Fuzzy Logix: data crunching should be parallelised, optimised, standardised & native-ised

Fuzzy Logix (like fuzzy logic, but with an X, get it?) has announced availability of its analytics suite DB Lytix on the Cloudera Enterprise 5 data platform.

So what is this? Essentially it’s a chance to perform predictive statistical analysis using analytics running natively on a cloud cluster.

The firm claims that DB Lytix for Cloudera could (for example) optimise a supply chain across thousands of stores in multiple countries taking into account factors such as local weather patterns and differences in demographics at each of its stores.

IoT volume overload

The huge data volumes generated by Internet of Things (IoT) devices can be effectively handled for large-scale applications in transportation, health, manufacturing, agriculture and utility industries.

Fuzzy Logix’ analytics solutions go for the parallelised and scalable play (fast too, obviously) that work where the data resides.

Why should developers care?

The CWDN blog asked Fuzzy Logix why software application development professionals should care about data analytics presented in these flavours. The firm reminded us that there is a big ease of use factor here i.e. developers won’t need to code their own data models.

Plus, with model standardisation – certified models guarantee accurate results (vs. self-coded models) and allow standardisation across the enterprise.

It’s that kind of ‘software packaging’ theme that we are hearing so much more about now in terms of.

Fuzzy Logix chief marketing officer Aashu VIrmani is on the record saying that his firm is seeing customer interest in running high-performance analytics on the Cloudera platform.

“Already, we have several large customers evaluating this joint solution for business-critical problems. This product is the result of close cooperation between Cloudera and Fuzzy Logix engineering teams to produce a tightly integrated and high quality solution,” he said, in a kind of press release style packaged comment.

There’s also optimized performance here – the algorithms exploit data and computational parallelisation. This is not easy to do. Coding for Hadoop is one thing. Parallelizing maths to take advantage of specific platform capabilities is another.

Cross platform support is also strong – the same models can run in Hadoop and traditional EDW.  This is important for those companies who keep some data in EDW and a bunch more in Hadoop.