Principal Investigator C Dewey
Singapore-MIT Alliance (SMA) researchers are developing a new system for scientific database management that may revolutionize how datasets of the future are handled.
The system. dubbed Da Vinci's Notebook, will make it much faster and easier for scientist to search, store, explore, query and visualize their data. White conventional Relational Database Management Systems (RDBMS) are well developed for the business world, those systems are said to be a 'flop' In the scientific community. This is due to several key reasons.
Firstly, the cost of learning the tool is very high due to the need for mastering unknown query languages (such as SQL XQuery) and dealing with complex and often unknown structure of databases (schema). Secondly, traditional RDBMS do not provide efficient support for ontologles. complex data structures (such as XML), data provenance, and annotation management. Finally. often many scientists view databases to be too slow for processing large volume of scientific data.
Da Vinci's Notebook will have several advantages over RDBMS. First, it will be completely intuitive and require zero programming knowledge, thus addressing the problem of a steep learning curve for scientists. Next, it aims to provide efficient support for complex data such as XML ontologles, annotations, and support data provenance. Lastly. it starts fetching data even while the scientist formulates his query, unlike conventional systems which often remain idle at the time, leading to a boost In performance.
Da Vinci's Notebook potentially improves performance by 200 times. The research team's tests showed that a typical query performance improved from 83 seconds in a traditional framework to a blazing 0.4 seconds in the new system.
Scientists will appreciate that they no longer have to learn programming and manually type in code just to gel the data they want, freeing them to focus on what they do best. Looking ahead, Da Vinci's Notebook can potentially be applied to business and commercial applications too -- wherever there is a huge amount of data involved.