mobile node is routed using multi-hop communication via solutions is that can not deal with individual node mobility the other nodes of the PAN. We consider that there is a routing and are clearly beneficial when applied to mobile 6LoWPANs.
protocol that is suitable for multi-hop routing inside the PAN Bag et al. [14] use an adaptation layer packet format for mobility signaling from 6LoWPAN. The proposed mobility When the MN is switched on for the first time it has to support mechanism reduces the Inter-PAN handover time by establish L2 connectivity by receiving beacons and associate providing some extra information about the frequencies of with a specific channel. After that, it should run the Neighbor the surrounding PANs at the border nodes. The performance Discovery (ND) procedure to retrieve information from the evaluation is based on the mobility signaling cost incurred to network like the network prefix. Finally, the Duplicate Address provide mobility in the network, which is given by analytical Detection (DAD) procedure should be run to ensure that the modeling. The results clearly indicated that less number of global address obtained by the MN is unique. The DAD signaling (no of bits) is required as compared to HMIPv6 procedure should be run by the backbone Router to unload the when speed and packet arrival rate of the mobile device MN from time and resource consuming operations. The above RS = Router Solicitation, RA=Router Advertisement, NR= Node Registration NC=Node Confirmation, NS =Neighbor Solicitation, NA=Neighbor Advertisement of node mobility, where a single node moves between pointsof attachment.
While the MN is moving inside the same PAN the 16-bit short address that is used to reach the node is not changed.
Thus, in the context of this paper we did not consider theintra-PAN mobility movement as we suppose that there is acapable routing protocol that is used to handle this kind ofmobility. In this section, we study the procedure of how theMN parent proxy is changed inside the same PAN.
Our proposed mobility model is based on the following 1) 6LoWPAN PAs located in the same network will adver- three steps can be grouped in one phase called bootstrapping.
During the ND procedure the device is auto-configured and 2) The address of the MN inside the same network is not gets a global IPv6 address. The global address is constructed based on the network prefix and the 64-bit interface identifier 3) All the data from/to the MN are not routed via the PA (IID). In 6LoWPAN networks it is assumed that the IID has but from the other neighbors nodes based on the routing a direct mapping to the link-layer address; as a result address resolution is avoided. Inside each PAN we use a unique 16- 4) The proxy node and MN exchange packets when the bit address instead of using the IPv6 address. The edge router is responsible to map this 16-bit address to the global IPv6address in case there is a need to communicate with a node New proxy-based Intra-PAN mobility procedure: outside the 6LoWPan network. This procedure has a strong 1) The 6LoWPAN PA measures the RSSI value of the link relationship with security and an appropriate key management option has to be followed. The ND procedure signaling is 2) If the RSSI is below a predefined threshold then the current 6LoWPAN PA will inform the surrounding PAs In this paper we deal with the following two types of about MN movement and will instruct them to start 1) Device movement within a single Wireless PAN domain 3) When the PAs capture packets from the MN they will (Intra-PAN or micro-mobility): This scenario is probably create a Join message where in the payload field they the most common in WSNs architectures. The sensor will include the MN 16-bit address, PAN ID and the node moves within the domain without losing the con- RSSI value of the message that they captured. This message will be sent to the parent proxy via the MN’s 2) Device movement between multiple Wireless PAN do- mains (Inter-PAN or macro-mobility): Sensor nodes 4) When the parent PA receives a join packet it will check if move between different sensor networks, each one with the included (in the payload) RSSI value is acceptable its Sink node responsible to configure and manage all and if the PAN ID is the same. If yes, it will answer the aggregated devices. In this type of mobility, the IPv6 to the proxy node with an accept message (Join Ack packet). In case more than one surrounding proxies Both mobility types described above belong to the category exist,the parent proxy will receive more than one Join packets and will accept the one with the stronger RSSI value. This means that it must be pre-configured to wait for a specific time interval before sending a Join Ack message in order to make sure that all possible Joinmessages arrive.
5) The new proxy will then inform then the edge router that the MN is now served by him (location update).
The previous proxy will inform the MN with a Proxy Confirmation message about the new attachment point.
B. The Inter-PAN Mobility Scenario FCBU = Fast Binding Update FCBA = Fast Binding Acknowledgement In this section, we study the mobility of the MN between BU=Binding UpdateBA= Binding Acknowledgement Proxy Conf = Proxy confirmation PAN networks. Our proposed mobility model is based on thefollowing assumptions : 1) 6LoWPAN proxy agents located in different networks 2) Home Agents have direct access between them.
3) The address of the MN when is moving between differ- 4) The parent proxies have all the necessary information to perform the Duplicate Address Detection (DAD) 5) The Home Agents are responsible to create the 16-bit address of the MN that will be used to route packetsinside the 6LoWPAN network.
6) The Home Agents will maintain a binding table where the Binding acknowledgement will be sent to the HA by the Global IPv6 address, the Care-of Address, the 16- the backbone router and then finally the fast compressed bit address, the parent proxy address and a timestamp is BA will be sent to the parent proxy. The HA will start buffering the data that are destined to MN untilit receives the Location Update message.
