RDISK prototype

The RDISK system can be seen as a specialized cluster for information retrieval in large datasets. It has been developed primarily for the extraction of genomic information, but other types of data can be manipulated without restriction. The key idea is to be able to rapidly scan large datasets, and extract on the fly any useful information for further, more complex processing. The following two examples illustrate our point: alignment search and pattern search.

As outlined above, an alignment can be searched for on the basis of an anchor point corresponding to a W-character word present both in the query sequence and in the bank sequences. This detection can be carried out very quickly, as the data is read from the storage medium. All bank sequences...