Hortonworks-Euskaltel

Big Data

Euskaltel's customers consumption analyzed in real time

Euskaltel - Telecommunications service provider

Intro

It is undeniable that we live stuck to our mobile phone and for Euskaltel's costumers it is vital to have access to the Internet at the maximum available speed at all times, so the real time query of the mobile data consumption especially relevant to them.

In collaboration with Euskaltel, a use case was developed to improve the quality of information and its availability. Although the case was simple at a technical level, it helped to maximize the value obtained from the analytical Big Data in a short time. It was intended: 

To be able to determine the consumption of the mobile network in real time in a completely decoupled way of productive machines, guaranteeing a scalable growth in the number of requests, and at the same time knowing both upload and download bandwidth.

Liberating the productive machines from the task of analizing the real time consumption implies a substantial improvement of the hardware's capacity to process requests.

Gustavo Fernández - Zylk CTO

Solution

In order to do that, we opted to use the following components: Nifi, Kafka, Flink, Hbase, API.

Once all the data has been processed, this data is exposed for the exploitation through an API-GATEWAY. Mainly, what is done is to process log files using a NIFI cluster. Subsequently, this data is sent to a Kafka tail from where they are consumed by Flink Jobs. These Jobs correlate different consumption types and they store in a set of Hbase tables.

This system made it possible to achieve the objective. The exploitation of the obtained data is used by other systems, which invoke it via API REST, without compromising productive services. As we discussed at the beginning of this success story, what we were looking for was:

  • Avoid compromising productive services.
  • Create a case that contributes added value quickly and easily.
  • Enable greater scalability in customer data consumption queries, which reduces the impact of queries growth derived from customers growth.