Available online at www.sciencedirect.com
Computer Communications 31 (2008) 896–914 www.elsevier.com/locate/comcom
Efficient data propagation strategies in wireless sensor networks using a single mobile sink Ioannis Chatzigiannakis a,b, Athanasios Kinalis a,b,*, Sotiris Nikoletseas a,b b
a Computer Technology Institute (CTI), N. Kazantzaki Str., Rio, Patras, 26500, Greece Computer Engineering and Informatics Department, University of Patras, Rio, Patras, 26500, Greece
Available online 25 December 2007
Abstract Data collection is usually performed in wireless sensor networks by the sensors relaying data towards a static control center (sink). Motivated by important applications (mostly related to ambient intelligence and remote monitoring) and as a first step towards introducing mobility, we propose the basic idea of having a sink moving in the network area and collecting data from sensors. We propose four characteristic mobility patterns for the sink that we combine with different data collection strategies. Through a detailed simulation study, we evaluate several important performance properties of each approach. Our findings demonstrate that by taking advantage of the sinks mobility and shifting work from sensors to the powerful sink, we can significantly reduce the energy spent in relaying traffic and thus greatly extend the lifetime of the network. Ó 2007 Elsevier B.V. All rights reserved. Keywords: Data propagation; Mobility; Performance evaluation; Wireless sensor networks; Distributed computing
1. Introduction Wireless Sensor Networks are visioned as very large collections of smart sensor nodes that form ad hoc distributed sensing and data propagation networks that collect quite detailed information about the ambient environment. In a usual scenario, these networks are largely deployed in areas of interest for fine grained monitoring in different classes of applications [1]. The sensor devices are battery powered; energy is the most precious resource of a wireless sensor network, since replacing the battery of the nodes in large scale deployments is infeasible. The collected data is disseminated to a static control point – data sink in the network, using node to node – multi-hop data propagation, [2,3]. However, in this setting sensor devices consume significant amounts of energy due to the execution of a rout*
Corresponding author. Address: Computer Engineering and Informatics Department, University of Patras, Rio, Patras, 26500, Greece. Tel.: +30 2610 960459; fax: +30 2610 969002. E-mail addresses:
[email protected] (I. Chatzigiannakis), kinalis@ceid. upatras.gr (A. Kinalis),
[email protected] (S. Nikoletseas). 0140-3664/$ - see front matter Ó 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.comcom.2007.12.011
ing protocol which also increases implementation complexity. Also, a point of failure emerges in the area near the control center where nodes relay the data from other nodes that are farther away. 1.1. Mobility Additionally, a new category of important sensor networks applications emerges where motion is a fundamental characteristic of the examined system. In such applications sensors are attached to vehicles, animals or people that move around large geographic areas. Data exchange between individual sensors and infrastructure nodes will drive applications such as traffic and wild life monitoring, smart houses and hospitals and pollution control. Clearly, the usual approach of having a statically placed control center is unable to operate efficiently in such scenarios. Recently, a new approach has been developed that shifts the burden from the sensor nodes to the sink. The main idea is that the sink is mobile, has significant and easily replenishable energy reserves and moves inside the area the sensor network is deployed. When moving inside the
I. Chatzigiannakis et al. / Computer Communications 31 (2008) 896–914
network area the sink is constantly in close proximity to a (usually small) subset of the sensor devices and can acquire the data collected by these nodes at very low energy cost. By travelling in the whole network area, the control center is capable of collecting all the available data. The mobility assumption may be especially useful in particular application classes, e.g., in emergency preparedness [4], a set of nodes called actors aggregate and evaluate the collected data in order to detect and react to emergency situations. In this setting mobile elements can move in close proximity to an area and better assess an emergency situation. Traversing the network area in a timely and efficient way is critical since failure to visit some areas of the network will result in data loss, while infrequently visiting some regions will result in long delivery delays. Also, routing and localization problems in the case of mobile sinks become more difficult to cope with, efficient solutions in the state of the art become inefficient or even inoperable in networks with mobility. Thus new techniques and algorithms need to be developed. Despite the apparent difficulties, this data collection paradigm has many attractive properties (see also [5]). The major advantage of having a mobile sink (or more) is an increase in system lifetime. A mobile agent that moves closer to the nodes can help conserve energy since data is transmitted over fewer hops thus reducing the number of transmitted packets. The extra energy spent for the operation and movement of the sink does not affect overall sensor network lifetime since the mobile sink is considered an external to the network factor, in fact it could be a man navigated vehicle or an unmanned robot that periodically returns to a support center in order to recharge itself. Another important advantage is that sparse and disconnected networks can be better handled, since full connectivity of the network nodes is not required. A mobile sink allows to monitor a region with fewer sensor devices thus decreasing the operational cost of the network. Also, the sensor devices can reduce their transmission range to the lowest value required to reach the mobile infrastructure, further decreasing the energy consumption. Additionally, the mobile sink can navigate through or bypass problematic regions where sensor devices cannot operate, such as small lakes, or obstacles, e.g., large boulders, that block the propagation path. Conventional multi-hop protocols either fail or spend too many resources in order to overcome such obstacles, thus the monitoring of much harsher environments is possible. Moreover, increased throughput and data fidelity can be achieved with the use of a mobile sink. By reducing the number of hops, the probability of transmission errors and collisions also reduces. Also, security can be enhanced as the data does not traverse multiple hops across potentially compromised nodes, while an adversary sniffing the data packets from an area will receive only the information regarding that area. A very nice and complete categorization and discussion of several different mobility models
897
and their relevance to different application types and real world scenarios can be found in [6]. 1.2. Our contribution In this work, we propose and investigate sink mobility as a method for efficient and robust data delivery in wireless sensor networks. Our work is one of the first few attempts in the relevant state of the art that introduces mobility of the sink. We propose four mobility patterns for the sink, mostly randomized (such as the simple random walk, biased random walks and walks on spanning subgraphs) as well as predictable mobility (moving on a straight line or cycle). These patterns assume and exploit different degrees of freedom, simplicity and network knowledge. To get data from the sensors, the sink movement is combined with three data collection strategies: a passive, a multi-hop and a limited multi-hop. Each approach has different advantages and disadvantages and achieves different trade-offs, mostly between energy dissipation and time efficiency. We investigate several important performance properties of our protocols through a detailed, large scale simulation evaluation considering non-uniform positioning of sensor nodes. We conduct an experimental rather than theoretical analysis, because of the added complexity due to the number of the proposed protocols and the diversity between them as well as the complications that arise due to the non-uniform evaluation settings. Our findings demonstrate that having a mobile sink can greatly extend the lifetime of the network. In fact, even by having very limited sink mobility, the overall success rate can be improved by more than 50% and the energy dissipation is reduced by more than 30%. If the sink can be fully mobile (i.e., be able to visit all the areas of the network), we can further reduce the energy consumption and achieve very high success rates, very close to 100%. Our work differentiates from the state of the art and extends it, in at least the following ways: (a) We assume very weak models, our protocols use only locally obtainable knowledge and impose minimum overhead to the sensor nodes. (b) We present the only work, to the best of our knowledge, that proposes and evaluates comparatively several different methods to exploit mobility. The examined methods span across the gamut of proposed mobility categories and data propagation techniques. (c) We propose elaborate and novel randomized mobility techniques that adapt to network conditions. (d) We perform extensive evaluation of our protocols in realistic settings and present in detail qualitative results in comparison to the static case. Moreover, we feel that our work is a first step towards introducing mobility in wireless sensor networks, e.g., having many/multiple sinks collecting data in a network composed of mobile sensors (or combinations of mobile and static sensors). Though efficient and important in practice, mobility introduces complications and new challenges for protocol design that should be investigated in future
