Monday, June 10, 2013

NoSQL Based Middleware for Sensor Data Management

Along with the rapid development of technology, sensors are more and more intelligent, autonomous, small and cheap. Different kinds of sensors are already being used in various domains such as medical, environmental, military, urban and industrial. These applications require different kinds of data such as temperature, humidity, pressure, GPS values. For managing these types of different sensor data and retrieving information, we need to have heterogeneous middleware system. We noted that the NoSQL approach which does not rely on the ACID properties is a better match for a database abstraction for sensor networks. Thus we proposed NoSQL based middleware for managing these heterogeneous distributed sensor data.

Sensor networking is an emerging technology and it has become increasingly attractive area for researchers. The applications for sensor networks are varied; typically those applications are involving tasks such as monitoring, tracking or controlling systems. Area monitoring, environmental monitoring, green house monitoring, landside detection, machine health monitoring and fleet monitoring are some of the areas for sensor networks [1].

Retrieving information from these different kinds of heterogeneous data is more challenging task. In order to retrieve data we have to use middleware platform [2, 3]. However the problem is; we have to deal with application dependent data to retrieve information. As a solution to these problems we proposed NoSQL based middleware for dealing with these application dependent data and retrieving information from heterogeneous sensor data, triggered from various applications. 

Dealing with large set of heterogeneous data, using a table based database systems, it needs lot of resources to store such a massive data and the operations are time consuming to retrieve information. With regards to NoSQL databases [4], it handle massive amount of data in much easier manner, and the performances wise it is very faster comparing with relational database management systems.

Problem and Motivation
Even though there are middlewares developed for sensor data management, all those are depending on relational aspect based on ACID properties [2]. However the current middleware systems in sensor data management faced scalability, mobility and heterogeneity problems. Therefore with NoSQL Middleware approach we can overcome those limitations, because NoSQL guarantees the scalability of data by retrieving information from heterogeneous data [5].

Literature Review
Existing sensor data management systems can be categorized as distributed sensor networks [2, 3, 6], centralized sensor networks [7, 8] and hybrid sensor networks [9, 10]. In distributed sensor networks, sensors are usually supposed to be homogeneous and they used same communication protocol [3]. This is a limitation for deploying large scale sensor network. In centralized sensor network systems, heterogeneity is not much explored. Also data retrieving queries are expensive and that is a limitation for the lifetime of the sensor network. In hybrid approach, we can use heterogeneous sensors. Due to that, scalability of the sensor data is very much higher than other sensor network categories.

For the time being there is a NoSQL query processing system for wireless ad-hoc and sensor networks [11]. It was proven that NoSQL based query processing system is better than other query processing systems such as TinyDB [2], TikiriDB [12], Cougar [13, 14] and Vector Programming database abstraction [15].

Proposed Methodology
Due to the scalability of data in hybrid sensor network approach, we have to introduce new kind of middleware for sensor data management. NoSQL (eg: - RedisDB [16], CouchDB [5], Cassandra [4], LevelDB) is the best data management system for highly scalable data. As discussed under problem and motivation sections; limited power resources, scalability, mobility and dynamic network topologies are the major challenges to design middleware for sensor data management. For overcome these problems, NoSQL based middleware is a better match for sensor data management.

Going through NoSQL based middleware for sensor data management we can maintain the high scalability of data to retrieve information from distributed heterogeneous sensor data, quick response to queries and retrieve information quickly, low energy consumption and we can use it within limited memory capacity.

Expected Outcome
Using NoSQL based middleware for sensor data management, we can achieve,
  • Quickly retrieving information from highly scalable heterogeneous sensor data
  • Low energy consumption
  • Low memory usage
  • Increase sensor network lifetime
[1] C. Weyer, V. Turau, A. Lagemann, and J. Nolte, “Programming wireless sensor networks in a selfstabilizing style,” in Sensor Technologies and Applications, 2009. SENSORCOMM’09. Third International Conference on, vol. 0, pp. 610–616, IEEE, 2009.

[2] S. R. Madden, M. J. Franklin, J. M. Hellerstein, and W. Hong, “TinyDB: an acquisitional query processing system for sensor networks,” ACM Transactions on Database Systems, vol. 30, pp. 122–173, Mar. 2005.

[3] P. Bonnet, J. Gehrke, and P. Seshadri, “Towards sensor database systems,” in Mobile Data Management, pp. 3–14, Springer, 2001.

[4] A. Lakshman and P. Malik, “Cassandra: a decentralized structured storage system,” ACM SIGOPS Operating Systems Review, vol. 44, no. 2, pp. 35–40, 2010.

[5] M. Seeger and S. Ultra-Large-Sites, “Key-value stores: a practical overview,” Computer Science and Media, Stuttgart, 2009.

[6] M. Chen, “Energy-efficient differentiated directed diffusion (EDDD) in wireless sensor networks,” Computer Communications, vol. 29, pp. 231–245, Jan. 2006.

[7] D. Abadi, D. Carney, U. Cetintemel, M. Cherniack, C. Convey, S. Lee, M. Stonebraker, N. Tatbul, and S. Zdonik, “Aurora: a new model and architecture for data stream management,” The VLDB Journal, vol. 12, no. 2, pp. 120–139, 2003.

[8] D. P. Arvind, A. Arasu, B. Babcock, S. Babu, M. Datar, K. Ito, I. Nishizawa, J. Rosenstein, and J. Widom, “Stream: The stanford stream data manager,” IEEE Data Engineering Bulletin, vol. 26, pp. 19–26, 2003.

[9] S. Madden and M. Franklin, “Fjording the stream: An architecture for queries over streaming sensor data,” in Data Engineering, 2002. Proceedings. 18th International Conference on, pp. 555–566, IEEE, 2002.

[10] D. Abadi, W. Lindner, S. Madden, and J. Schuler, “An integration framework for sensor networks and data stream management systems,” in Proceedings of the Thirtieth international conference on Very large data bases-Volume 30, pp. 1361–1364, VLDB Endowment, 2004.

[11] T. A. M. C. Thantriwatte and C. I. Keppetiyagama, “Nosql query processing system for wireless ad-hoc and sensor networks,” in Proceedings of the International Conference on Advances in ICT for Emerging Regions (ICTer), 2011, ICTer 2011, pp. 78–82, IEEE, 2011.

[12] N. M. Laxaman, M. D. J. S. Goonathillake, and K. D. Zoysa, “TikiriDB : Shared Wireless Sensor Network Database for Multi-User Data Access,” CSSL 2010.

[13] Y. Yao, “The cougar approach to in-network query processing in sensor networks,” ACM SIGMOD Record, vol. 31, p. 9, Sept. 2002.

[14] W. F. Fung, U. Hall, D. Sun, U. Hall, J. Gehrke, and U. Hall, “COUGAR : The Network is the Database,” System, no. July 2001, pp. 14853–14853, 2002.

[15] T. Sookoor and K. Whitehouse, “Vector Programming Abstraction for Sensor Networks,” 2010.

[16] Citrusbyte, “Redis.” Website, August 10 2012.

No comments:

Post a Comment