GSN Cloud Storage and Processing

Description

GSN (Global Sensor Networks ) is capable of managing configurable virtual sensors through a wide range of wrappers, and is able to manage one-shot and continuous queries, even in a distributed environment with several GSN instances. However, each GSN instance runs on a single machine, and uses a relational-based data storage underneath. While in most medium-size sensor deployments this is just enough, when it comes to process very large numbers of sensor observations, and at very high incoming rates, scalability can become a problem at various stages. Cloud infrastructures and parallelized processing frameworks have been developed in the latest years, simplifying the use of multi-node computing infrastructures. In the case of GSN, this type of technologies can help boosting the processing and querying times at various stages:

– Wrappers: huge datasets are imported into GSN through wrappers (e.g. Alpine 3D datasets). The wrapping tasks can be transformed and parallelized in a MapReduce-like workflow in the cloud.

– Storage: GSN currently supports only relational DBMS for storage. It can be extended to support other NoSQL systems such as HBase for sensor data storage.

– Processing: Query processing, can be delegated to cloud-ready infrastructures (like Storm or Spark), which are able to process streams in a multi-node environment.

The project proposed can include one of these aspects or a combination of them, depending on the time schedule.

Contacts

In case of any questions, please drop us an email or come to our offices:

• Jean Paul Calbimonte Perez (BC 118)