898
I. Chatzigiannakis et al. / Computer Communications 31 (2008) 896–914
research. A preliminary version of this work has been presented in [7]. 1.3. Related work and comparison Several protocols that implement energy efficient information collection, have been proposed for static wireless sensor networks. Probabilistic forwarding schemes (like PFR, [8]) perform redundant optimized multi-path transmissions to combine energy efficiency and fault-tolerance. Such protocols, although well suitable for deployment in sparse networks, tend to spend a lot of energy in the case of high densities. Protocols specifically aimed at achieving energy efficiency at a lower level, have been proposed in [9]. The protocols alternate the operation of the sensors between a low power sleep mode and normal operation. The protocols continually monitor local network conditions, and adapt the length of each period accordingly. Thus, great energy savings are achieved even on highly dynamic networks with incremental and heterogeneous node deployment. Also, in [10] an adaptive randomized algorithm is proposed that balances the energy consumption of the sensor devices. As we will shortly see, mobility of the sink also tends to balance energy consumption. In the context of mobile ad-hoc networks, mobility and routing have been studied extensively. An introduction in mobile ad-hoc networks as well as detailed descriptions of several routing protocols can be found in [11]. A survey of proposed mobility models can be found in [12]. The results in this area can not be directly applied in wireless sensor networks since, sensor nodes have different, more constrained capabilities. Also, the communication patterns are much different, in wireless sensor networks usually many nodes transmit data to one whereas in mobile adhoc networks many nodes communicate with each other. The usual scenario, examined in most research papers in the context of mobile ad-hoc networks, is that nodes move in an uncontrollable fashion and a solution is proposed that establishes and maintains routing paths in a resilient, towards mobility, manner. For example, in [13] an algorithm for constructing a routing backbone out of dedicated powerful or under-utilized nodes is presented. The resulting network uses fewer links and is connected as long as the original network is connected. Another attempt to counter mobility is presented in [14] where the authors propose an algorithm that maintains routes by observing the relative movement between nodes. Here we examine in unison the issues of mobility and routing, proposing joint methods for directly improving network performance exploiting mobility. Recently, several applications that motivate mobility in wireless sensor networks appeared, [15] presents a case study of applying peer-to-peer techniques in mobile sensor networks designed for wildlife position tracking for biology research. In [16], a data sink was mounted on a public transportation bus. The sensor nodes learn the times at
which they have connectivity with the bus, and wake up accordingly to transfer their data. In [17], a three-tier architecture is proposed that exploits the random motion of mobile entities, such as humans or animals, in order to collect information from the sensors and relay that information to a central control center. In [5], the authors discuss several interesting issues of mobility and implement and evaluate a small sensor network with one mobile entity that moves back and forth on a straight line. This idea is further extended in [18] where the possibility of having multiple mobile sinks is examined. The trajectories of the mobile entities are fixed by the network operator; sinks move in parallel following linear trajectories. An algorithm to load balance the data collection process by assigning sensor nodes to sinks is proposed. Note that the algorithm operates only under the assumption of full coverage of the network by the mobile entities. Here, we examine various types of random movement as well as predefined trajectories of a single sink under weaker assumptions. Luo et al. [19], nicely investigate how to minimize energy consumption in a wireless sensor network with a mobile sink while also considering multi-hop propagation effects. They conduct an analysis assuming a completely uniform network, both in terms of node distribution and generation of events. They estimate that the overall energy consumption is minimized when the trajectory of the mobile sink is a cycle and calculate its optimal positioning and radius. Optimization of the data propagation process using mobility is examined in [20]. Under the assumption that all sensor nodes can move, the authors propose an algorithm for rearranging the position of nodes in a multi-hop propagation path in order to minimize the energy consumption along this path. In our work, we also focus on increasing network performance by proposing solutions that couple sink movement with efficient data propagation. Additionally, we investigate not only multi-hop but additional propagation methods, like single hop and limited multi-hop. We also examine the behavior of our protocols in more realistic settings assuming networks with non-uniform node distribution. In [21], the authors investigate the cover time of a graph when performing a random walk that uses deterministic choice. Their proposed scheme selects randomly two nodes and then chooses deterministically to move to the least frequently visited of the two. Note that here, we also propose random walks that perform choices based on various network criteria. We also favor less frequently visited areas, but our method is purely randomized and was developed independently. In [22], the authors propose a routing scheme suitable for networks with one mobile sink. The sink visits certain anchor points in the network area and remains still while collecting data at each one of them. The sink samples the global power consumption of all nodes while stationed in an anchor point. It uses this data to create power consumption profiles and calculate the optimal resting time at each
I. Chatzigiannakis et al. / Computer Communications 31 (2008) 896–914
anchor point. Another approach that optimizes the sink’s trajectory is presented in [23]. The authors assume a mobile sink that initially follows a linear trajectory and collects network information from the sensors. The sink uses this information to break its trajectory into separate line segments that are closer to the network nodes, thus minimizing the data propagation cost. Collecting network knowledge incurs a significant overhead on the sensor nodes. In contrast to these approaches our protocols don’t require global network knowledge and can adapt to changing network conditions using only local observation. 2. The model Sensor networks are comprised of a vast number of ultra-small homogeneous sensor devices, which we also refer to as sensors (see also [2]). Each sensor is a fullyautonomous computing and communication device, characterized mainly by its available power supply (battery) and the energy cost of data transmission and by the (limited) processing capabilities and memory. Sensors (in our model here) do not move. The network area A is a flat square region of size D D; this assumption can be easily relaxed to include general network areas of arbitrary shapes. The positions of sensor nodes within the network area are random and in the general case follow a uniform distribution. Let n be the number of sensors spread in the network area and let d be the density of sensors in that area (usually measured in numbers of sensors=m2 ). However, in several scenarios the network implementors are expected to deploy more sensors in areas where fine grained monitoring is required. We model such scenarios by assuming P n ‘‘pockets”, i.e., regions in the network area with high sensor node density. For the sake of simplicity, each pocket is a circular area of radius rP , pockets do not overlap and the density of sensors in these areas is d P . For the rest of the network area, the density is d n . We denote AP ¼ pr2P P n the area occupied by the pockets. Let nP the number of sensors contained in pockets, clearly nP ¼ d P AP . Likewise, the number of sensors contained in the rest of the network is nn ¼ d n ðA AP Þ. Then the total number of sensors is given by n ¼ nP þ nn . Sensor devices are equipped with a set of hardware monitors that can measure several environmental conditions. Each device has a broadcast (digital radio) beacon mode of fixed transmission range R, and is powered by a battery. Also a sensor is equipped with a general purpose storage memory (e.g., FLASH) of size C. Let E i be the available energy supplies of sensor i at a given time instance. At any given time, each sensor can be in one of three different modes, regarding the energy consumption: (a) transmission of a message, (b) reception of a message and (c) sensing of events. In our model, for the case of transmitting and receiving a message, we assume that the radio module dissipates an amount of energy proportional to the message’s size. To