5) Finally, the Proxy Confirmation will be sent to the MN from the previous proxy so that to confirm the 16-bit IPv6 address to use. At this time, the MN is routable 6) In order to avoid any security threats we aim to use The handoff procedure is initiated by the proxy and is 1) The first 3 steps are the same as in new proxy intra proactive. The basic idea behind this is to leverage information from the link-layer technology to either predict or rapidly 2) When the parent PA receives a join packet it will check respond to a handover event. In that way, we can reduce the if the RSSI value is acceptable and if the PAN ID is the handoff time and, subsequently, any packet losses and delays.
same. If the PAN ID is different it will assume that the Moreover, the MN should be left out of any mobility signaling.
The signaling of the proposed solution is presented in Fig. 4.
3) In case more than one surrounding proxies exist, the Mainly, the networks (PANs) use different frequencies in parent proxy will receive more than one Join packets order to avoid interference. For that reason, the proxy agents and will accept the one with the stronger RSSI value.
at the edges of PANs that is configured to serve the MN should This means that it must be pre-configured to wait for a be able to capture packets from different frequencies and PANs specific time interval before sending a Join Ack message as well, if any in range. In order to handle the scenario where in order to make sure that all possible Join messages the MN is moving without communicating, the proxies must be configured to send advertisements to their serving MNs in a 4) After the expiration of the predefined interval the PA predefined interval (every second) so that to guarantee that the will decide which proxy is the best for the MN and it node mobility can be detected. Additionally, the MN will be will create and send the fast compressed binding update aware of that interval. If the advertisement is received inside (FCBU) to the HA. The HA will create a Binding Update that interval the MN will continue its normal behaviour, as message and sent it to the backbone router. After that, shown in Fig. 5 (a). In case that the advertisement interval The main overhead in the proposed solution is the Frame Header where is 22 bytes. The Frame Header structure is As discussed previously, the packet formats are based on the LOWPAN IPHC encoding compression. In order to support the 3 different types of communication (multicast, link-local and global) we define the following cases for the values of LOWPAN IPHC as shown in Table II. The LOWPAN IPHC utilizes 13 bits and uses the 5 rightmost bits of the dispatch type. In addition to the LOWPAN IPHC encoding we definedthe mobility header using the LOWPAN NHC encoding so that to support the mobility signaling of the proposed solu- tion. The mobility header is compressed based on [11]. The proposed mobility header is shown in Table III.
value is expired without receiving any advertisement the MN will transmit an alive message to the parent proxy and will wait for receiving an acknowledgement. If not acknowledged then the MN wait to receive advertisement packets from another proxy node as shown in Fig. 5 (b). With this solution, we can assure that it is possible to control the maximum disconnection time during handoff and that the MN movement can alwaysbeen detected.
In this section, we evaluate the handover delay of the In this section, we define the format of the mobility mes- proposed mobility model analytically. We consider the inter sages of our proposed solution. The format is recommended mobility scenario where the MN is moving from the home based on the RFC 4944 [9] and the relevant IPv6 header com- network PAN1 to the visited network PAN2. For that purpose, pression format for IPv6 packet delivery in 6LoWPAN net- we will use the topology shown in Fig. 2. A mobile sensor works [16]. The compression format relies on shared context node is unable to receive IP packets on its new association to allow compression of arbitrary prefixes. For that purpose point until the handover process finishes. The period between the 6LoWPAN working group defined a new encoding for the transmission (or reception) of its last packet through the compressing IPv6 header, called LOWPAN IPHC. In addition, old connection and the first packet through the new connection we used a new encoding format for arbitrary next header, is the handover latency. The handover latency in mobility protocols, in general, is affected by several components: The proposed messages format is shown in Fig. 6. Based 1) Link Layer Establishment Delay (DL2): The time re- on the proposed mobility model the communication can be quired by the physical interface to establish a new separated in two categories: intra communication when nodes inside the same PAN exchange messages and inter com- 2) Bootstrapping(DBS): The time required for the mobile munication when the communication is between two nodes node to run the ND (DND) and DAD (DDAD) proce- that belong to different PANs. In the former case the link local unicast addresses are used and the IPv6 header can be 3) BU/Registration Delay (DREG): The time elapsed be- compressed down to 6 bytes. In the latter case the global tween the sending of the FCBU from the proxy sensor addresses are used and the header can compressed down to 10 to the HA and the arrival/transmission of the first packet.
bytes. Additionally, the compression gain of the IPv6 header is 4) Processing and queuing delay (DPROC): The time re- significant also in the case of well-known multicast addresses quired form processing the messages and the delay that as LOWPAN IPHC can compress the header down to 7 bytes.
The general handover delay for all the protocols can analyti- bytes without the proxy node thus all the signaling is handled by the MN and the case where the proxy exists and takes careof the mobility signaling (as proposed).
