There's a bright cloud on the horizon ... and it will transform the economics of BI - The Database Column

There's a bright cloud on the horizon ... and it will transform the economics of BI - The Database Column

Cloud computing is ushering in a new era of analytic data management for business intelligence (BI) by enabling organizations to analyze terabytes of data faster and more economically than ever before. The key change: It's delivered in an on-demand basis.

Organizations no longer need to justify spending hundreds of thousands of capital expense budget dollars for upfront hardware and software purchases or spend weeks waiting for hardware delivery and installation. Instead, they can sign up to tap into a computing cloud, such as Amazon's Elastic Compute Cloud (Amazon EC2), and have a dedicated, high-performance analytic database cluster provisioned and hosted for them. They can then use it on a pay-per-use basis, usually for a monthly fee.

This shouldn't be confused with software as a service (SaaS) models. Cloud customers are, in effect, renting dedicated servers and the people needed to house, secure, and manage them. These cloud offerings are more secure than multi-tenant SaaS models in which data from one customer may co-exist with data from another customer within the same application. Cloud customers have full control over server and firewall settings to ensure security.
 

Transforming BI

This alternative to traditional, in-house data analytics infrastructure will transform the economics of BI and open up many new possibilities for organizations of all sizes. I expect cloud-based analytics to impact BI in the following ways:

• New BI technology adoption will accelerate. The cloud will become the de facto platform for evaluating new software. The cloud enables software companies to make new technology available to many more evaluators on a self-service basis. Unlike free software downloads, evaluators are spared the time and expense of finding hardware and going through installation and setup and the other tasks required to get the software up and running. As a result, the adoption of new BI software technology should increase much faster than it has in the past.
  
  
• Organizations will conduct more short-term ad-hoc analysis. The need for data marts often arises suddenly, usually in response to new business conditions or events. The need may also last only a short time -- maybe just a few weeks or months. For example, a company might need to suddenly analyze manufacturing data in the wake of a quality or safety breakdown, or it may need a new price plan in response to a new competitor or market condition. The cloud gives companies a way to respond to these requests immediately -- get a mart created in a few hours or days, have business people slice and dice to their hearts' content for as long as they need to, then cancel the cloud cluster, and it goes away with no leftover hardware or software licenses. The cloud makes it economically feasible to conduct more of these short-lived projects.
  
  
• Lines of business will have the flexibility to fund more data mart projects. Because there are no long-term financial commitments required, lines of business can pay monthly cloud-based analytic database usage fees out of the operating expense budgets they directly control rather than going through a lengthy capital expenditure approval process. Companies can fund departmental, proof of concept, and ad-hoc analytic data projects on-demand, giving them the agility to respond to BI needs faster than their competitors and increase the quality of their strategy setting and execution.
  
  
• Data warehousing will increase within medium-size businesses. Despite their size, many midmarket companies have very large volumes of data they would like to analyze. Hedge fund companies with only a handful of IT people at their disposal need to analyze tens of terabytes of stock market history data to hone their trading strategies. Young bio-techs are in similar situations -- they have hundreds of gigabytes of genomic data to cull through. Cloud-based analytic databases will enable them to warehouse and analyze terabytes of data even though their BI budgets and staff are a small fraction of larger enterprises.
  
  
• The analytic SaaS market will develop faster. Companies that collect economic, market, advertising, scientific, and other data and then offer customers the ability to analyze it on line -- analytic SaaS -- will come to market faster and in greater numbers. They will be able to bring their solutions to market with much less risk and cost by basing them on the cloud during the early stages of growth. The companies can use the hundreds of thousands of dollars saved on in-house data center development to invest in customer acquisition, product development, and other market development activities. After the viability of the business model is proven, analytic data can be migrated to internal databases from the cloud if needed.
  

In order for these pioneering analytic cloud projects to succeed -- especially as data volumes grow -- they will require a database architecture that is designed to function efficiently in elastic, hosted computing environments like the cloud. At a minimum, such databases must include the following architectural features:

• "Scale-out" shared-nothing architecture to handle changing analytic workloads as elastically as the cloud
  
  
• Aggressive data compression to keep storage costs low
  
  
• Automatic grid replication and failover to provide high availability in the cloud
  

Technorati タグ: cloud,database,bi