899
transmit a k-bit message, the radio expends ET ðkÞ ¼ trans k and to receive a k-bit message, the radio expends ER ðkÞ ¼ recv k where trans ; recv are constants that depend on the radio module and the transmission range of the sensors. For the idle state, we assume that the energy consumed for the circuitry is constant for each time unit and equals Eelec . Overall, there are three different types of energy dissipation: (a) ET , the energy dissipation for transmission, (b) ER , the energy dissipation for receiving and (c) Eidle , the energy dissipation for idle state. For the idle state, we assume that the energy consumed for the circuitry is constant for each time unit and equals Eelec (the time unit is 1 s). We note that in our simulations we explicitly measure the above energy costs. There is a special node within the network area, which we call the sink S, that represents a control center where data should be collected. An important modelling assumption that differentiates our approach from most standard models in the state of the art is the mobility of the sink. The sink is not resource constrained, i.e., it is assumed to be powerful in terms of computing, memory and energy supplies. The sink can calculate accurately it’s position (e.g., by using navigational equipment, such as GPS) and is aware of the dimensions of the network area. The sink moves following a high level mobility function which we symbolize by M. If pt is the position of the sink in a given moment then Mðpt Þ will return a new position ptþ1 towards which the sink should move. This defines a trajectory for the sink as a series of points p0 ; p1 ¼ Mðp0 Þ; p2 ¼ Mðp1 Þ; . . . ; pt ¼ Mðpt1 Þ. Also, the function M defines the speed st ¼ Mðst1 Þ by which the sink moves from position pt1 to position pt . The speed is bounded by a maximum value which is depended on the scenario and models the mobility capacity of the sink; we call this limit smax . A valid definition of M returns positions that are within the network area and s 6 smax . The mobility function can be invoked at anytime even before reaching the designated point. The actual mechanism that moves the mobile entity from position pt1 to position pt is beyond the scope of this paper since it can be a human driver or an automated navigation system. However, in order to simplify our model we assume that all changes in speed and direction can be done instantly. Finally, we assume that a specific, high-level, application is executed by the sensors that form the network. Applications for wireless sensor networks fall in three major categories [24]: (i) Periodic Sensing (the sensors are always monitoring the physical environment and continuously report the recorded measurements to the control center S), (ii) Event driven (to reduce energy consumption, sensors operate in a silent monitoring state and are ‘‘programmed” to notify about events) and (iii) Query based (queries can be propagated to the sensors arbitrarily by the control center S, according to the application and/or user’s will). We model in an abstract way the kind of high-level application by the message generation rate in each sensor ki :
900
I. Chatzigiannakis et al. / Computer Communications 31 (2008) 896–914
3. Protocols for data collection with sink mobility There are many different approaches when considering the mobility pattern that the mobile sink should follow. Depending on the application scenario and the network size and conditions, different approaches may yield diverse results and affect drastically the achieved network performance. A coarse-grained categorization of mobility patterns will identify the following different approaches (see also [6]): Random mobility. The movement of the sink is done in a random manner regarding the position and the speed of movement. The main characteristic of this pattern is simplicity and unpredictability of the future position of the sink. Different random mobility methods define various degrees of freedom to the movement such as allowing only predefined locations or moving only in certain directions (i.e., north, south, etc). Predictable mobility. The movement of the mobile element follows a certain pattern that can be computed. Such movements may be periodic movements along a predefined trajectory. In this case sensor nodes can learn, with some precision, the time the sink will be in their vicinity and optimize the sensing task and data collection process. Controlled mobility. The mobile element can vary its movement in a deterministic way in order to achieve better results. The movement variation may include trajectory as well as speed modifications. This movement pattern is unpredictable but the deterministic variation function can guarantee certain conditions, such as visiting a certain number of nodes within a fixed time. The differences between the categories of mobility suggest that different data collection protocols are appropriate in each case. There is certainly a large space of options for protocol design, depending on which mobility strategy and which data collection mechanism are used. We partially explore this space, by proposing and investigating below a few characteristic protocols composed of a mobility pattern and an appropriate data collection process, along with a summary discussion of their potential. 3.1. P1: random walk and passive data collection The simplest of all possible mobility patterns is the random walk, where the mobile sink can move chaotically towards all directions at varying speeds. We define Mrandom as a function that implements random walks in our scenarios. At each invocation Mrandom selects a random uniform angle in ½p; p radians. This angle defines the deviation from the mobile sink’s current direction. In our version of random walk the speed of the movement is constant srandom and predefined by the network implementors. To determine the new position, Mrandom selects a uniform random distance d 2 ð0; d max Þ which is the distance to tra-
vel along the newly defined direction. If the new position falls outside the network area, Mrandom crops the position to fall on the boundary of the area. This is the simplest possible movement; no network knowledge at all is assumed. Furthermore, this method is very robust, since it guarantees visiting all sensors in the network and thus collecting data even from disconnected areas of few/faulty sensors or under the presence of obstacles. However in some network structures it may become inefficient, mostly with respect to latency. Data is collected in a passive manner, sensors cache all recorded data. Periodically a beacon message is transmitted from the sink. Each sensor node that receives a beacon attempts to acquire the medium and transmit the cached data to the sink. Transmitted data is then removed from a sensor’s cache to free memory for new recordings. This method may lead to many collisions, when the sink visits a dense area. Thus, an appropriate MAC layer protocol with an efficient backoff function is essential for the proper deployment of this protocol. Clearly, this approach minimizes energy consumption since only a single transmission per sensed event is performed; however time efficiency may drop due to long intervals between visits to the sensors. 3.2. P2: partial random walk with limited multi-hop data propagation Here, another form of random walk is performed by using a set of predefined areas and random transitions between the areas according to their connectivity. More specifically, we propose the mobility function Mgraph that uses a transition graph. The graph is constructed during a graph formation phase that is executed by the sink during the network initialization. The network area is partitioned in j j equal square regions, the center of each region is considered as a vertex in a graph that is connected with unidirectional edges only to the 4 vertices corresponding to adjacent regions. Thus, a lattice graph Go is created which is overlayed over the network area as depicted in Fig. 1. Initially the mobile sink is positioned on or near one of the nodes of Go , Mgraph calculates the next position by selecting uniformly randomly one of the neighbors of the current vertex in the graph. The speed of the movement is constant and predefined by the network implementors. In this way, the area covered by the sink in a single move covers the maximum number of sensor nodes and thus fewer distance needs to be traveled by the mobile element. Full coverage of the network is achieved when pffiffiffi j ¼ dD 2=Re since every sensor node in a D=j D=j area can communicate with the center of the area and thus the sink. In this partial walk setting, a data collection protocol may need to communicate with nodes outside of the immediate reach of the sink. The protocol presented here forms propagation trees with the sink as root. The sink periodically broadcasts a beacon message, which carries a hop counter hc (initially hc ¼ 0) and a time to live counter ttl.
