Clusters have emerged as the most cost-effective solution to design high performance and dependable data centers, which are increasingly being deployed for a wide variety of Web-based services. However, data centers contribute to a significant part of the overall delay, which is likely to grow with the increasing use of dynamic Web contents. Furthermore, (24x7) server availability to host critical services appears to be a far-fetched objective. Therefore, design of high performance and dependable data centers has become a critical issue.
The objective of this research is to investigate the design of a three-tier data center consisting of a front-end Web server, a middle-tier application server, and a back-end database server using the industry standard, InfiniBand Architecture (IBA) communication paradigm. A prototype data center will be implemented on a 96-node IBA-connected Linux cluster to investigate three research issues using real applications and benchmarks. The issues are quantifying the performance benefits of using a user-level communication technique such as IBA instead of the standard TCP/IP stack, proposing new performance enhancement techniques for data centers, and characterizing the impact of different faults on server dependability.
The research will enable effectively utilizing the IBA communication mechanism for designing multi-tier data centers, quantifying the end-to-end performance benefits of using IBA, developing fault-tolerant mechanisms for highly available data centers, and understanding the design tradeoffs between high performance and dependability. In addition to fostering new research directions in several areas of data center design, the outcome of this research is likely to have significant commercial interests.
