"According to our own internal testing, Sybase is offering a 30-40% performance benefit over competing vendors on the same platform."
BioData Platform Manager
Founded in 1982 as a spin-off of the University of Ghent, Belgium, Bayer BioScience is one of the main innovation and research centers for Bayer CropScience. It is now a division of the multinational chemical, medical and biotechnology concern Bayer AG. With its frequent mergers and acquisitions, Bayer BioScience required seamless evolution and future-proof software development, along with a means of meeting the exponentially growing performance requirements caused by scaling local operations into a multinational group.
- Bayer's Sybase solution ensures seamless synchronization and extensive consistency checks that monitor the purity of the data, allowing researchers to conduct studies and collaborate internationally.
- Improves data quality and tracking of biological material
- Enables migrations and platform redesigns without changes to the data model
- Reduces highly complicated ad-hoc queries from 20 minutes to 20 seconds
- Life Sciences
Moving the Boundaries
Over the past 10 years Bayer BioScience has seen constant expansion from a once small-scale operation to today’s multinational environment. During this evolution Sybase’s database platforms and development tools have proved their versatility, flexibility and considerable scalability. Bayer BioScience is still constantly moving the boundaries and is now delivering critical database applications in the field. With Sybase iAnywhere technology, observations are no longer recorded using pen and paper but are input directly into a local copy of the database. As a result, researchers’ efficiency is enhanced while the likelihood of errors and inconsistencies is substantially reduced.
Consistent, Flexible and Scalable
The former university research laboratory, which has built up unique, worldwide expertise in improving seed, plant and biological materials, has been taken over by large international organizations no fewer than three times in the past 10 years. With each of these transitions, Bayer BioSciences central IT applications were suddenly linked to those of an international group with numerous users and sites in several countries. This created a challenge to manage data across mixed environments and more complex infrastructures. The IT team needed the right solution to address each new adaptation
“Scalability was one of the key reasons why we initially opted for Sybase,” said Pierre Lens, BioData Platform Manager at Bayer BioScience. The data model developed at Plant Genetic Systems is still largely intact – albeit with a great deal of additional data. Later, the company opted for PowerBuilder, and shortly after, PowerDesigner, to develop its front-end applications. This decision is still paying off, as the latest generation of mobile front-end applications was rewritten in record time. “What took us ten working days to develop in Visual Basic, I rewrote in PocketBuilder in one day,” confirms Lens.
The key driver for mobile database applications at Bayer BioScience is ensuring consistent data input. It is crucial to know the detailed background of seeds and plants used in experiments. Thanks to the BioData platform a mass of data is recorded, stored, and made available for interpretation in the field. The quality of data input is monitored via convenient drop-down menus. This access to detailed data on the experiments and materials requirements is critical when applying for patents and demonstrating ‘prior art.’
Mobility Meets Data Quality
The latest generation of the BioData platform links dozens of user applications in 26 locations worldwide. This is achieved through a three-part structure: the central database (Adaptive Server Enterprise), the local database servers and the mobile database (both SQL Anywhere). An “always available” architecture was deliberately chosen because in some situations a permanent connection between the different servers cannot be guaranteed and to ensure absolute data quality.
A copy of the data set required for each experiment is stored locally and the scientists can work on this at any time — in the laboratory, in the greenhouses or anywhere in the field. The Sybase solution ensures seamless synchronization and extensive consistency checks, which monitor the purity of the data. “We didn’t opt for a real-time solution, based on wireless LAN for example, because of the extensive monitoring of data we require at the database interface,” says Lens.
The mobile solution is equipped with a barcode scanner for fast, accurate identification of plants. The data set for each experiment is set up by the scientists themselves: they select the columns and fields to be used from a list. “The entire environment is dynamically conceived. Nothing is hard-coded,” stresses Lens. However, he continues, the front-end input screens are only the tip of the iceberg. “The supervision of data exchanged with the back-end is where Sybase makes a real difference.”
Evolution, integration and performance
Due to its dynamic multinational operating environment, Bayer BioScience needs a flexible IT infrastructure. Other parts of the Bayer group use different platforms and needs change continuously. “It’s vitally important to anticipate change. We don’t always have to stick with Sybase; the important thing is the openness and the modular structure of our environment, allowing us to link up easily with other platforms,” stresses Lens. In short, the IT group could have their pick of technology solutions. Nevertheless, Sybase remains the preferred platform: “According to our own internal testing, Sybase is offering a 30-40% performance benefit over competing vendors on the same platform. Speed is still the prime demand made by our users,” says Lens.
Performance is an extra challenge for some highly complex search tasks that may link data from tens of tables, and for this, Bayer BioScience uses Sybase IQ. “We make a copy of the relational database in Sybase IQ for specific search tasks. Sybase IQ allows us to reduce the processing of highly complex queries from around 20 minutes to less than 20 seconds,” concludes Lens.