By Phil Jones
Sand Technology Systems International Inc, the overseas arm of Montreal, Canada-based data warehousing analytics developer Sand Technology Inc, has kicked off a new marketing campaign by announcing its first European direct sales push since setting up shop in Potters Bar in the UK in April.
Even before ramping up its marketing and sales effort fully, Duncan Painter, the company’s European general manager says Sands’ core Nucleus Exploration Warehouse has chalked up seven early local customer wins, bringing the product’s global installed base to 47. Now, he says, the company is tightening its focus for a direct sales push against the burgeoning European telecoms service provider sector, where he claims Nucleus’ ease to use, complex querying capabilities will make it a hot ticket product.
In fact, Nucleus is just the latest of a raft of analytic engine tools that have been pointed at Europe’s mobile and new carrier operator community this year. However, unlike Tantau’s Infocharger, which is touted as an in memory processing solution to big data set crunching, or Norkom’s Alchemist, a neural processing approach to data grooming, Nucleus is positioned to be a ‘quick and dirty’ (if still very powerful) tool for throwing ad hoc complex queries at constantly changing data sets.
Painter claims Nucleus’ column-oriented database engine combines the ease of use of Microsoft Access, with some of the scalability and complexity processing capabilities of major mainstream products. This makes it more accessible than most on-line analytical processing tools (OLAP) and much quicker to implement, he claims – making it suitable for one-off marketing exercises where the time normally spent cleaning data and setting up the analytics legislate against using higher end tools.
Sand also claims there is little performance compromise in choosing its Exploration Warehouse, and in some circumstances it may be able to out-perform rival products. In terms of data loading for instance, Nucleus is claimed to use a proprietary technology which enables this to be done online without interfering with queries currently being run by other users. A bit freezing technique means that current queries are able to continue against their original data set, while new data is loaded ready for a different query, Painter said.
The volume of data held in the warehouse can also be controlled by Nucleus’ characteristic of not indexing redundant data. In most circumstances, Painter claims this produces a smaller data set than would normally be required when using a conventional OLTP engine. In one recent bench test, he said, a 250GB data set employed by a rival 16-way massively parallel processing (MPP) system was condensed to 64GB when executed against a Nucleus system running on a four-way Intel server.
