Background and Aim: VDW system architecture varies from one institution to another. This provides an opportunity for the development of best practice in system architecture design. The word architecture describes the underlying structure of networks, computer hardware and software. Well-designed systems-level architectures are critical to the success of large-scale projects. The explosion of technological opportunities and customer demands has driven up the size, complexity, cost, and investment risks of projects. Without solid architectures, these projects lack the firm foundation and robust structure on which to build. Solid systems architecture and design is a required phase following data gathering and analysis which should precede any design prior to any implementation. With functional requirements established, the system architect can design a system to perform in an optimized environment, ultimately producing robust data and customer satisfaction. Architecture choices Hardware platform DBMS –Relational Database Management System (RDBMS)Query/analysis tools Maintenance/ service strategy. For bioinformatics research, the most important qualities of system architecture are: extensibility for new functions/features; interoperability for cross-site collaboration; security due to regulatory and ethical requirements; serviceability/maintainability to support independence from enterprise/business computing; and performance for analysis of large datasets. Aim of this study was to determine best practice architecture choices resulting in performance improvements per institutions data volume. Methods: Performance benchmarking analysis was gathered from KPNC and S&W comparing the effect on VDW performance of different design choices and implementations of database software and hardware. Our performance measurement focused on the delivery speed of data return. Measurements were made on various data areas at both peak and off-peak times. Results: KPNC has measured via benchmarking on VDW tables that Oracle DBMS can increase performance compared to SAS datasets. KPNC was able to measure separately the performance improvement from migration to a more powerful hardware platform. Conclusions: A proactive, best practice approach to systems architecture can create benefits for research databases, particularly in performance speed. Improvements in performance can be achieved through software, by adopting a relational database management system such as Oracle, that allows table indexing and partitioning, and through investment in high-performance hardware. The two strategies can be implemented alone or in conjunction for maximum benefits.