I. Chatzigiannakis et al. / Computer Communications 31 (2008) 896–914
901
to P1 and leads to improved time efficiency. On the other hand, it is also more expensive in terms of communication and computational cost on the sensor devices. 3.3. P3: biased random walk with passive data collection
Fig. 1. Example of the overlay graph Go and possible sink movement.
Each sensor node maintains a hop distance hd from the sink (initially hd ¼ 1), a timestamp tu indicating the last time hd was updated and the network address of a parent sensor node (initially null). When a beacon is received at time tb , the sensor decides if it needs to update his parent node according to the following rules: (1) If hc < hd , the node sets the sender of the message as it’s parent, sets hd ¼ hc and tu ¼ tb . Then the sensor node decides whether to further propagate the beacon or not. It increases the hop counter of the beacon hc ¼ hc þ 1, reduces ttl ¼ ttl 1 and if ttl > 0 broadcasts the beacon message to it’s neighbors. (2) If hc P hd and tb > tu þ tthres , then the node sets the sender of the message as it’s parent, sets hd ¼ hc and tu ¼ tb . Note that the value tthres is set by the network operator and expresses the time a propagation tree is considered valid.Then the sensor node decides whether to further propagate the beacon or not as above. (3) If hc P hd and tb 6 tu þ tthres , the node simply discards the beacon. This process creates a number of propagation trees within the network with the roots of these trees being one hop away from the sink. Sensor nodes that belong to a propagation tree may begin immediately forwarding their data to the sink. The depth of these trees is determined by the ttl value which is an operational parameter of the mobile sink. In this way, the network operator can tune the trade-off between reduced delay and increased energy consumption. The value of ttl should be set according to the selected j, when the sink is positioned in the center of a square the maximum hop distance of a sensor node is pffiffi dD=jR 2e. Note that when ttl ¼ 1 this method is equivalent to the passive collection described earlier. As the sink moves, whole propagation trees may become disconnected. When a node with hd ¼ 0 can no longer communicate to the sink, it simply caches all data, both generated and relayed. Then it waits to receive another beacon message with hc ¼ 0 to begin the propagation process again or after tthres time the node can participate to a new tree. This protocol assumes and uses more knowledge of the network, it can accelerate times to visit network nodes, reduces the distance traveled by the sink when compared
The idea of using a logical graph can be extended in a way that certain areas of the network are favored (i.e., more frequently visited) by the sink in order to improve the data collection process or to overcome problems that arise from the network topology. More specifically, the mobility pattern Mbias forms a logical overlay graph (as described in 3.2, see also Fig. 2) Go ¼ GðV ; EÞ with pffiffiffi j ¼ dD 2=Re; it associates two counters cv and d v for every vertex v, initially cv ¼ d v ¼ 0 8v 2 V . These counters store data about the network conditions for each area visited. When the mobile sink enters the area corresponding to vertex u it increases the associated counter cu by 1. Also, while the sink remains in the area corresponding to vertex u, it increases the counter d u according to the number of new nodes, i.e., sensor nodes that has received a message from for the first time. Thus, the frequency of visits and the density of each area can be estimated by the sink. The selection of the next area to visit is done in a biased random manner depending on these two variables as described below. 3.3.1. Frequency biased If the mobile element is currently P on vertex u of degree degu , then we define cneigh ðuÞ ¼ v cv for all v : ðu; vÞ 2 E. Then the probability pðcÞv of visiting a neighboring vertex 1cv =cneigh ðuÞ v is calculated as pðcÞv ¼ deg when cneigh 6¼ 0. When u 1 cneigh ¼ 0 we have pðcÞv ¼ 1=degu . Thus, less frequently visited areas are more likely to be visited when the sink is located at a nearby area.
u1
u2
u4
u3
Fig. 2. Example of the overlay graph Go and the logical partitioning (dotted lines) of the network in regions.
902
I. Chatzigiannakis et al. / Computer Communications 31 (2008) 896–914
3.3.2. Density biased If the mobile element is currently P on vertex u of degree degu , then we define d neigh ðuÞ ¼ v d v for all v : ðu; vÞ 2 E. Then the probability pðdÞv of visiting a neighboring vertex 1þd v =d neigh ðuÞ when d neigh 6¼ 0. When v is calculated as pðdÞv ¼ deg u þ1 d neigh ¼ 0 we have pðdÞv ¼ 1=degu . Thus, areas with many sensor devices are more likely to be visited when the sink is located at a nearby area. After calculating pðcÞv and pðdÞv the final probability of transition to the area corresponding to vertex v is given by pv ¼ a pðcÞv þ b pðdÞv where a, b are positive real numbers such that a þ b ¼ 1. An example of such probability calculation can be seen
in Table 1. Data is collected in a passive manner as described in 3.1. This protocol uses knowledge collected by the sink in order to speed up the coverage of new areas (when
Table 1 Example of probability calculation for transition to the adjacent areas, when a ¼ b ¼ 0:5 Node
cv
dv
pðcÞv
pðdÞv
pv
u1 u2 u3 u4
3 5 2 2
10 8 6 10
0.25 0.1944 0.2778 0.2778
0.2588 0.2472 0.2352 0.2588
0.2544 0.2208 0.2565 0.2683
100 97.5 95 90
Success Rate %
85 80
60
40 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology s=4
s=8
Fig. 3. Success rate of protocol P1 for topologies with 0, 2 and 4 pockets with
nP nn
2 f1=2; 2=3g.
4.5
Energy Dissipation (Joules)
4.4 4.3 4.2 4.1 4 3.9 3.8 3.75 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology s=4
s=8
Fig. 4. Energy dissipation of protocol P1 for topologies with 0, 2 and 4 pockets with
nP nn
2 f1=2; 2=3g.
I. Chatzigiannakis et al. / Computer Communications 31 (2008) 896–914
903
1700 1600
Average Delivery Delay (seconds)
1500 1400 1300 1200 1100 1000 900 800 700 600 500 400 300 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology s=4
s=8
Fig. 5. Average delay of protocol P1 for topologies with 0, 2 and 4 pockets with
nP nn
2 f1=2; 2=3g.
100 97.5 95 90 85
Coverage %
80
60
40 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology s=4
s=8
Fig. 6. Covered set (d) of protocol P1 for topologies with 0, 2 and 4 pockets with
a > b) or increase data delivery in areas with many nodes (when a < b). Overall, when compared to P1 we expect to achieve faster network coverage, higher delivery rates and lower latency with an increase in computational overhead at the sink. 3.4. P4: deterministic walk with multi-hop data propagation Here, we use a simple form of controlled mobility where the mobile entity moves on a predefined trajectory. We examine the cases where the trajectory is a line (we call the mobility function Mline ) or a circle ðMcircle Þ that is fully contained in the network area. The trajectory is characterized by its length l. In particular, the linear trajectory con-
nP nn
2 f1=2; 2=3g.
sists of a horizontal or vertical line segment passing through the center of the network. The sink moves from one edge of the line to other and returns along the same path. For the case of the circular trajectory the circle is centered at the center of the network and its radius is defined as r ¼ 2pl (for comparison fairness with the line case). Initially, the sink is positioned on the circumference of the circle and continues along this path. Since the mobile sink covers only a small network area it is necessary to collect data with a multi-hop data propagation protocol. We use a tree formation protocol similar to the one presented in 3.2, without the timeout and time-to-live mechanism. In this way paths are created according to minimum hop distance and span in the whole network area. Note that we
904
I. Chatzigiannakis et al. / Computer Communications 31 (2008) 896–914 100 97.5 95 90
Success Rate %
85 80
60
40 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology hops=2, s=4
hops=2, s=8
hops=4, s=4
hops=4, s=8
Fig. 7. Success rate of protocol P2 for different partitioning and hop limit, for topologies with 0, 2 and 4 pockets with
nP nn
2 f1=2; 2=3g.
