Elsevier

Computer Networks

Volume 55, Issue 5, 1 April 2011, Pages 1069-1082
Computer Networks

Distributed geographic service discovery for mobile sensor networks

https://doi.org/10.1016/j.comnet.2010.09.015Get rights and content

Abstract

Many of the recent sensor network applications assume node mobility. The service provisioning protocol should therefore be designed to be stateless and scalable for continuously changing link states of typical mobile sensor nodes. In this paper, we propose DSDP, Distributed Service Discovery Protocol, for mobile sensor networks. DSDP organizes the network area into hierarchical regions to reduce service update cost. Regions at the lowest hierarchy cover a two-hop range area, and nodes freely move within the larger area without updating the service coordinators. Due to the low hierarchy architecture, the increased size of the minimal regions helps to reduce the runtime overhead. Node mobility is specifically considered in DSDP. Depending on the nodes’ mobility, nodes in DSDP work as service coordinators for requested services, and the discovery protocol runs in a distributed manner to lessen the overhead. Both the simulation and experimental results show that the proposed mechanism achieves a high service query success rate and low service update overhead in mobile sensor networks.

Introduction

In traditional sensor network applications such as environment monitoring, sensor nodes are statically deployed to monitor specific events. Recent applications, however, include Mobile Sensor Networks (MSNs) where sensor nodes move freely throughout the network, carried by people, robots, or vehicles. The mobility of nodes has raised issues in handling dynamic link changes or mobility patterns that have not been considered in conventional sensor networks. In particular, the mobility of the nodes requires related algorithms to be stateless because link states between mobile nodes are valid for only a short duration. In addition, the underlying communication protocols in MSNs need to be scalable and robust to endure dynamic environments. Many technical challenges are, therefore, newly addressed in MSNs.

Sensor nodes may contain various resources such as storage, processing power, GPS receivers, cameras, or other sensors. A service discovery protocol needs to be provided so that nodes can discover and use the desired service among a wide variety of resources. Service coordinators play the service discovery role, coordinating the queried service to the corresponding node. A service discovery protocol in MSNs should be designed to coordinate the network for which sensor nodes provide services in a timely manner with high reliability and minimum infrastructure. The node mobility causes several issues. First, query results could be invalid if the query latency is long or a destination node continuously moves to another location. Shortening the query path length is therefore an important issue in MSNs. Furthermore, high mobility causes frequent service updates that lead to severe congestion and hinder the proper delivery of packets. Finally, the movement of the service coordinators should be considered. When every node works as a service coordinator, query packets cannot reach their destinations if the coordinators move while the packets are in transit. This problem could seriously deteriorate the reliability of the system, especially for systems with many mobile nodes.

Conventional approaches [1], [2], [3], [4], [5], [6] for service discovery protocols in MANET (Mobile Ad-hoc Network) are not suitable for MSNs, since nodes in MANET are usually considered to have powerful hardware and support reliable communications. Since flooding-based protocols [1], [3] cause unnecessary overhead, specific protocols were developed to reduce the flooding overhead by centralizing the service coordinators [2], [4]. This approach, however, causes congestion around the coordinators, and consequently leads to severe power consumption and communication errors in the nodes around the service coordinators. Beacon Location Service (BLS) [6] works without pre-configuring the network layout by integrating Beacon Vector Routing (BVR) [7]. The mechanism, however, requires a number of centralized location servers. To overcome the scalability problem of the above protocols, geographic service discovery protocols [8], [9] were proposed. However, these protocols hardly consider node mobility [8] or the characteristics of sensor networks [9] that are normally equipped with limited computing and communication resources. Scalable Geographic Service Provision (SGSP) [9], in fact, considered node mobility, but the protocol is not applicable for sensor networks.

In this paper, we propose DSDP, Distributed Service Discovery Protocol, for mobile sensor networks. Application scenario of DSDP includes a variety of sensor network applications. One example is habitat monitoring [10]. A Service Discovery Protocol (SDP)-based habitat monitoring provides information on objects themselves, as well as on environments where objects move around. Moreover, the information is shared with anyone who is interested in while not knowing which device provides what kind of information. For example, a person may look for the number of sightings of a particular animal, the highest temperature recorded so far, or the humidity level in a particular area, and so on. The flexibility to look for any desired service comes from standardization of the service discovery framework. SDP-based applications potentially save the developers a significant amount of time over developing dedicated systems for each application [11]. Hence, DSDP is designed to overcome the drawbacks of previous work on service discovery protocols for sensor networks: i.e. the lack of support for both mobility and scalability. In contrast to other geographic service discovery protocols that confine the location to store the payload within a single-hop range, DSDP supports a two-hop range for the lowest hierarchy region. Compared to previous work, DSDP lowered the number of hierarchy levels of the service coordination; hence, service queries and updates are performed with low overhead. The location addressing scheme in DSDP supports non-rectangular shapes by using coordinates based on the best-fit rectangular region. DSDP exploits node mobility to differentiate roles as service coordinators; hence, node mobility is efficiently handled to improve performance in mobile networks. Most sensor nodes in DSDP act as service coordinators to distribute the overhead of maintaining the service discovery protocol.

The rest of this paper is organized as follows. In Section 2, we discuss related work. Sections 3 Distributed service discovery protocol, 4 Mobility support in DSDP describe the design of DSDP and details of the mobility policy, respectively. Section 5 evaluates DSDP through simulation experiments. Section 6 discusses the experimental results based on a real implementation of the system. Section 7 concludes the paper.

