We are pleased to announce that Principled Technologies has acquired Demartek.
Demartek Evaluation of High Performance MySQL Cluster Reference Architecture using 16 Gbps Fibre Channel and SSD Technology
30 May 2012
As CIOs and datacenter managers consider increasing stresses on existing computing infrastructure and the seemingly insatiable demand for storage and rapid access to that storage, new infrastructure concepts must be investigated. The combination of recently announced 16 Gbps Fibre Channel (16GFC) infrastructure and Solid State Drive (SSD) technology offers a compelling way to significantly improve performance, lower costs and dramatically condense server and storage infrastructure footprints.
Many data centers today are struggling to maintain their SLAs for database performance because of conflicting workloads such as transactions and large data warehouse queries, un-optimized database code and a host of other reasons. One of these causes can be that large database server clusters, such as those running MySQL, are most often built with local disks and large amounts of system RAM in order to get maximum performance. This architecture, while somewhat simple to design, has huge risks with it that can impact performance, scalability and availability. For example, if these local disks fail, an instance of the database server may go down. In addition, this type of architecture can be scaled in only a linear fashion, which severely restricts the number of users per cabinet or rack, and the size of the database per server.
Advances in technology are now available which can help overcome these limitations and dramatically improve performance, scalability and availability. With the advent of 16 Gbps Fibre Channel from Brocade and innovative SSD technology from GridIron Systems, the limitations of the existing database cluster architectures can be overcome, servicing significantly more users per cabinet, growing the maximum database size for each server, while significantly lowering capital expense and operating expense, including server costs, power and cooling. In some cases, the footprint savings can be significant, reducing many racks of equipment down to one, while dramatically increasing database transaction rates.