4.5
Energy Dissipation (Joules)
4.4 4.3 4.2 4.1 4 3.9 3.8 3.75 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology hops=2, s=4
hops=2, s=8
hops=4, s=4
hops=4, s=8
Fig. 8. Energy dissipation of protocol P2 for different partitioning and hop limit, for topologies with 0, 2 and 4 pockets with
disable the timeout threshold that invalidates a tree since the sink performs a periodic movement. When the root of a propagation tree disconnects from the sink, it simply caches all data, both generated and relayed, and waits to hear another beacon message with hc ¼ 0 to resume the propagation process. Effectively this means that the node waits for the sink to be within transmission range again. Note that here we assume that nodes that received a beacon with hc ¼ 0, are within a single hop distance from the trajectory of the sink. Since the movement of the sink is deterministic and periodic, and the radio channel maintains the same properties over time, these nodes will communicate directly to the sink everytime the sink passes near them. Another approach is to invalidate
nP nn
2 f1=2; 2=3g.
that tree and let the tree formation process reassign the nodes to a new tree, which will happen only if the sensor network is connected. However, this approach has a maximum hop distance equal to that of the static sink approach. The kind of mobility presented here models situations where the mobile sink cannot execute complex movements, movement on the terrain is possible only in certain areas or the mobile sink does not have enough energy to traverse the whole network area. The deployment of this protocol imposes a high cost on the sensor devices that perform tree formation and multi-hop propagation, however the delivery latency is lower than any of the three previous protocols.
I. Chatzigiannakis et al. / Computer Communications 31 (2008) 896–914
905
1700 1600
Average Delivery Delay (seconds)
1500 1400 1300 1200 1100 1000 900 800 700 600 500 400 300 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology hops=2, s=4
hops=2, s=8
hops=4, s=4
hops=4, s=8
Fig. 9. Average delay of protocol P2 for different partitioning and hop limit, for topologies with 0, 2 and 4 pockets with
nP nn
2 f1=2; 2=3g.
100 97.5 95 90 85
Coverage %
80
60
40 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology hops=2, s=4
hops=2, s=8
hops=4, s=4
hops=4, s=8
Fig. 10. Covered set of protocol P2 for different partitioning and hop limit, for topologies with 0, 2 and 4 pockets with
Furthermore, the selection of the trajectory length l introduces a trade-off between the cost at the sink and the cost at the sensors, e.g., increasing l makes the sink move more and get closer to the sensors, thus reducing energy dissipation. Finally, it is interesting to investigate which shape (line or cycle) is best with respect to cost and performance (for equal shape lengths). 4. Experimental evaluation We implement and evaluate the above discussed protocols in the network simulator platform ns 2 version 2.26. We also included for comparison a well known static multi-hop data propagation paradigm, the Directed
nP nn
2 f1=2; 2=3g.
Diffusion (see [3]). We considered different simulation setups for various network sizes, number of nodes and mobility parameters. We here present the results for the set of experiments that consider several network topologies. In particular, the size of the network area is 200m 200m and 300 sensor nodes are deployed. We consider a uniform deployment and deployment of nodes in P n 2 f2; 4g pockets of radius rP ¼ 2R whose positions are identical to the placement of dots on the facet of a dice. The ratio of the number of nodes in pockets over the number of nodes in the rest of the network is nP 2 f1=2; 2=3g. The transmission range of the sensors nn and the sink S is set to R ¼ 15m and the speed of the movement of S is set to 4 and 8 m/s.
906
I. Chatzigiannakis et al. / Computer Communications 31 (2008) 896–914 100 97.5 95 90
Success Rate %
85 80
60
40 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology α=1, β=0, s=4 α=1, β=0, s=8
α=0.5, β=0.5, s=4 α=0.5, β=0.5, s=8
α=0, β=1, s=4 α=0, β=1, s=8
Fig. 11. Success rate of protocol P3 for different values of a; b for topologies with 0, 2 and 4 pockets with
np nn
2 f1=2; 2=3g.
4.5
Energy Dissipation (Joules)
4.4 4.3 4.2 4.1 4 3.9 3.8 3.75 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology α=1, β=0, s=4 α=1, β=0, s=8
α=0.5, β=0.5, s=4 α=0.5, β=0.5, s=8
α=0, β=1, s=4 α=0, β=1, s=8
Fig. 12. Energy dissipation of protocol P3 for different values of a; b for topologies with 0, 2 and 4 pockets with
We evaluate our protocols under different configurations, more specifically for Mrandom of P1, we set d max ¼ 50m. For P2 we examine two settings where j ¼ 10, ttl ¼ 2, tthres ¼ 10 ss and j ¼ 5, ttl ¼ 4, tthres ¼ 10 s. For parameterizing P3 we set j ¼ 20 and present the results for various values of a and b, setting a ¼ 1 and b ¼ 0, a ¼ 0:5 and b ¼ 0:5 and a ¼ 0 and b ¼ 1. For P4 we try different trajectory lengths, we set l 2 fD; D=2; D=3g. For protocols P1, P2, P3 the mobile sink is initially positioned at point ðD=2; 0Þ, while for P4 the sink is initially located at a point on its trajectory. In every tested setting and for all protocols the sink transmits a beacon message every 0.5 s. The initial energy reserves of the nodes are high enough to ensure that nodes remain operational for all
np nn
2 f1=2; 2=3g.
the duration of the simulation, also here we do not consider the possibility of node failures that would make harder to investigate the behavior of our protocols. The values of trans , recv and Eidle were set to match as close as possible the specifications of the mica mote platform [25]. We assume that a high level periodic monitor application is executed by the sensor devices, the application is triggered at the beginning of the simulation and registers data about the network region. The data is generated at random times in packets of 36 bytes and at an average rate of 1 message every 10 s, the size of a beacon message is 24 bytes. Each sensor node has a cache of 256 kbyte. Each sensor transmits 100 messages before the monitored phenomenon ends, meaning that the data generation
I. Chatzigiannakis et al. / Computer Communications 31 (2008) 896–914
907
1700
Average Delivery Delay (seconds)
1600 1500 1400 1300 1200 1100 1000 900 800 700 600 500 400 300 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology α=1, β=0, s=4 α=1, β=0, s=8
α=0.5, β=0.5, s=4 α=0.5, β=0.5, s=8
α=0, β=1, s=4 α=0, β=1, s=8
Fig. 13. Average delay of protocol P3 for different values of a; b for topologies with 0, 2 and 4 pockets with
nP nn
2 f1=2; 2=3g.
