NoSQL and BigData

Some words you hear more and more often and you wonder : what is it really in reality? We were also wondering, so we launched an internal project based on these questions “is it that simple, is it that powerful? What are the conditions to go fast?”

We configured and used ElasticSearch on more or less structured data but in large amounts:

  • Internally, with the ELK (ElasticSearch / Logstash / Kibana) stack to manage the logs of our infrastructure:
      Reporting and visualization along several axes (time, equipment, services, …)
      and search: for example, in two clicks, one can know whether all of our backups have succeeded at the first try on last Sunday.
  • To index MDM contents into ElasticSearch and allow near-instant search on completely different criteria from the original Data Model, including geopositioning attributes for geographic search. Application: with a responsive website, on a mobile + GPS, it instantly finds references near its position.
  • Extension to the previous application, we used ElasticSearch to index the geographical public database of OpenStreetMap. With the correct OpenSource import module, it can be ready and launched in a few hours!

 

The results are very impressive in terms of response time and ease of implementation. There are obviously some points to watch on the data format, but the development time and required infrastructure are within reason.

Please note: configuration is important to operate successfully, especially for API requests.

To go further, come to see us for a demonstration and technical explanations.