File Server Auto-Classification

Our Professional Services for Automated Content-Based File Classification

for File Classification Infrastructure (FCI) and Dynamic Access Control (DAC)
in Windows Server 2012

We classify the text content of your existing file repositories automatically (bulk classification). We provide the service at your location or ours. Fully or partially, depending on your security needs. Then, based on our classification results your IT-security expert will automatically generate individual access rights per each single file of your repository.

It can take a while to process millions of text files, automatically generate classifications and apply the individual access rights based on a Central Access Policy (CAP). Better to start preparations earlier rather than later, even if your Windows Server 2012 has not yet been deployed ...

All right, and your future files will automatically be classified also. You simply use our classification engine (having been trained during processing your legacy files) in your daily business.

With our "d-Class" technology we can automatically analyze the text content of a file and assign an appropriate class (category). Knowing the class of a file - for example, order, invoice, identification letter, termination agreement – one can generate the access rights for that file based on a Central Access Policy (CAP).

You could, for example, derive a file permission from content which has automatically been inferred as belonging to class CONTRACT, so that only employees of the commercial department have access to it. It could even be taken further, so that for instance internal company agreements can have separate access rights to customer agreements...

DocStream GmbH is primarily a technology manufacturer for content analysis technology. In recent years, our team of mathematicians and classification experts has developed a high quality text-based classification technology (SVM - support vector machine). The content classification of text repositories has already been successfully tested in a pilot project of 3.8 million files and 30 classes.

more Information here or contact us info(at)