100 97.5 95 90 85
Coverage %
80
60
40 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology α=1, β=0, s=4 α=1, β=0, s=8
α=0.5, β=0.5, s=4 α=0.5, β=0.5, s=8
α=0, β=1, s=4 α=0, β=1, s=8
Fig. 14. Covered set of protocol P3 for different values of a; b for topologies with 0, 2 and 4 pockets with
phase lasts for about 1000 s. During the data generation phase the mobile sink is collecting data and another 4000 s of simulation time are given in order to collect all the data, leading to 5000 s of total simulation time. Conducting these experiments, we measure several metrics that depict the behavior of the protocols. We call success rate the percentage of data messages that were received to the sink over the total number of generated messages. We measure the energy consumed at the sensor network (i.e., we do not measure the energy consumption of the mobile entity), as an absolute value in Joules. The delivery delay is defined as the time interval between the creation of a message and the time when it is delivered to the sink. Also, another set of metrics which is of particular interest
nP nn
2 f1=2; 2=3g.
for the experiments with mobility functions that traverse the whole network are the covered set, which is the percentage of the visited nodes (i.e., nodes that communicated with the sink directly or indirectly) over the total number of nodes. Figs. 3–6 depict the performance of protocol P1, we first observe that increased mobility speed s affects positively the performance of P1, since for s ¼ 8m=s it achieves success rate close to 95%, only a small increase in energy consumption, lower delay and almost full coverage of the network (95–97%). When s ¼ 4m=s the covered set is about 85% meaning that about 15% of nodes remain unvisited. This explains the lower success rate; it is evident that protocol P1 needs a lot of time to cover the whole network
908
I. Chatzigiannakis et al. / Computer Communications 31 (2008) 896–914 100 97.5 95 90
Success Rate %
85 80
60
40 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology l=200.0, s=4 l=200.0, s=8
l=100.0, s=4 l=100.0, s=8
l=66.66, s=4 l=66.66, s=8
Fig. 15. Success rate for a circle trajectory of variable circumference l, for topologies with 0, 2 and 4 pockets with
nP nn
2 f1=2; 2=3g.
4.5
Energy Dissipation (Joules)
4.4 4.3 4.2 4.1 4 3.9 3.8 3.75 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology l=200.0, s=4 l=200.0, s=8
l=100.0, s=4 l=100.0, s=8
l=66.66, s=4 l=66.66, s=8
Fig. 16. Energy dissipation for a circle trajectory of variable circumference l, for topologies with 0, 2 and 4 pockets with
area depending on the movement speed. This is essential for collecting all of the data but it also has a significant impact in increasing the perceived delivery delay. The protocol seems mostly unaffected from different topology settings which is expected given the nature of random walk. Figs. 7–10 present the experimental results for P2, the first thing we observe is that for a greater number of hops (when ttl ¼ 4) the performance drops significantly in terms of success rate and increased energy consumption. For 2 hops, P2 achieves high success rate (90%), compared to 60–70% for 4 hops and low energy consumption, around 4 J compared to 4.2–4.3 J for 4 hops. The delay in both cases is very low, especially when ttl ¼ 4 which is almost 45% faster than the case of ttl ¼ 2. This is due to the fact
nP nn
2 f1=2; 2=3g.
that when we increase the ttl counter, the constructed routing trees become deeper. The sink essentially collects data from a larger area, thus in shorter time the sink can cover more sensor nodes. In both cases P2 achieves full coverage (>97%) of the network, meaning that almost all nodes forwarded at least one packet to the sink. When s increases the success rate drops and the consumed energy increases, especially when ttl ¼ 4. This is because in this limited multi-hop setting, nodes that are one hop away from the sink relay traffic from the rest of the nodes. When the sink moves rapidly, they do not have enough time to forward all messages and this results in re-transmissions and some packet loss. Note, that even though all nodes got a chance to communicate to the sink (coverage >97%), due to the
I. Chatzigiannakis et al. / Computer Communications 31 (2008) 896–914
909
1700
Average Delivery Delay (seconds)
1600 1500 1400 1300 1200 1100 1000 900 800 700 600 500 400 300 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology l=200.0, s=4 l=200.0, s=8
l=100.0, s=4 l=100.0, s=8
l=66.66, s=4 l=66.66, s=8
Fig. 17. Average delay for a circle trajectory of variable circumference l, for topologies with 0, 2 and 4 pockets with
nP nn
2 f1=2; 2=3g.
100 97.5 95 90 85
Coverage %
80
60
40 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology l=200.0, s=4 l=200.0, s=8
l=100.0, s=4 l=100.0, s=8
l=66.66, s=4 l=66.66, s=8
Fig. 18. Covered set for a circle trajectory of variable circumference l, for topologies with 0, 2 and 4 pockets with
multi-hop propagation several messages were not delivered. This is because updating propagation paths on time is difficult in the case of a mobile sink, thus until a new path is devised, messages follow outdated paths leading to a decrease in success rate. These results dictate that for proper selection of ttl, j and s, good trade-offs between success rate, energy consumption and delay can be achieved. When examining topologies with pockets we notice that the energy consumption and delay increase, especially for 2 pockets and when nnPn ¼ 2=3. In high density areas multihop protocols suffer from many collisions which explains our results. In Figs. 11–14, we examine the performance of our biased protocol P3. We first notice that when the bias is
nP nn
2 f1=2; 2=3g.
based on the frequency of visits only ða ¼ 1:0; b ¼ 0:0Þ, the protocol performs very well. Success rate reaches 99.7%, energy consumption and delay is low, while full coverage is achieved. When the protocol is biased towards both frequency and density ða ¼ 0:5; b ¼ 0:5Þ, the protocol achieves good success rate (92–98%), but slightly lower than the previous case, almost identical energy consumption and higher delay. When the protocol is biased towards density ða ¼ 0:0; b ¼ 1:0Þ, the protocol achieves success rate ranging from 75% to 95%, almost identical energy consumption to the previous cases, higher delay and lower coverage than the previous two cases. In all cases, an increase in s results in improvement of the success rate and delay, especially when a ¼ 0:0; b ¼ 1:0; this
910
I. Chatzigiannakis et al. / Computer Communications 31 (2008) 896–914 100 97.5 95 90
Success Rate %
85 80
60
40 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology l=200.0, s=4 l=200.0, s=8
l=100.0, s=4 l=100.0, s=8
l=66.66, s=4 l=66.66, s=8
Fig. 19. Success rate of protocol P4 for a line trajectory of variable length for topologies with 0, 2 and 4 pockets with
nP nn
2 f1=2; 2=3g.