D = DL2 + DBS + DREG + DPROC In the case of our model the DDAD and the DREG delays Critical applications require mobile sensor nodes to be the are not part of the handoff time since they are performed uniquely addressable. To support the movement of those sensor prior to the disruption point; the address test performed during nodes a new mobility protocol is needed, since the existing handover to reduce handover latency. The handoff procedure is solutions can not be applied to such networks. In this paper started at the time the MN receives the Router Confirmation we proposed a mobility model that promises soft handoffs and message and finished by the arrival of the first data packet controlled disconnections between different PANs. The model routed via the visited 6LoWPAN network. Thus the handoff is based on the involvement of a proxy node that is responsible delay is equal to DL2 + DPROC. In order to evaluate the to handle, on behalf of the MN, the mobility-related messages.
signaling cost in bytes we firstly calculate the number of bytes This work has performed a detailed decomposition and anal- of each message that are sent inside the 6LoWPAN networks.
ysis of the handover delay and has shown the validity and Table IV depicts the values. When a MN is moving from one applicability of the proposed solution.
This work has been conducted under the European Union Project GINSENG funded under the FP7 Program (FP7/2007- [1] M. Durvy, J. Abeill´e, P. Wetterwald, C. O’Flynn, B. Leverett, E. Gnoske, M. Vidales, G. Mulligan, N. Tsiftes, N. Finne, and A. Dunkels, “Making sensor networks ipv6 ready,” in SenSys ’08: Proceedings of the 6th ACM conference on Embedded network sensor systems. New York, NY, USA:ACM, 2008, pp. 421–422.
[2] 6LoWPAN, “6lowpan charter.” [Online]. Available: http://datatracker.
network to another the signaling cost in bytes is equal to the [3] C. Perkins, D. Johnson, and J. Arkko, “Mobility support in ipv6, request total transmit bytes plus the total received bytes. The total for comment 3775,” Internet Engineering Task Force, RFC 3775, 2004.
transmit bytes are equal to trigger bytes + FCBU bytes + Join [4] H. Soliman, C. Castelluccia, K. El-Makri, and L. Bellie, “Hierarchical Ack bytes = 103 bytes. The total received bytes are equal to mobile ipv6 mobility management (hmipv6),” Internet Engineering TaskForce, RFC 4140, 2005.
T otal = #joins ∗ 37bytes + F CBAckbytes [5] R. Koodli, “Fast handovers for mobile ipv6,” Internet Engineering Task +P roxyConfbytes = 109bytes [6] V. Devarapalli, R. Wakikawa, A. Petrescu, and P. Thubert, “Network mobility (nemo) basic support protocol,” Internet Engineering TaskForce, RFC 3963, 2005.
[7] S. Gundavelli, K. Leung, V. Devarapalli, K.Chowdhury, and B. Patil, “Proxy mobile ipv6,” Internet Engineering Task Force, RFC 5213, 2008.
[8] N. Kushalnagar, G. Montenegro, and C. Shumacher, “Ipv6 over low- power wireless personal real networks (6lowpans),” Internet EngineeringTask Force, RFC 4919, 2004.
[9] G. Montenegro, N. Kushalnagar, J. Hui, , and D. Culler, “Transmission of ipv6 packets over ieee802.15.4 networks,” Internet Engineering Task [10] G. Mulligan, C. Williams, and D. Huo, “6lowpan architectural consid- eration for mobility,” Internet Engineering Task Force, Internet-Draft draft-williams-6lowpan-mob-03.txt, July 2010, work in progress.
[11] R. Silva and J. S. Silva, “An adaptation model for mobile ipv6 support in lowpans,” Internet Engineering Task Force, Internet-Draft draft-silva- [12] G. Kim, C. Hong, and T. Shon, “A lightweight nemo protocol to support 6lowpan,” ETRI Journal, vol. 30, pp. 685–695, 2008.
[13] J. Kim, C. Hong, and K. Okamura, “A routing scheme for supporting network mobility of sensor network based on 6lowpan,” in Managing Next Generation Networks and Services, ser. Lecture Notes in ComputerScience, S. Ata and C. Hong, Eds. Springer Berlin / Heidelberg, 2007,vol. 4773, pp. 155–164.
[14] G. Bag, H. Mukhtar, S. M. S. Shams, K. H. Kim, and S. wha Yoo, “Inter-pan mobility support for 6lowpan,” vol. 9, 2009, pp. 5844–5877.
[15] G. Bag, M. Raza, K.-H. Kim, and S.-W. Yoo, “Lowmob: Intra-pan Thus, if we consider one join inter message the total bytes mobility support schemes for 6lowpan,” CSensors, vol. 1, pp. 787–792,2008.
that are required for the proposed solution are equal to 212 [16] J. Hui and P. Thubert, “Compression format for ipv6 datagrams in bytes. Based on our solution only the reception of the 44 bytes 6lowpan networks,” Internet Engineering Task Force, Internet-Draft involves the MN. Fig. 8 depicts the case of signaling cost in draft-ietf-6lowpan-hc-07, 8pr 2008, work in progress.

Source: http://wmn.csie.ntpu.edu.tw/attachments/article/136/Inter-Mobility%20Support%20in%20Controlled%206LoWPAN%20Networks.pdf