## Cache placement in sensor networks under update cost constraint (2005)

### Cached

### Download Links

Venue: | In Proc. of AdHoc-Now |

Citations: | 7 - 3 self |

### BibTeX

@INPROCEEDINGS{Tang05cacheplacement,

author = {Bin Tang and Himanshu Gupta},

title = {Cache placement in sensor networks under update cost constraint},

booktitle = {In Proc. of AdHoc-Now},

year = {2005}

}

### OpenURL

### Abstract

In this paper, we address an optimization problem that arises in context of cache placement in sensor networks. In particular, we consider the cache placement problem where the goal is to determine a set of nodes in the network to cache/store the given data item, such that the overall communication cost incurred in accessing the item is minimized, under the constraint that the total communication cost in updating the selected caches is less than a given constant. In our network model, there is a single server (containing the original copy of the data item) and multiple client nodes (that wish to access the data item). For various settings of the problem, we design optimal, near-optimal, heuristic-based, and distributed algorithms, and evaluate their performance through simulations on randomly generated sensor networks. 1

### Citations

1787 | Directed Diffusion: A Scalable and Robust Communication Paradigm for Sensor Networks
- Intanagonwiwat, Govindan, et al.
- 2000
(Show Context)
Citation Context ...ncy a critical design goal. There has been a significant interest in designing algorithms, applications, and network protocols to reduce energy usage of sensors. Examples include energy-aware routing =-=[18]-=-, energy-efficient information processing [10, 13], and energy-optimal topology construction [29]. In this article, we focus on designing techniques to conserve energy in the network by caching data i... |

1616 | GPSR: greedy perimeter stateless routing for wireless networks
- Karp, Kung
- 2000
(Show Context)
Citation Context ...ory resources at each sensor node, a proactive routing protocol [27] that builds routing tables at each node is unlikely to be feasible. In such a case, a location-aided routing protocol such as GPSR =-=[22]-=- is sufficient for our purposes, if each node is aware of its location (either through GPS [17] or other localization techniques [2, 7]). Distributed Greedy-Steiner Algorithm. The distributed version ... |

1520 | Highly dynamic destination-sequenced distancevector routing (dsdv) for mobile computers
- Perkins, Bhagwat
- 1994
(Show Context)
Citation Context ...zed Greedy-Steiner Algorithm. Here, we assume the presence of an underlying routing protocol in the 6ssensor network. Due to limited memory resources at each sensor node, a proactive routing protocol =-=[27]-=- that builds routing tables at each node is unlikely to be feasible. In such a case, a location-aided routing protocol such as GPSR [22] is sufficient for our purposes, if each node is aware of its lo... |

749 | GPS-less Low Cost Outdoor Localization for Very Small Devices
- Bulusu, Heidemann, et al.
- 2000
(Show Context)
Citation Context ...le. In such a case, a location-aided routing protocol such as GPSR [22] is sufficient for our purposes, if each node is aware of its location (either through GPS [17] or other localization techniques =-=[2, 7]-=-). Distributed Greedy-Steiner Algorithm. The distributed version of the centralized Greedy-Steiner Algorithm consists of rounds. During a round, each non-cache node A estimates its benefit per unit up... |

431 | Directed diffusion for wireless sensor networking
- Intanagonwiwat, Govindan, et al.
- 2003
(Show Context)
Citation Context ...ion cost can be optimized by caching generated data at appropriately selected intermediate nodes. In particular, the datacentric sensor network applications which require efficient data dissemination =-=[6, 8]-=- will benefit from effective data caching strategies. In our model of the sensor network, there is a single data item at a given server node, and many client nodes. (See Section VI for a discussion on... |

319 | Approximation algorithms for metric facility location and k-median problems using the primal-dual schema and Lagrangian relaxation
- Jain, Vazirani
(Show Context)
Citation Context ...1, 20], under the assumption that the edge costs in the graph satisfy the triangular inequality. Without the triangular inequality assumption, either problem is as hard as approximating the set cover =-=[19, 25]-=-, and therefore cannot be approximated better than O(log |V |) unless NP ⊆ ˜P. Here, |V | is the size of the network. Several papers in the literature circumvent the hardness of the facility-location ... |

