2007 in review: scaling business analytics applications.

The growing reliance on business analytics to improve strategic operational decision making and increase an organization's competitive advantage is driving an explosive market. IDC estimates that the overall market for business analytics development tools and packaged applications will top $20 billion for the year 2007. Furthermore, companies are aiming to expand the reach of their business analytics or business intelligence applications within their organizations:

According to Louella Fernandes, Principal Analyst, Quocirca, in August 2007, 'Various research projects have shown that business intelligence (BI) continues to be one of the top investment priorities for CIOs. But what is changing is its shift in focus from complex tools for a few users to more flexible, affordable and accessible tools for a larger audience. As such, BI is moving away from being an exclusive tool for power users, or 'information producers', to empowering the 'information consumers' in accessing, analyzing and sharing data.'

In the quest to make business analytics and intelligence applications more accessible to more users, companies often spend unnecessarily on underlying infrastructure to support these goals.

Too many users simultaneously analyzing a single data set can cause contention and slow application performance, particularly when the underlying infrastructure relies on a disk-based storage system. The impact of simultaneous use simply brings the system to a crawl and this can cost organizations substantially in terms of lost revenue opportunities. For example, if bank tellers are bogged down by 'screen freeze' or glacial system response time when trying to up sell a customer new financial services they are not familiar with, the sales opportunity window could close at the point of contact. Making these applications readily available to more users typically requires copying or replicating the underlying databases.

This adds cost and complexity in two areas. First, the replication or copying of databases takes time, adds infrastructure, and requires new copies for each additional group working on the data set. Second, analytics applications working on a copy of the primary database are inherently out-of-date due to the real-time nature of most databases. But working on the primary database increases concurrent use, often a recipe for further application slowdowns, so copies are often the only choice.

Now new solutions based on centralized storage caching allow more users to concurrently access a single data set and increase overall performance without requiring multiple database copies. By caching frequently accessed data in high-speed, high-capacity memory, scalable caching appliances enhance existing storage systems to deliver massive simultaneous access without requiring additional copies. The use of cache also enables real-time analysis not possible with traditional disk-only systems.

By operating within the network and connecting with IP and Ethernet, scalable caching appliances deliver high I/O rates with ultra-low latency to any client from any storage server or data set. This enables the dynamic allocation of cache resources across multiple systems and applications' delivering dramatic performance improvements data center wide. This network-centric approach is ideally suited for NAS and NFS environments across new and existing NAS users looking to boost performance for key applications.

By Gary Orenstein, Vice President of Marketing, Gear6

www.gear6.com