Cognos Analytics

Interactive Dashboarding on Big Data (Billions of Rows)

By Jason Tavoularis posted Thu January 16, 2020 04:43 PM


As a Senior Offering Manager for IBM Cognos Analytics, I regularly get asked the question “How does Cognos work with Big Data”.

My response has not changed much in the past several years – we exploit your data infrastructure for all its worth. The primary data access strategy for Cognos software is to delegate data processing, as much as possible, to a data server. Therefore, in most cases the volume of data is bounded by your data server’s capacity to respond to analytical queries within the threshold of your users’ wait time tolerance (typically users will not like waiting more than a few seconds for a request).

If I need to break this down further, I’ll expand that when connecting to a SQL interface of a data server for example, the Cognos software will generate SQL tailored for the type and version of the data server technology, optimized to minimize user wait time.

Usually, only as many rows as there are values to be displayed within Cognos need to be transferred from the data server to Cognos. So even if your data server is storing billions of records, if a bar chart in a Cognos dashboard is displaying five bars then there should only be five rows of data retrieved since it’s the data server computing all joins, aggregations, calculations, filters, etc., that result in the 5 values that get displayed in the Cognos visualization.

Of course Cognos has contingency solutions, like data sets, if you only have slow or overloaded data servers at your disposal. But here we’re talking Big Data, which means you will need Big Data infrastructure if you want to enable interactive analytics on large volumes of records.


Enter Varada

Some friendly folks at Varada recently reached out to us to collaborate on validating that Varada can indeed to deliver on its mission statement of enabling interactive analytics on large data volumes in conjunction with IBM Cognos Analytics.

Since Varada is based on Presto which Cognos already attenuates SQL generation for, and Varada embraces our philosophy of maximizing SQL push-down and in-database processing, there was nothing for us to do but try them out together … with billions of records of course 😊

The following video shows the result of Varada + Cognos proof of concept. It’s a visually engaging, highly interactive dashboard where behind-the-scenes Varada is accessing a data store with over two billion records and responding to all the user’s analytical gestures within a few seconds.

Despite the large data volumes, the snappy response times on the Varada-powered Cognos dashboard ensure users won’t be impeded from cross-examining intersections of data points that may lead them to uncover patterns which allow them to make smarter decisions.