289 |
Global Positions System: Theory and Practice, Fourth Edition
- Wellenhoff, Lichtenegger, et al.
- 1997
(Show Context)
Citation Context ... at each node is unlikely to be feasible. In such a case, a location-aided routing protocol such as GPSR [22] is sufficient for our purposes, if each node is aware of its location (either through GPS =-=[17]-=- or other localization techniques [2, 7]). Distributed Greedy-Steiner Algorithm. The distributed version of the centralized Greedy-Steiner Algorithm consists of rounds. During a round, each non-cache ... |

288 | On the placement of Web server replicas
- Qiu, Padmanabhan, et al.
- 2001
(Show Context)
Citation Context ...hich consist of a large number of energyconstrained nodes, the constraint on the number of cache nodes is of little relevance. Cache placement problem has also been widely used in the web environment =-=[3, 5, 28]-=- and peer-to-peer networks [12, 16, 26] to alleviate problems such as server overloading, delayed respond time, and inadequate bandwidth. In particular, Qiu et al. [28] have addressed effective placem... |

242 | Replication strategies in unstructured peer-to-peer networks
- Cohen, Shenker
(Show Context)
Citation Context ...rgyconstrained nodes, the constraint on the number of cache nodes is of little relevance. Cache placement problem has also been widely used in the web environment [3, 5, 28] and peer-to-peer networks =-=[12, 16, 26]-=- to alleviate problems such as server overloading, delayed respond time, and inadequate bandwidth. In particular, Qiu et al. [28] have addressed effective placement of web server replicas over the Int... |

210 | Scalable information-driven sensor querying and routing for ad hoc heterogeneous sensor networks
- Chu, Haussecker, et al.
- 2001
(Show Context)
Citation Context ...ignificant interest in designing algorithms, applications, and network protocols to reduce energy usage of sensors. Examples include energy-aware routing [18], energy-efficient information processing =-=[10, 13]-=-, and energy-optimal topology construction [29]. In this article, we focus on designing techniques to conserve energy in the network by caching data items at selected sensor nodes in a sensor network.... |

205 | Improved combinatorial algorithms for facility location problems
- Charikar, Guha
(Show Context)
Citation Context ...mber constraint, i.e., that at most k nodes can be selected as caches. Both problems are NP-hard, and a number of constant-factor approximation algorithms have been developed for each of the problems =-=[9, 11, 20]-=-, under the assumption that the edge costs in the graph satisfy the triangular inequality. Without the triangular inequality assumption, either problem is as hard as approximating the set cover [19, 2... |

140 |
Steiner minimal trees
- Gilbert, Pollak
- 1968
(Show Context)
Citation Context ...ata item from the nearest node (either a cache or the server) having a copy of the data item. The update cost incurred in updating a set of caches M is modeled as the cost of the optimal Steiner tree =-=[14]-=- spanning the server and the set of caches. This problem is obviously NP-hard, as even the Steiner tree problem is known to be NP-hard [4]. In this article, we look at the above problem in various sta... |

125 | Data-centric storage in sensornets
- Shenker, Ratnasamy, et al.
(Show Context)
Citation Context ... cost which is the main source of energy consumption. However, cache placement problem in sensor networks has several faces considering the specific characteristics of sensor networks. Shenker et al. =-=[30]-=- propose data centric storage (DCS) as a data dissemination paradigm for sensor networks. In DCS, data is stored, according to event type, at corresponding sensornet nodes. Data is also replicated to ... |

118 | The Cache Location Problem
- Krishnan, Raz, et al.
- 2000
(Show Context)
Citation Context ...m optimally in a tree of n nodes. In other related works on cache placement in trees, Xu et al. [30] discuss placement of “transparent” caches to minimize the sum of reads and writes, Krishnan et al. =-=[23]-=- consider a cost model based on cache misses, and Kalpakis et al. [21] consider a cost model involving reads, writes, and storage. In sensor networks, which consist of a large number of energyconstrai... |

115 |
Effective replica allocation in ad hoc networks for improving data accessibility
- Hara
- 2001
(Show Context)
Citation Context ...rgyconstrained nodes, the constraint on the number of cache nodes is of little relevance. Cache placement problem has also been widely used in the web environment [3, 5, 28] and peer-to-peer networks =-=[12, 16, 26]-=- to alleviate problems such as server overloading, delayed respond time, and inadequate bandwidth. In particular, Qiu et al. [28] have addressed effective placement of web server replicas over the Int... |