Section snippets

Related work

Traditional solutions for service discovery protocols can be classified into two approaches: centralized protocols and distributed protocols [12]. The centralized protocols, such as UDDI [13], use a central lookup server, whereas the distributed protocols rely on multicast or flooding. Most of the traditional MANET service discovery protocols are distributed and outperform centralized protocols in terms of scalability. Moreover, distributed protocols better suit the ubiquitous environments. In

Distributed service discovery protocol

Fig. 1 illustrates an overall view of DSDP. A node requests service information via DSDP by using specific keys such as target ID, service ID, or even attributes such as node position. Nodes in the network are mobile and move constantly. Various mobility policies and considerations for mobile service coordinators are proposed to support full mobility. The following sections explain the basic components of DSDP: region addressing scheme, service update, service query, and its coordination.

Mobility support in DSDP

DSDP classifies nodes according to their mobility level. This is reflected when DSDP distributes service information or manages the service information database. Nodes can communicate with each other only within a single hop in minimal regions. Currently, DSDP uses a modified GPSR [22] as an internal routing algorithm. The algorithm searches neighbors within two hops to maximize efficiency by utilizing two-hop neighbor management in DSDP. However, two-hop neighbor search cannot guarantee

Simulations

We performed simulations to evaluate DSDP based on the ns-2 network simulator [24]. The IEEE802.15.4 radio model with a maximum transmission range of 20 m was employed for our simulations. The simulation duration was 1000 s, and the simulation was performed in different topologies of an 120 × 80 m area with 75 nodes, an 160 × 120 m area with 125 nodes, an 180 × 160 m area with 165 nodes, and an 200 × 180 m with 230 nodes. In the static network no node was mobile, whereas 67% of the nodes were mobile in the

Implementation

To validate the simulation results, we implemented DSDP on Tmote Sky [27] hardware running the RETOS operating system [28]. DSDP is developed as a kernel module of RETOS. The service query is supported by basic primitives for the application layer. By implementing DSDP as a kernel module, applications run the service provision framework without modifications.

Fig. 13 illustrates the message types used in DSDP: the neighbor message, the service update message, and the service query message. Node

Conclusion

DSDP is a service provisioning framework designed to support node mobility in wireless sensor networks. DSDP can be used for geographic routing and for various sensor network applications such as data dissemination or aggregations. Unlike previous studies of service provisioning in sensor networks, our work supports mobile networks. DSDP suppresses service updates as much as possible by extending the coverage of minimal regions to a two-hop communication range. With mobility policies, nodes

Acknowledgement

This work was supported by the National Research Foundation (NRF) of Korea (Grant Nos. 2009-0079878, 2009-0066418).

Choonha Hwang is a research engineer in Xener Systems, Inc. His research interests include wireless sensor networks, embedded operating systems, and wireless communication protocols. He received his B.Sc. and M.Sc. degrees in computer science from Yonsei University, Korea.

References (28)

  • C. Campo et al.

    PDP: a lightweight discovery protocol for local-scope interactions in wireless ad hoc networks

    Computer Networks

    (2006)
  • S. Helal, N. Desai, V. Verma, C. Lee, Konark – a service discovery and delivery protocol for ad-hoc networks, in:...
  • U.C. Kozat, L. Tassiulas, Network layer support for service discovery in mobile ad hoc networks, in: IEEE INFOCOM 2003,...
  • O.V. Ratsimor, D.C. Chakraborty, S. Tolia, Allia: alliance-based service discovery for ad-hoc environments, in:...
  • M. Storey et al.

    MARE: resource discovery and configuration in ad hoc networks

    Mobile Networks and Applications

    (2002)
  • G. Karumanchi, S. Muralidharan, R. Prakash, Information dissemination in partitionable mobile ad hoc networks, in:...
  • J. Ortiz, C.R. Baker, D. Moon, R. Fonseca, I. Stoica, Beacon location service: a location service for point-to-point...
  • R. Fonseca, S. Ratnasamy, J. Zhao, C.T. Ee, D. Culler, S. Shenker, I. Stoica, Beacon vector routing: scalable...
  • S. Ratnasamy, B. Karp, L. Yin, F. Yu, D. Estrin, R. Govindan, S. Shenker, GHT: a geographic hash table for data-centric...
  • X. Xiang, X. Wang, A scalable geographic service provision framework for mobile ad hoc networks, in: PerCom’07, White...
  • P. Juang et al.

    Energy-efficient computing for wildlife tracking: design tradeoffs and early experiences with ZebraNet

    ACM Sigplan Notices

    (2002)
  • M. Kaur, S. Bhatt, L. Schwiebert, G.G. Richard, An efficient protocol for service discovery in wireless sensor...
  • A.N. Mian et al.

    A survey of service discovery protocols in multihop mobile ad hoc networks

    IEEE Pervasive Computing

    (2009)
  • Universal Description Discovery and Integration Platform. Available from:...
  • Cited by (8)

    View all citing articles on Scopus

    Choonha Hwang is a research engineer in Xener Systems, Inc. His research interests include wireless sensor networks, embedded operating systems, and wireless communication protocols. He received his B.Sc. and M.Sc. degrees in computer science from Yonsei University, Korea.

    Elmurod Talipov is currently a Ph.D. student at the Department of Computer Science, Yonsei University, Korea. He received his M.Sc. degree in computer science from Yeungnam University in 2007. His research interests include ad hoc and wireless sensor networks, delay tolerant networks, mobile computing, and wireless communication systems.

    Hojung Cha is currently a professor in computer science at Yonsei University, Seoul, Korea. His research interests include wireless and mobile systems, embedded operating systems and sensor network systems. He received his B.S. and M.S. in computer engineering from Seoul National University, Korea, in 1985 and 1987, respectively. He received his Ph.D. in computer science from the University of Manchester, England, in 1991.

    View full text