4.5
Energy Dissipation (Joules)
4.4 4.3 4.2 4.1 4 3.9 3.8 3.75 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology l=200.0, s=4 l=200.0, s=8
l=100.0, s=4 l=100.0, s=8
l=66.66, s=4 l=66.66, s=8
Fig. 20. Energy dissipation of protocol P4 for a line trajectory of variable length for topologies with 0, 2 and 4 pockets with
result is consistent with our findings so far. These results show that when a ¼ 1:0; b ¼ 0:0 the sink visits the whole network area fast. The protocol attempts to equalize the visits in each area, passing multiple times from each area thus collecting all the available data. On the contrary, when a ¼ 0:0; b ¼ 1:0 the sink remains around areas of relatively higher density and it takes a lot of time to visit new areas. When a ¼ 0:0; b ¼ 1:0, the covered set is around 80–90% for s ¼ 4 and about 95% for s ¼ 8, thus the sink never visits all areas in the network. This effect can be also seen when there are 2 pockets in the network for s ¼ 4; the sink stays close to the areas containing the pockets. Since the pockets contain many nodes, it collects more data than in the case of 0 or 4 pockets when the
nP nn
2 f1=2; 2=3g.
bias is not that strong. When a ¼ 0:5; b ¼ 0:5 an intermediate behavior is observed but the effect of the frequency bias appear to dominate, since the results are closer to the case of a ¼ 1:0; b ¼ 0:0. We examine the effect of deterministic mobility in Figs. 15–18 and 19–22 for circular and linear trajectories, respectively. We notice that the length of the trajectory greatly affects the performance of P4. Both for line and circle trajectories when l decreases also the success rate decreases and the energy dissipation increases. This is because messages need to be propagated over multiple hops, possibly traveling over great distance when compared to other solutions. Also, the delay decreases when l decreases which is expected since the round trip time of the sink is reduced.
I. Chatzigiannakis et al. / Computer Communications 31 (2008) 896–914
911
1700
Average Delivery Delay (seconds)
1600 1500 1400 1300 1200 1100 1000 900 800 700 600 500 400 300 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology l=200.0, s=4 l=200.0, s=8
l=100.0, s=4 l=100.0, s=8
l=66.66, s=4 l=66.66, s=8
Fig. 21. Average delay of protocol P4 for a line trajectory of variable length for topologies with 0, 2 and 4 pockets with
nP nn
2 f1=2; 2=3g.
100 97.5 95 90 85
Coverage %
80
60
40 0
2-1/2
2-2/3
4-1/2
4-2/3
Topology l=200.0, s=4 l=200.0, s=8
l=100.0, s=4 l=100.0, s=8
l=66.66, s=4 l=66.66, s=8
Fig. 22. Covered set of protocol P4 for a line trajectory of variable length for topologies with 0, 2 and 4 pockets with
Coverage ranges from 75% to 90%, as we can see from Fig. 18, thus whole areas of the network are disconnected from the sink. The effect of increased mobility speed is mostly negative since in most configurations success rate drops, while energy consumption and delay increases. This is because, as we already observed when examining P2, relay nodes have reduced communication time to the sink, collisions increase due to multi-hop communication and parts of the network are disconnected especially in nonuniform distributions where density is low in some areas. When examining different topologies, the behavior of P4 for a circle trajectory is better (with respect to consumed energy and success rate) when 2 and 4 pockets are present especially when l ¼ 200. This is because, due to the circular
nP nn
2 f1=2; 2=3g.
trajectory the average minimum distance of the sink from the nodes is reduced and the sink traverses regions near the pockets. Overall, P4 with the circle trajectory is better when l ¼ 200 but as l decreases P4 with the line trajectory performs better. In the final set of experiments we evaluate the overall advantages and disadvantages of our protocols (that exploit sink mobility) with ‘‘classic” multi-hop data propagation protocols (that assume a static sink). We select five different configurations of our protocols and compare them to a well known static multi-hop data propagation paradigm, the Directed Diffusion [3]. In this set of experiments we assume random uniform distribution of nodes (i.e., no pockets appear), the selected configurations of our proto-
912
I. Chatzigiannakis et al. / Computer Communications 31 (2008) 896–914 100 97.5 95 90 85
Success Rate %
80
70
60
40 0 P1 s=8 P2 hops=2, s=4 P3 α=1.0, β=0.0, s=8
P4 circle s=4 P4 line s=8 Directed Diffusion
Fig. 23. Success rate for selected configurations of protocols P1, P2, P3, P4 and Directed Diffusion for topology with 0 pockets.
6.5 6.4 6.3 6.2
Energy Dissipation (Joules)
6
5
4.4 4.2 4 3.9 3.75 0 P1 s=8 P2 hops=2, s=4 P3 α=1.0, β=0.0, s=8
P4 circle s=4 P4 line s=8 Directed Diffusion
Fig. 24. Energy dissipation for selected configurations of protocols P1, P2, P3, P4 and Directed Diffusion for topology with 0 pockets.
cols are such that the success rate is maximum for the uniform topology. For the case of Directed Diffusion we fix the static sink at position ðD=2; 0Þ. As shown in Fig. 23, our protocols achieve much higher success rate (up to 140% higher) and significantly lower, about 40%, energy consumption than Directed Diffusion (see Fig. 24). This is somewhat expected since Directed Diffusion needs to construct routing paths that consist of several hops over a large area, thus increasing energy consumption. Also, due to the relatively sparse network setting, Directed Diffusion may not always discover alternative paths towards the nodes or some nodes may be disconnected, thus bottlenecks appear that decrease success rate. This indicates that having a mobile sink (and by taking advantage of its mobil-
ity) can greatly extend the lifetime of the network. On the other hand, the time-efficiency of our protocols is related to the mobility of the sink (speed, mobility pattern), hence they exhibit much higher delivery delays than Directed Diffusion that achieves a propagation delay of about 1.5 s (see Fig. 25). 5. Conclusions and future work In this work, we investigate the impact of having a sink moving in the network area and collecting data. We have presented a collection of mobility patterns and data collection strategies that can be employed in applications where the sink is mobile (mostly related to ambient intelligence).
Average Delivery Delay (seconds)
I. Chatzigiannakis et al. / Computer Communications 31 (2008) 896–914
913
1700 1600 1500 1400 1300 1200 1100 1000 900 800 700 600 500 400
1 0 P1 s=4 P2 hops=2, s=4 P3 α=1.0, β=0.0, s=8
P4 circle s=4 P4 line s=8 Directed Diffusion
Fig. 25. Average delay for selected configurations of protocols P1, P2, P3, P4 and Directed Diffusion for topology with 0 pockets.