108 | Supporting cooperative caching in ad hoc networks
- Yin, Cao
- 2006
(Show Context)
Citation Context ...e cache placement problem in the specific context of ad hoc networks. Hara [16] addresses replica allocation methods for mobile ad hoc networks that can experience frequent disconnection. Yin and Cao =-=[31]-=- design and evaluate three simple cooperative caching techniques to efficiently support data access in ad hoc networks. In particular, they propose that intermediate nodes either cache data and/or nea... |

90 | Opportunity-based topology control in wireless sensor networks
- Liu, Ni
- 2010
(Show Context)
Citation Context ...ations, and network protocols to reduce energy usage of sensors. Examples include energy-aware routing [18], energy-efficient information processing [10, 13], and energy-optimal topology construction =-=[29]-=-. In this article, we focus on designing techniques to conserve energy in the network by caching data items at selected sensor nodes in a sensor network. The techniques developed in this paper are ort... |

78 | Improved approximation algorithms for capacitated facility location problems
- Chudak, Williamson
- 2005
(Show Context)
Citation Context ...mber constraint, i.e., that at most k nodes can be selected as caches. Both problems are NP-hard, and a number of constant-factor approximation algorithms have been developed for each of the problems =-=[9, 11, 20]-=-, under the assumption that the edge costs in the graph satisfy the triangular inequality. Without the triangular inequality assumption, either problem is as hard as approximating the set cover [19, 2... |

69 | Approximation algorithms for geometric median problems
- Lin, Vitter
- 1992
(Show Context)
Citation Context ...1, 20], under the assumption that the edge costs in the graph satisfy the triangular inequality. Without the triangular inequality assumption, either problem is as hard as approximating the set cover =-=[19, 25]-=-, and therefore cannot be approximated better than O(log |V |) unless NP ⊆ ˜P. Here, |V | is the size of the network. Several papers in the literature circumvent the hardness of the facility-location ... |

67 |
On the optimal placement of web proxies in the Internet
- Li, Golin, et al.
- 1999
(Show Context)
Citation Context ...P ⊆ ˜P. Here, |V | is the size of the network. Several papers in the literature circumvent the hardness of the facility-location and k-median problems by assuming that the network has a tree topology =-=[24, 30]-=-. In particular, Li et al. [24] address the optimal placement of web proxies in a tree topology, essentially designing an O(n 3 k 2 ) time dynamic programming algorithm to solve the k-median problem o... |

63 | On the optimal placement of web proxies - Li, Golin, et al. - 1999 |

44 | Energy-Conserving Data Placement and Asynchronous Multicast in Wireless Sensor Networks
- Bhattacharya, Kim, et al.
- 2003
(Show Context)
Citation Context ...ion cost can be optimized by caching generated data at appropriately selected intermediate nodes. In particular, the datacentric sensor network applications which require efficient data dissemination =-=[6, 8]-=- will benefit from effective data caching strategies. In our model of the sensor network, there is a single data item at a given server node, and many client nodes. (See Section VI for a discussion on... |

27 | Placement problems for transparent data replication proxy services
- Xu, Li, et al.
- 2002
(Show Context)
Citation Context ...P ⊆ ˜P. Here, |V | is the size of the network. Several papers in the literature circumvent the hardness of the facility-location and k-median problems by assuming that the network has a tree topology =-=[24, 30]-=-. In particular, Li et al. [24] address the optimal placement of web proxies in a tree topology, essentially designing an O(n 3 k 2 ) time dynamic programming algorithm to solve the k-median problem o... |

25 | Energy-efficient caching strategies in ad hoc wireless networks
- Nuggehalli, Srinivasan, et al.
- 2003
(Show Context)
Citation Context ...rgyconstrained nodes, the constraint on the number of cache nodes is of little relevance. Cache placement problem has also been widely used in the web environment [3, 5, 28] and peer-to-peer networks =-=[12, 16, 26]-=- to alleviate problems such as server overloading, delayed respond time, and inadequate bandwidth. In particular, Qiu et al. [28] have addressed effective placement of web server replicas over the Int... |

