Enterprise level, high-concurrency BI on Hadoop is possible with Kognitio

Enterprise Level BI is not simply about supporting complex SQL syntax. It is about supporting the needs of 100s of end users all with different needs. So a mixed workload and concurrency. In this video we prove that enterprise-level BI on Hadoop is possible – using 5.2billion rows of Transport for London (TfL) data as an example.

Video Transcript

Enterprise Level BI is not simply about supporting complex SQL syntax.

It is about supporting the needs of 100s of end users all with different needs. So a mixed workload and concurrency.

Our previous demo based on TfL open data showed interactive BI over 5.2bn rows for a single Tableau dashboard user.

Now, in this video we prove Kognitio can support a true enterprise workload using the same data and similar queries.

We executed a 100 scripts all simultaneously connected to Kognitio. Each script submits the next SQL query when its results are returned. All scripts were run continuously.

On the map every dot shows the location of a TfL bus stop used in a single query. As you can see there are lots of queries returning results These are used to colour the dots alerting us to stops with a long waiting time for example.

The timeline shows the average number of queries per second over each 10 second interval. This is consistently much higher than 100 so many of our scripts are able to run more than 1 query a second.

The continuous SQL submission from the scripts emulates the workload of 100s of end-users that would be submitting a query then digesting results before moving to a new query.

TfL open data is available from: tfl.gov.uk/infor-for/open-data-users

If you want to find out more about how Kognitio can help you support your BI requirements on Hadoop data then visit our website.