Our experimental comparison demonstrates the relative advantages and disadvantages and that different trade-offs can be achieved by each approach. Our results show that for applications where time efficiency is not critical it is better to let the sink suffer the burden of traversing the whole network area (as in protocols P1 and P3), since the greater energy savings are achieved in this way. By trading off some energy efficiency the delivery delay can be significantly reduced if a limited multi-hop approach is used as in protocol P2. For applications that the mobility capabilities of the sink are limited, but can tolerate some loss of information and increased energy consumption, the best approach is to follow a fixed trajectory with multi-hop data propagation. We plan to further investigate the effect of the sink mobility on the performance of wireless sensor networks for different motion rates, network area sizes and sensor node densities. Also we plan to look into different ways to perform a biased walk, possibly by also adjusting the movement speed and/or a walk over regular spanning regions of the original area. It is also interesting to further integrate the multi-hop approaches with the passive strategy leading to more efficient hybrid solutions. Also, a very interesting new line of research is to assume many mobile sinks. In this case several new challenges arise; some of the algorithms presented in this paper may achieve better performance in the existence of many sinks, but more advanced techniques that coordinate the sinks may be necessary in order to fully benefit from the extra mobility. A first attempt to this direction is presented by our group in [26], a work developed in parallel to the presented ideas. Acknowledgements This work has been partially supported by the IST Programme of the European Union under contract number
IST-2005-15964 (AEOLUS), the Programme PYTHAGORAS under the European Social Fund (ESF) and Operational Program for Educational and Vocational Training II (EPEAEK II). Also, by the Programme PENED under contract number 03ED568, co-funded 75% by European Union – European Social Fund (ESF), 25% by Greek Government – Ministry of Development – General Secretariat of Research and Technology (GSRT), and by Private Sector, under Measure 8.3 of O.P. Competitiveness – 3rd Community Support Framework (CSF). References [1] I. Akyildiz, W. Su, Y. Sankarasubramaniam, E. Cayirci, Wireless sensor networks: a survey, Journal of Computer Networks 38 (2002) 393–422. [2] I. Chatzigiannakis, S. Nikoletseas, P. Spirakis, Smart dust protocols for local detection and propagation, in: 2nd ACM International Annual Workshop on Principles of Mobile Computing (POMC), 2002, pp. 9–16, also, in the ACM Mobile Networks (MONET) Journal, Special Issue on Algorithmic Solutions for Wireless, Mobile, Ad Hoc and Sensor Networks, Vol. 10 (1) (2005). [3] C. Intanagonwiwat, R. Govindan, D. Estrin, Directed diffusion: a scalable and robust communication paradigm for sensor networks, in: 6th ACM/IEEE Annual International Conference on Mobile Computing (MOBICOM 2000), 2000, pp. 56–67. [4] A. Boukerche, F.H.S. Silva, R.B. Araujo, Context interpretation based wireless sensor networks for the emergency preparedness class of application, in: 2nd International Workshop on Algorithmic Aspects of Wireless Sensor Networks (ALGOSENSORS 2006), 2006. [5] A. Kansal, A. Somasundara, D. Jea, M. Srivastava, D. Estrin, Intelligent fluid infrastructure for embedded networks, in: 2nd ACM/ USENIX International Conference on Mobile Systems, Applications, and Services (MobiSys04), 2004. [6] C. Schindelhauer, Mobility in wireless networks, in: 32nd Annual Conference on Current Trends in Theory and Practice of Informatics, Czech Republic, 2006. [7] I. Chatzigiannakis, A. Kinalis, S. Nikoletseas, Sink mobility protocols for data collection in wireless sensor networks, in: 4th ACM/IEEE International Workshop on Mobility Management and Wireless Access Protocols (MobiWac), Torremolinos, Spain, 2006, pp. 52–59.
914
I. Chatzigiannakis et al. / Computer Communications 31 (2008) 896–914
[8] I. Chatzigiannakis, T. Dimitriou, S. Nikoletseas, P. Spirakis, A probabilistic forwarding protocol for efficient data propagation in sensor networks, Ad Hoc Networks Journal 4 (5) (2005) 621–635, also, in: 5th European Wireless Conference on Mobile and Wireless Systems beyond 3G (EW 2004), pp. 344–350, 2004.. [9] I. Chatzigiannakis, A. Kinalis, S. Nikoletseas, Adaptive energy management for incremental deployment of heterogeneous wireless sensors, Theory of Computing Systems (TOCS) Journal (2007) in print to appear in 2007, also, in: Proceedings of the Seventeenth (17th) Annual ACM Symposium on Parallelism in Algorithms and Architectures (SPAA 2005), ACM Press, 2005, pp. 96–105. [10] P. Leone, J. Rolim, S. Nikoletseas, An adaptive blind algorithm for energy balanced data propagation in wireless sensor networks, in: IEEE International Conference on Distributed Computing in Sensor Networks (DCOSS), Lecture Notes in Computer Science (LNCS), vol. 3267, Springer-Verlag, 2005, pp. 35–48. [11] C.E. Perkins, Ad Hoc Networking, Addison-Wesley, 2001. [12] V.D. Tracy Camp, Jeff Boleng, A survey of mobility models for ad hoc network research, Wireless Communications and Mobile Computing 2 (2002) 483–502. [13] S. Basagni, D. Turgut, S. Das, Mobility-adaptive protocols for managing large ad hoc networks, in: IEEE International Conference on Communications (ICC 2001), vol. 5, 2001, pp. 1539–1543. [14] S. Basagni, I. Chlamtac, V.R. Syrotiuk, B.A. Woodward, A distance routing effect algorithm for mobility (DREAM), in: MobiCom’98: Proceedings of the 4th Annual ACM/IEEE International Conference on Mobile Computing and Networking, ACM, New York, NY, USA, 1998, pp. 76–84. [15] P. Juang, H. Oki, Y. Wang, M. Martonosi, L. Peh, D. Rubenstein, Energy-efficient computing for wildlife tracking: design tradeoffs and early experiences with zebranet, in: 10th International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS02), 2002. [16] A. Chakrabarti, A. Sabharwal, B. Aazhang, Using predictable observer mobility for power efficient design of sensor networks, in: IPSN, 2003. [17] R.C. Shah, S. Roy, S. Jain, W. Brunette, Data mules: modeling a three-tier architecture for sparse sensor networks, in: IEEE
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25] [26]
Workshop on Sensor Network Protocols and Applications (SNPA), 2003. D. Jea, A. Somasundara, M. Srivastava, Multiple controlled mobile elements (data mules) for data collection in sensor networks, in: 1st IEEE Conference in Distributed Computing in Sensor Systems (DCOSS05), 2005, pp. 244–257. J. Luo, J.-P. Hubaux, Joint mobility and routing for lifetime elongation in wireless sensor networks, in: 24th IEEE INFOCOM, Miami, USA, 2005. D. Goldenberg, J. Lin, A. Morse, B. Rosen, Y. Yang, Towards mobility as a network control primitive, in: 5th ACM International Symposium on Mobile Ad-Hoc Networking and Computing (MobiHoc04), 2004. C. Avin, B. Krishnamachari, The power of choice in random walks: an empirical study, in: MSWiM’06: Proceedings of the 9th ACM International Symposium on Modeling Analysis and Simulation of Wireless and Mobile Systems, ACM Press, New York, NY, USA, 2006, pp. 219–228. J. Luo, J. Panchard, M. Piorkowski, M. Grossglauser, J.-P. Hubaux, Mobiroute: routing towards a mobile sink for improving lifetime in sensor networks, in: IEEE International Conference on Distributed Computing in Sensor Networks (DCOSS), Lecture Notes in Computer Science (LNCS), vol. 4026, Springer Verlag, 2006, pp. 480–497. M. Ma, Y. Yang, Sencar: an energy efficient data gathering mechanism for large scale multihop sensor networks, in: IEEE International Conference on Distributed Computing in Sensor Networks (DCOSS), Lecture Notes in Computer Science (LNCS), vol. 4026, Springer-Verlag, 2006, pp. 498–513. A. Boukerche, R. Pazzi, R. Araujo, A supporting protocol to periodic, event-driven and query-based application scenarios for critical conditions surveillance, in: 1st International Workshop on Algorithmic Aspects of Wireless Sensor Networks (ALGOSENSORS 2004), 2004. Crossbow technology inc., mica motes, http://www.xbow.com/. A. Kinalis, S. Nikoletseas, Scalable data collection protocols for wireless sensor networks with multiple mobile sinks, in: 40th Annual Simulation Symposium (ANSS), Norfolk, VA, USA, 2007, pp. 60–69.
