Case Studies

Greenfield Online.  September 2005.

Problem:

Greenfield Online has multiple Windows 2003 clusters running SQL Server 2000 in an OLTP environment.   The access to the data was slow.   There was no method of creating an instant backup or near line backup device.   The analysis of this data on a daily basis was critical to the company.    Creating large cubes on the production server was impossible due to already peaked usage.   A performance hit could not be tolerated.   Moving large amounts of data for data mining and data analysis to another DSS server across the LAN would take hours.

Solution:  

Architected and implemented a 10 Terabyte SAN solution using EVA 8000 with dual redundant fabrics and Brocade 4/32 SAN Switches.  Installed and configured multiple HP DL 585 Opteron servers running 4 dual core processors and Windows 2003 Enterprise Edition with 64 bit extensions, and SQL Server 2000/2005 clusters.   Backups and near-line storage included an HP Virtual Tape Library 6510 with 6 Terabytes of virtual storage, HP MSL 6060 libraries with E1200-160 Network Storage Router cards and Veritas software.  Configured Replication Solutions Manager and wrote scripts to automate the creation of Snapclones and snapshots for both Decision Support Systems running cubes, and for backup purposes. The DSS system is running Microsoft Yukon/SQL 2005 Clusters on Itanium servers.  Designed the system for Continuous Access.  This solution has resulted in the automated transfer of snapshots and clones to the DSS servers and the automatic mounting/dismounting of volumes to backup servers for a true 2 Gb SAN based backup operation using RSM scripting.   In addition, there is a near line backup that exists on a HP Virtual tape library VLS 6510. 
 
 
  
Redesigned and in the process of migrating existing SQL Clusters running People Soft from MSA 1000 to EVA with zero downtime reducing disk queue lengths from 60 times maximum tolerable to acceptable  limits.   Merging SAN islands (MSA) into a single SAN environment.



XL Weather and Energy (a division of XL Global Capital which is a 38 billion dollar company) July 2003-2005. 

Problem:

The company was running simulations of derivatives which involved the import of millions of rows into tables for analysis.  During the process of the import, the servers would freeze and performance would be severely impacted.   Certain critical complex queries took six hours to run.  In general, the performance of the system was unacceptable to a rapidly growing company.  In addition, file server performance was unacceptably slow.

Solution:

Implemented a $350,000 EVA SAN solution resulting in over 500% improvement in performance.  Solution was designed and implemented following consulting services and troubleshooting.  The problems were identified and isolated.  The infrastructure involved several SQL Servers, development servers, and Novell File and Print servers.  The initial storage solution involved 1.4 Terra bytes of data and the use of snapshots and clones.   The backup solution for the SAN involved the configuration of a NSR card and a MSL 5026 tape library.  SSP configured at a hardware level for SAN security in addition to Zoning.  The solution was then designed and implemented successfully in August, 2003 with no downtime or loss of productivity.  XL Global had unsuccessfully tried to implement a 2.5 million dollar Hitachi SAN in another division for three years.
 
 
Configured RSM scripting to automate the transfer of snapshots and clones automatically to development servers for testing and configured automatic mounting/dismounting of volumes to backup servers for a true 2 Gb SAN based backup operation.
 
 
Implemented a HP NAS 4000 running Windows 2003 Storage Server for XL Global using SAN/NAS fusion resulting in performance benefits of 100%.  The use of this technology along with Alacritech TOE cards allowed optimization of file serving.  Network applications which could not run due to performance issues now work well.  Completed October 2004.  Zero downtime implementation.
 
 
Designed a “Disaster Tolerance” solution using Continuous Access by HP and data replication using both FC over IP and direct Fiber connections for XLWE.  February 2004.
 
 
 
HP Labs Rocky Hill

Problem:

Demonstrating Disaster Tolerance and Remote Replication Solutions to clients as “Proof of Concept . “ 

Solution:
 
Architected and configured a $2 million Disaster Recovery demo center for HP at Rocky Hill using Continuous Access.  The equipment involved two EVA 2C6D, a NAS 7000, an MSL 5030 tape library, an MSA 1000, and two fabrics.  This demo was used as “Proof of Concept” for Life Care Inc. and XL Global.  February 2004.