DBTA recently hosted a webcast on NoSQL database technology to explain what it is and how these databases meet new requirements created by the surging data growth and concurrent demand for speed and analytics.
The first in a series of educational webcasts on NoSQL technology, the webcast was hosted by Tom Wilson, president of DBTA, and featured Warren Davidson, director of business development and strategic alliances at InfiniteGraph, a business unit of Objectivity; Darren Wood, InfiniteGraph lead developer; and Jonathan Ellis, project chair of Apache Cassandra, and co-founder of Riptano.
Setting the stage for the dialogue about this emerging technology, Wilson cited the high rate of data growth reaching levels not previously measured at DBTA. For example, he noted, a new study on data growth, conducted by Unisphere Research among IOUG (Independent Oracle Users Group) members, and underwritten by Oracle, found data growing at 9 out of 10 organizations and growing at rates of over 50% a year at 16% of the respondents' organizations, or one in six. Nearly two-thirds of the survey respondents' organizations have over 5TB of data and 20% report having more than 100TB of online data, Wilson added.
While explaining their organizations' technologies and where they fit, the presenters also sought to shed light on the larger picture of what NoSQL actually means and why there is interest in it now.
There is not a group of companies building NoSQL technologies that are trying to replace SQL technologies, emphasized InfiniteGraph's Davidson. Instead, they view technologies as not only SQL because there is a recognition when considering what is happening in terms of data management that there are requirements that may not fit into what SQL databases do well, and one of them is cloud computing, he said. "When you think about NoSQL, think about it in terms of people trying to put technology into the space that makes sense for that technology but not to try to take over areas that just don't make technical sense for the way that you would manage your data," he noted. The NoSQL landscape is commonly split among four bucket areas, including key/value pair, column databases, document databases, and graph databases, said Davidson. "For us, InfiniteGraph fits into the graph database sector."
In addition, it is useful to think about NoSQL technologies in terms of what problems they are solving, pointed out Ellis. There are three main categories of problems that people are tackling with NoSQL technologies -- scaling problems, speed problems, and the ability to use a data model that is a better fit for certain applications than a relational tabular model, he explained. "Cassandra is primarily interested in tackling the scaling model with a secondary emphasis on speed, and not really interested in bringing a new data model to the table," he observed. "It is still a row and column data model with a couple of twists." Listen to the full one-hour introductory NoSQL webinar on-demand.