We IT professionals have gotten used to putting together an ROI for a proposed project or initiative. We sometimes struggle with coming up with an ROI that makes sense to the person controlling the money…the CFO. Our analyses are sometimes said to contain too many soft returns…measurements like quality. Quality is a wonderful thing, but rarely does it positively impact the bottom line.
The latest innovations in business intelligence offer the opportunity to deliver a solid ROI to the business and if approached holistically can show positive improvements across the board.
Long have users of business intelligence technologies complained about the lack of speed in getting answers to their questions. In many cases users never get answers to their questions, but instead learn to navigate slow BI environments by customizing their crucial questions into yet smaller and smaller queries with barely acceptable response times.
IT staff members have tried techniques to reduce this burden like caching subsets of data, creating indexes or aggregate tables in the relational database…which also contains a subset of the data that a user actually access to. Ingenious and hardworking as the IT may be, this type of effort is expensive and not proper way to make use of your talented and limited IT resources.
Business Intelligence is a rapidly evolving discipline and its latest evolutionary step (in-memory analytics) is a major advancement towards achieving a higher, industry wide adoption rate of the technology. One (but not the only) of the key advantages of in-memory analytics is speed. The size and complexity of the query/information are not impediments to the increase in speed users of in-memory analytics will experience.
The pace of business today demands faster access to information and easy analysis, and usage of tools that do not require extensive IT hand holding. This is a positive development for both the business user who will gain near self-service analytical capabilities and for IT staff which can decrease the amount of time spent on query customization, OLAP building, and other performance tuning tasks.
So how does it work you may ask…the key difference between conventional BI and in-memory analytics is that with conventional BI the user runs a query against a typical data warehouse. The query goes to the database that reads the information stored on the data warehouse server’s hard disk.
With in-memory analysis, all the information is initially loaded into memory. If the in-memory tool is server based, an administrator may initiate the load, if it’s a desktop tool, the user can initiate the process on his or her workstation. Users then query and interact with data loaded into the machine’s memory. Accessing data in memory is millions of times faster than accessing from a disk providing the user exponentially increased query and reporting processing speed.
The decreased costs in memory have lowered the costs barrier that initially held back this innovation. 1GB of RAM in years past costs $150. Today it will cost ~$35. An analytics server built at the earlier price point would run you ~$64,000, where as today it will cost ~$13,000.
In-memory innovator QlikTech has seen a major increase of in-memory deployments with a year over year increase in its sales by 80%. Other major vendors such as IBM with it’s Cognos TM1 solution, Microstrategy, TIBCO Spotfire, and SAP have begun to take notice of the market’s demand for this new way of analyzing data and making quicker, informed business decisions.
In-memory analytics offers the capability of decreasing an organization’s time to value, increasing the adoption rate of Business Intelligence users within your organization by simply making it easier to use, and decreasing your internal IT costs by either reducing the amount of IT necessary to deploy and maintain the environment or allow you to make better use of your scarce resources. All of those are compelling reasons to explore its feasibility within your organization…at a minimum.