25 | Data storage placement in sensor networks
- Sheng, Li, et al.
- 2006
(Show Context)
Citation Context ...ata dissemination paradigm for sensor networks. In DCS, data is stored, according to event type, at corresponding sensornet nodes. Data is also replicated to avoid overloading. Recently, Sheng et al. =-=[29]-=- study the storage node placement problem to minimize the total energy for data collection and data query. Intanagonwiwat et al. [18] propose directed diffusion, a data dissemination paradigm for sens... |

23 |
An approximation algorithm for the Steiner tree problem
- Berman, Ramaiyer
- 1991
(Show Context)
Citation Context ... caches M is modeled as the cost of the optimal Steiner tree [14] spanning the server and the set of caches. This problem is obviously NP-hard, as even the Steiner tree problem is known to be NP-hard =-=[4]-=-. In this article, we look at the above problem in various stages – a tree topology, a graph topology with a simplified update cost model, a graph topology with the general update cost model – and pre... |

16 |
V.N.: Radar: An in-building RF-based user-location and tracking system
- Bahl, Padmanabhan
- 2000
(Show Context)
Citation Context ...le. In such a case, a location-aided routing protocol such as GPSR [22] is sufficient for our purposes, if each node is aware of its location (either through GPS [17] or other localization techniques =-=[2, 7]-=-). Distributed Greedy-Steiner Algorithm. The distributed version of the centralized Greedy-Steiner Algorithm consists of rounds. During a round, each non-cache node A estimates its benefit per unit up... |

13 |
SteinerOptimal Data Replication in Tree Networks with Storage Costs
- Kalpakis, Dasgupta, et al.
- 2001
(Show Context)
Citation Context ...ement in trees, Xu et al. [30] discuss placement of “transparent” caches to minimize the sum of reads and writes, Krishnan et al. [23] consider a cost model based on cache misses, and Kalpakis et al. =-=[21]-=- consider a cost model involving reads, writes, and storage. In sensor networks, which consist of a large number of energyconstrained nodes, the constraint on the number of cache nodes is of little re... |

11 | Energy-conserving data cache placement in sensor networks
- Prabh, Abdelzaher
- 2005
(Show Context)
Citation Context .../processing data in-network. Bhattacharya et al. [6] develop a distributed framework that improves energy consumption by application layer data caching and asynchronous update multicast. Prabh et al. =-=[27]-=- improve upon [6] by presenting and analyzing the optimality properties of Steiner data caching tree 5sTr Lv,u r V u Tu Rv,u (a) Subtree notations Cache Nodes Possible Cache Nodes No Cache X Lv,u V u ... |

9 |
Selection and maintenance of views in a data warehouse
- Gupta
- 1999
(Show Context)
Citation Context ... (d(i, M ∪ O1) − d(i, M ∪ O1 ∪ O2)) i∈V = β(O2, M ∪ O1) Now, we show that the Greedy Algorithm returns a solution with near-optimal benefit. The proof technique used here is 5ssimilar to that used in =-=[15]-=- for the closely related problem of selection of views in a data warehouse. Theorem 1: Greedy Algorithm (Algorithm 1) returns a solution M whose absolute benefit is of at least (1 − 1/e) times the abs... |

2 |
World wide web caching: Trends and technologies
- Barish, Obraczka
- 2000
(Show Context)
Citation Context ...hich consist of a large number of energyconstrained nodes, the constraint on the number of cache nodes is of little relevance. Cache placement problem has also been widely used in the web environment =-=[3, 5, 28]-=- and peer-to-peer networks [12, 16, 26] to alleviate problems such as server overloading, delayed respond time, and inadequate bandwidth. In particular, Qiu et al. [28] have addressed effective placem... |

2 |
Propagation, caching and replication on the web, http://www.w3.org/Propagation
- Berners-Lee, Mielsen
(Show Context)
Citation Context ...hich consist of a large number of energyconstrained nodes, the constraint on the number of cache nodes is of little relevance. Cache placement problem has also been widely used in the web environment =-=[3, 5, 28]-=- and peer-to-peer networks [12, 16, 26] to alleviate problems such as server overloading, delayed respond time, and inadequate bandwidth. In particular, Qiu et al. [28] have addressed effective placem... |