Storage Hypervisor boosts charity’s data mining speed by 300 Percent
DataCore Software today announced that the British Red Cross Society has deployed the SANsymphony™-V Storage Hypervisor to provide a significant performance acceleration on its new Online Analytical Processing (OLAP) system, dramatically shortening response times and increasing the reliability of data extraction. The performance improvements were achieved by installing SANsymphony-V on an HP Proliant DL 370 server. The DataCore™ software has reduced the time window needed to perform the Extract, Transform and Load (ETL) operations from an average of 12 hours down to four, with the load spread across half the original number of internal hard disk drives achieved through the efficiency of SANsymphony-V’s thin provisioning.
The British Red Cross Society is the United Kingdom’s registered charity arm of the worldwide humanitarian organization, the International Red Cross. Formed in 1870, the Red Cross has over 31,000 volunteers and 3,300 staff providing assistance and aid to all people in crisis, both in the UK and overseas, without discrimination and regardless of their ethnic origin, nationality or religion.
“In order to sustain the Charity’s considerable ongoing work worldwide, the Red Cross needs to continually generate additional income from new and existing donors,” said Kevin Bush, technical architect for the Charity’s MIS Enterprise Architecture Team in London. “It is our function in MIS to ensure the relevant departmental units have the appropriate infrastructure available to allow them to complete automated processes in time to fulfil marketing campaigns to drive further donations.”
To help facilitate ongoing fundraising, a new suite of hardware and business intelligence tools were deployed six months ago for the British Red Cross utilizing OLAP – an approach that swiftly answers multi-dimensional analytical queries through accurate Business Intelligence (BI) tools deployed on British Red Cross’ SQL Server database. BI data marts are created to track behavioral changes, creating campaign relevancy trends for business units. This level of data profiling, specifying individual campaigns with matched targets, entails significant I/O (Input/Output) processing demands and depends on a stable, optimized infrastructure.
Working in conjunction with the MIS Enterprise Architecture Team, the British Red Cross’s partner, Adapto, recommended that deploying DataCore’s SANsymphony-V software would significantly decrease I/O strain and dramatically increase performance in a cost effective, non-invasive way. The SANsymphony-V storage hypervisor could dramatically improve performance levels by increasing the speed of read/write requests across the entire British Red Cross storage infrastructure using the storage server memory as the caching engine. This caching could dramatically accelerate application response times, manifesting in a dramatic increase in the speed of database queries and data extraction for the business units.
Critical to the effectiveness of the Extract/Transform and Load (ETL) from the database is achieving ongoing consistency within a predefined extraction window. The speed of I/O to process workloads determines these two factors; a slow I/O equates to a long and erratic extraction window. In practice, prior to the performance caching gains, each ETL was taking between nine and 15 hours, being set to run overnight with the resultant data marts ready in time for the next working day.
Following Adapto’s suggestion, Kevin downloaded the easy to install SANsymphony-V test drive and right away ran a test ETL that displayed immediate benefits through DataCore’s mega caching ability, with the software recognizing I/O patterns to anticipate which blocks to read next into RAM from the back-end disks. Requests became fulfilled quickly from memory at electronic speeds, eliminating the delay associated with the physical disk I/O. The findings were impressive. The production-ready, easy to use GUI allowed the ETL to perform at a blistering pace, similar to that achieved by SSD but without the associated cost overheads. This manifested in a shorter four hour query extraction timeframe.
“From the point of evaluation onwards, we haven’t looked back with SANsymphony-V,” said Bush. “It’s caching and performance acceleration has certainly addressed the consistency of extraction, whilst reducing the window to an acceptable level, so that as a Charity, we can concentrate on effective fundraising to help those most in need. We are so impressed that we are now looking at installing another node of SANsymphony-V for high availability and mirroring.”