SiSense Tweaks Memory to Boost Data Analytics

Business intelligence software vendor SiSense is forecasting that soaring demand among small and mid-sized companies for data analytics could help triple it sales this year.

The Israeli company differentiates itself from competitors through its “In-Chip” technology that speeds up and scales data processing by tweaking existing processing resources in computers. This approach leverages CPU cache, generally the fastest available memory, along with RAM (in-memory) and disk memory to speed up data processing at rates it claims are up to 100 times faster than standard in-memory processing.

The company’s amped-up approach also includes a columnar database that it says “slices and stores” information as columns rather than as rows, the approach used in relational databases. The capability allows users to “pull” individual columns rather than an entire table during a standard query in which only a few columns of data are used. That’s the case, the company adds, regardless of the size of the data set.

The columnar database is said to boost performance using compression algorithms that minimize the amount of data handled by memory. It also decompresses data in CPU cache to conserve memory bandwidth, the company said.

Another tweak involves how business intelligence queries are handled. The feature creates and reuses query blocks, caching them individually in a “recycle bin.” In a new query, these blocks can be pulled from the bin and reused. The result, SiSense claims, is faster response times.

SiSense CEO Amit Bendov recently told Reuters that big data analysis requirements are outgrowing the capabilities of existing software. As a result, he said company sales have tripled three years in a row and it expects the same result this year.

U.S. customers include eBay and ESPN, Reuters reported. The Tel Aviv-based company plans to open a Silicon Valley office early next year.

The company also argues that its approach to business intelligence software that leverages all computing resources means a modern 64-bit chipset should be able to handle big data sets and complex queries. That could allow smaller users to crunch more data using commodity hardware.

Hence, the company maintains it can move its version of big data analytics – business intelligence software – into the mainstream. With that in mind, the company offers a centralized database running on commodity hardware along with front- and back-end tools used to access and consolidate data sources.

Competitors are offering in-memory approaches that are equally fast and simple to run. But SiSense argues that the in-memory approach can’t handle increasing amounts of business data. The In-Chip approach scales as the amount of data grows, the company claims.

The business intelligence software company reportedly raised $10 million in venture capital last year and is aiming for an initial public offering. Investors include Battery Ventures, the Israeli fund Genesis Partners and Silicon Valley investor Opus Capital.

Related Items:

SiSense Unveils Crowd Accelerated Analytics

Is “In-Chip” Data Processing the Next Revolution?

SiSense Announces Availability of Prism on AWS