Knowledge Integration Dynamics

Digging for gold - Get more out of what already exists at a disaster recovery site.


By Mervyn Mooi, Director at Knowledge Integration Dynamics. (KID)
[Johannesburg, 12 June 2012]

The disaster recovery (DR) site is a potential goldmine of BI productivity. Traditionally used to backup operational systems in the event of the primary live operational system breaking or becoming unavailable for whatever reason, its main objective used to be to ensure business or operational continuity, even if there was a pause in operation, sometimes being hours or days. Today, businesspeople rightly demand zero downtime and high availability (HA) of systems, to the extent that users don't even know that a break has occurred. Dedicated, real-time links to the primary site have enabled a shadow copy to be kept synchronised all the time.

Employing the DR site for BI offers huge advantages to the business, a few being:
* Load balancing of data processing and querying between the primary and secondary (DR) systems;
* Using the DR site's data as the source for other downstream applications, such as operational data stores, data warehouse, and data marts; and
* Allowing virtual metadata-based models, captured in front-end tools, to be placed over the DR system databases for BI applications, given the appropriate processing power and capacity of the DR servers.

Load balancing adds processing capacity, reducing the processing work or load on the primary servers, also known as nodes, as input/output, memory and central processing unit churn are spread over two or more nodes. Load balancing also affords IT departments the opportunity to fully leverage existing investments, such as those in the DR infrastructure, by utilising every server to create a larger networked computing resource. Processing uptime, or HA, is increased as there is backup of processing resources on the DR site and vice versa. Using the DR resources to load balance and increase capacity greatly increases data processing speeds and the time to delivery is rapid. It reduces or eliminates latency problems usually experienced by BI end-users when running queries and generating reports. Greater processing speeds may also eliminate the need for data marts and aggregated cubes, and eliminating that layer improves resource economy even further. HA also creates improved service level delivery, which in turn improves stakeholder confidence.

In fact, technologies such as storage area network, network-attached storage, redundant array of independent disks, services-oriented architecture, together with the cloud, provide a global network of computing resources where servers are inter-operating with each other to ensure HA at the very highest level.

Doesn't it increase risk?

Some may argue that using the DR site for HA and load balancing puts the business at risk. How is the DR site supposed to accept a sudden increase in load when the primary system fails? That too employs load balancing techniques and simply shares available resources to support operational applications. Risk is increased when the primary site is not shadowed, which means it isn't synchronised in real-time. Risk is also increased when the DR server or servers are used to shadow more than they're capable of supporting, but that's not an issue of making use of idle resources.

Businesses that administer critical and important applications that require uninterrupted services, such as the cellphone networks, for example, or the banks with automatic teller machines, terminals and client server systems, benefit enormously from HA and load balancing. Real-time monitoring and reporting applications also benefit, examples being vehicle and people tracking systems used by the cellular service providers, the police, and vehicle tracking businesses.

While some businesses already employ their DR sites for HA and load balancing – but many larger organisations still don't – many medium to small businesses haven't had the resources to do so in the past. It has typically been the domain of the larger organisations with the finances and technical expertise necessary to deploy and maintain the requisite infrastructure. The cloud, however, has changed that. It has eliminated one of the biggest issues, being the cost of infrastructure, because smaller organisations no longer need to own all the technology and have the skills to keep it running.

comments powered by Disqus