Alexander Forbes, a provider of financial and risk services, has recently completed a R7 million upgrade of its disaster recovery environment. Scoping and implementation of the enhanced DR system was done by Dimension Data with key storage solutions supplied by Drive Control Corporation (DCC).
Says Dion Fowles, IT Security and Compliance Officer at Alexander Forbes: "Our business continues to grow and we had to replace much of, and enhance, our DR systems to keep pace with the expansion of our production environment."
The company currently backs up 60 TB of data on a daily basis and has 500 servers, 200 of which are mission critical. The systems which the company uses are mostly custom developed.
Alexander Forbes was outgrowing its existing DR capacity and needed to secure its servers, data and tier one applications. Different recovery windows, ranging from 5 minutes to 48 hours, were specified. Dimension Data was contracted to scope and set up the DR infrastructure at the new host site, also consolidating the new and existing DR systems.
The DR environment was designed to be a mirror image of the live environment, albeit on a smaller scale. Says Grant Morgan from Dimension Data: "We determined the bandwidth, linking and power requirements and then planned and built the data centre at the Alexander Forbes' new Telkom NNOC DR site.
DCC delivered the solution to the site in Pretoria and Dimension Data handled the implementation. This DR solution is one of the most complete systems we have implemented to date. It is a fully integrated first-world best practice system comprising VMware, a SAN, numerous blade and Unix servers, Overland NEO 2000 and 8000 tape libraries, Overland REO 4000 disk-based storage appliance, and Veritas' Netbackup and Command Central for centralised management of backups and storage resources. In particular, the use of virtualisation was critical to meet stringent recovery requirements."
Notes Fowles: "Our current architecture was duplicated in the DR environment and since virtual server technology was being deployed in our live environment, this was mimicked in the DR environment to enable us to switch over easily if needs be. Time is of the essence and virtualisation enables a server to be built in minutes through automation of the deployment of the operating system to the server. In contrast, manually building a server can take many hours, not to mention the six to eight weeks we would have to wait for specific equipment to be delivered."
Server virtualisation enables Alexander Forbes to manage approximately 200 to 300 servers on just 16 blades. The numbers are significant. Notes Fowles: "In a worst case scenario, if both our key buildings collapse 212 key production servers need to be brought back online via the DR site."
The company has ensured it has sufficient disk and tape storage, as well as the software needed to ease management of backups and storage resources.
Says Dineo Mosupye, Overland product manager at DCC: "The Overland NEO 2000 tape backup and archive solution provides up to 24 TB of native capacity in just 0,34 square metres of floor space, while the NEO 8000 packs up to 800 TB of capacity in just 0,64 m² of floor space. The NEO 8000 can to scale two units together horizontally, providing access to as many as 1000 LTO tape cartridges and 24 tape drives."
"The Overland REO 4000 disk-based backup and recovery appliance with 30 TB of disk capacity adds speed, reliability and flexibility to the data protection environment. Used primarily for replication from the old DR site, it will enable Alexander Forbes to make use of the tiered disk-to-disk-to-tape backup methodology which enables better lifecycle management and faster recovery of critical data. The REO 4000 features innovative Protection OS software that delivers volume and device virtualisation, management and connectivity capabilities."
Veritas NetBackup 6.0 offers a single console for management of all backup and recovery operations on UNIX, Windows, Linux, and NetWare systems. It enables centralised management of the entire backup and recovery environment, including realtime monitoring, historical reporting, alert management and troubleshooting. The Veritas CommandCentral Storage 4.3 storage resource management solution enables capacity management, centralised monitoring, application to spindle mapping, and active management of storage components, increasing data availability, optimising resources and reducing capital and operational costs.
It has taken six months working to a tight schedule to get the new DR system installed. Says Fowles: "Challenges included getting the equipment into the country and the implementation commissioned to Dimension Data at the new host DR site we chose, namely Telkom's NNOC.
However, our new DR site is up and running and testing has begun. It has been a long haul and the dedication of Dimension Data project managers and technical staff must be applauded."
Alexander Forbes will look at expanding its virtual server environment and adding in more disk capacity to accommodate the growth of their data and systems.