Kognitio wins on query performance

Kognitio beats Vertica, Oracle, Hadoop and PIG on query performance


WX2 test results compared to competing technologies; queries return in a fraction of the time without any requirement to tokenize, index, columnize or perform any other unnatural acts on data

Chicago, IL, October 28, 2011 — A recent performance study using openly-available data and query language statements has shown how Kognitio’s in-memory analytic database satisfies queries faster than Vertica, Oracle, Hadoop and PIG. Specifically, when compared, the study shows that WX2 returned answers to questions in a fifth of the time it took Vertica and a quarter of the time it took Oracle, while Hadoop and PIG were off the chart in terms of how long they took to satisfy the queries that were run.

The study centers around counting triangles in an undirected graph with reciprocal edges. The idea is to run queries to find out the quantity of vertexes and edges, which edges are joined and which vertexes join which edges, which is regarded as a computationally intensive exercise. Such a concept can be used to understand social networks and how people are connected and interact with others.

Using data that was made available by Vertica on Github, Kognitio loaded it on to a 4 node cluster and ran the same queries that had been run by Vertica and Oracle on a 4 node cluster. Using Kognitio WX2, the queries were executed in 11.48 seconds, whereas the Vertica tests show a result of 97 seconds, Oracle 90 seconds and 14 seconds following an Oracle SQL rewrite. Using Hadoop, queries were returned in 3,900 seconds and using PIG in 2,151 seconds.

“Three constraints drive users to become disenfranchised with their analytic platform – performance, performance, performance,” explained Steve Millard, Chief Operating Officer at Kognitio. “This test proves that Kognitio is at the forefront of the analytic database market offering unprecedented query speeds to data by leveraging the power of in-memory analytics. Unlike other solutions on the market that call for data to be loaded “efficiently”, tokenized or that need to have other unnatural acts performed on it such as columnization, WX2 continues to offer the most straightforward and easy-to-use analytic database solution that allows users to get answers to their business questions in just seconds.”

About the perfomance study

Kognitio ran the tests using WX2 on 4 HP servers, each with 128GB of RAM and 24 cores per server.

Vertica ran the tests using Vertica, PIG and Hadoop on 4 HP servers, each with 96GB of RAM and 12 cores per server. Results can be found here

Tests were run using Oracle Exadata 2-2 hardware with 2 socket, 12 core Westmere-EP nodes and Oracle Database 11.2.0.2. Results can be found here

To download the data set used in all tests, go here

To download the query statements, go here

The results on a 86 million row data set:

Database Time (seconds)
Kognitio WX2 11.48
Oracle – post SQL re-write 14
Oracle 90
Vertica 97
PIG 2,151
Hadoop 3,900

About Kognitio

Kognitio is a long-standing innovator in the data warehousing, business intelligence and analytics markets. The company has pioneered many of the technologies now employed by state-of-the-art data warehouse/BI systems including the industry’s first in-memory analytic database, data warehouse appliance, MPP shared-nothing database, and Data warehousing as a Service (DaaS) cloud-based solution. Kognitio’s award-winning WX2 product is the industry’s fastest and most scalable analytic database, enabling firms to turn massive amounts of raw, complex data into valuable insight to solve their most urgent business problems. Its clients span many industries including market research, consumer packaged goods, retail, telecommunications, financial services, insurance, gaming, media and utilities. Kognitio is based in Bracknell, UK, with North American headquarters in Chicago and offices in New York, Raleigh, Dallas, Minneapolis and other U.S. cities. More information is available at kognitio.com on Twitter and Facebook.

Facebook

Twitter

LinkedId