image

In distributed environments that collect or monitor data, useful data may be spread across multiple distributed nodes, but users or applications may wish to access that data from a single location. One of the most common ways to facilitate centralized access to distributed data is to maintain copies of data objects of interest at central locations using replication. In a typical replication environment, illustrated abstractly in Figure 1, a central data repository maintains copies, or replicas of data objects whose master copies are spread across multiple remote and distributed data sources. Replicas are kept synchronized to some degree with remote master copies using communication links between the central repository and each source. In this way, querying and monitoring of distributed data can be performed indirectly by accessing replicas in the central repository. While querying and monitoring procedures tend to become simpler and more efficient when reduced to centralized data access tasks, a significant challenge remains: that of performing data replication efficiently and effectively. Ideally, replicas of data objects at the central repository are kept exactly consistent, or synchronized, with the remote master copies at all times

Recent advances in wireless technologies and microelectronics have made feasible, both from a technological as well as an economical point of view, the deployment of densely distributed sensor networks Although today’s sensor nodes have relatively small processing and storage capabilities, driven by the economy of scale, it is already observed that both are increasing at a rate similar to Moore’s law. In applications where sensors are powered by small batteries and replacing them is either too expensive or impossible designing energy efficient protocols is essential to increase the lifetime of the sensor network. Since radio operation is by far the biggest factor of energy drain in sensor nodes minimizing the number of transmissions is vital in data-centric applications. Even in the case when sensor nodes are attached to larger devices with ample power supply, reducing bandwidth consumption may still be important due to the wireless, multi-hop nature of communication and the short-range radios usually installed in the nodes. Data-centric applications thus need to devise novel dissemination processes for minimizing the number of messages exchanged among the nodes. Nevertheless, in densely distributed sensor networks there is an abundance of information that can be collected. In order to minimize the volume of the transmitted data, we can apply two well-known ideas.

image

Get In Touch

Technologies we work with