Data Analytics Research
NoSQL in Today’s Organizations
NoSQL, meaning “not only SQL,” is way in which data is stored in a non-relational format. It is intended to handle very large scale and distributed data in a relatively inexpensive way. Facebook, Twitter, LinkedIn, Amazon, and Google all use NoSQL somewhere within their business to handle the mass amounts of data they receive every second. NoSQL resulted from several needs of database managers. One was a need for more storage space. The amount of stored data has risen 500% between 2006 and 2010, a trend that is likely to continue. There was also a need for storing interconnected data that had very complex structures. Businesses wanted these functions without giving up system.
NoSQL has to do with our MIS 2502: Data Analytics course in that NoSQL is another way to store, manipulate, and examine data. In the course we learned about MySQL and how to extract data. There are a few similarities between the two such as the need to query. NoSQL is a different type of database management system. Nevertheless, it is still useful for analyzing large amounts of data in order to solve problems and enhance business developments.
For LinkedIn, managing and storing data is a big operation. Data originates from LinkedIn shares, tweets, profile information, and much more. In 2010, they implemented Scala, JRuby, and Voldemort, all types of NoSQL data management systems. Scala and JRuby were useful for fast ad hoc operations. Voldemort was chosen because it fit LinkedIn’s current key-value store query access pattern (Harrison 1). The most challenging part of their process was in implementing a search system. After this, LinkedIn had to adjust specific of the ways in which their data was ranked. Problems arose when “share count” consumed too much weight in the ranking formula and “spam” data became inaccurately measured. By adjusting the formula, results became great. Scala, JRuby, and Voldemort especially required time to mature with LinkedIn’s systems. However, these systems seemed to have been preferred by analytics team at LinkedIn in term of design, testing, and stability.
References:
Harrison, Guy. “10 Things You Should Know about NoSQL Databases.” TechRepublic. CBS
Interactive, Inc., 26 Aug. 2010. Web. 18 Nov. 2012.
Perdue, Tim. “NoSQL: An Overview of NoSQL Databases.” About.com New Tech. About.com,
n.d. Web. 18 Nov. 2012.
Synodinos, Dio. “LinkedIn Signal: A Case Study for Scala, JRuby and Voldemort.” InfoQ.
Contegix, 11 Oct. 2010. Web. 18 Nov. 2012.