February 24th, 2014


“We saw that Apache Cassandra integrates everything we looked for: easy scalability and set up, fault tolerance, great performance by writing and native integration with Hadoop.”

– Marcos Rosado , University of Extremadura, SmartPoliTech


Marcos Rosado



Marcos Rosado SmartPoliTech




SmartPoliTech is a project initiated at the Polytechnic School at the University of Extremadura (Spain) with the goal of turning our school into a large experimental ecosystem: a living-lab for the design, implementation, integration and validation of systems capable of creating smart spaces by means of SmartX technologies development. Our purpose is to create an energy efficient space and to make social and academic life easier on campus by appealing to the personal development of its inhabitants and users.


Sensor data with Cassandra 

We have been using platforms like Xively, to store the data produced by the sensors installed in our school and to visualize the information produced by the network of sensors.

Our goal is to create our own platform to store and process all the data involved in the project, which includes, not only data coming from sensors, but also information related to the school building model, the use of the building and so on. To support the expected volume of data that will have to be stored and managed, we are using Apache Cassandra, which properly fits the scalability and availability requirements for SmartPoliTech.


The perfect fit

We needed a system capable of handling large amounts of data, such as time series, with a high performance when writing more than reading (logging). We also wanted it to be easily scalable because it must grow with the project, and it must be compatible with Hadoop for data analysis.

After studying different alternatives, we understood that SQL databases aren’t the best option and it was necessary to use a NOSQL system. Finally, we saw that Apache Cassandra integrates everything we looked for: easy scalability and set up, fault tolerance, great  performance by writing and native integration with Hadoop. Also, it supports multiple datacenter or clusters, which is a nice feature because our university isn’t located in a single place and we may initiate new clusters in others departments or faculties.


Ready to scale

We have our own datacenter which is hosted here at the University. We’re using virtual machines on a blade server (Xeon machine with 24GB RAM) with spinning disks. Because we are still in an early stage of SmartPoliTech, the data volume is not too large, so that we have started with a simple Apache Cassandra configuration consisting of 3 nodes (virtual machines) with 4GB RAM and 1TB of HD storage. As new rooms (classrooms, laboratories, offices, departments, etc…) are incorporated to SmartPoliTech, new nodes will be added to the platform.


Making the difference

Our experience with the community is great! Planet Cassandra, IRC channels, people on Stack Overflow are very helpful. We’re grateful to all of them because at first we had many questions (about data design, Hadoop integration…) and small problems with the configuration.

There are smart and hardworking people behind Apache Cassandra and that makes all the difference.