NoSQL Databases in Big Data Management is Increasingly More Commonplace

To gain a competitive edge more and more companies are building externally focused mobile and social applications. These applications complement their legacy transactional systems. They involve Big Data that doesn’t lend itself well to MySQL style relational database (RDBMS).

A variety of NoSQL (not only SQL) databases are complementing and or consolidating data from existing RDBMS applications. The importance of these NoSQL databases are growing within the Big Data Management realm as the underlying technology is more relevant to the insights companies are seeking from Big Data today. Success stories from early adopters of these NoSQL databases are providing confidence that insights from Big Data can be realized relatively quickly and economically.

Why NoSQL instead of RDBMS?

When using NoSQL databases, data can be messy, unstructured and connected (social media). New data types can be easily accommodated to make the applications more insightful and useful. There are no schemas to mess around with and hence no application downtime. These are typically more modern web scale databases that power web applications used 24/7. Since NoSQL databases are distributed by nature, they can be scaled out easily and cost effectively through clusters of commodity servers. This allows the application itself to scale and accomodate more users and usage scenarios.

Success Stories

Metlife in the insurance industry used NoSQL (MongoDB) to build a Facebook wall like application for it’s call centers and claims admin staff. The application allows call center agents to more efficiently and quickly answer customer questions on coverage and claims. Data from 70 different databases covering approximately 45 million agreements (entire US customer base) was integrated for this application. They have been trying to do this with RDBMS for several years but completed the application using NoSQL in just three months.This application runs across six servers in two data centers and stored 24 terabytes of data at the time of release. You can read the entire story here.

As Netflix moved into the cloud, they adopted multiple NoSQL solutions (SimpleDB, Hadoop/HBase and Cassandra) and moved beyond RDBMS to achieve their goal of building fast, fault tolerant systems at internet scale. Each NoSQL solution was a best fit for certain set of use cases and contributes to Netflix’s goal of  achieving infinite scale and building the leading global content streaming platform used by over 33 million subscribers in over 40 countries. Today Netflix stores 95 percent of all its data in over 50 Cassandra clusters with over 750 nodes.  The data includes customer account info, movie ratings, metadata, bookmarks and logs. You can read the entire story here and here.

Companies like Cisco, Craigslist, Goldman Sachs, Intuit, Carfax, Constant Contact, Disney, Comcast and many others also use NoSQL.

Understanding the NoSQL Landscape

As with any new technology, implementing NoSQL involves a learning curve that includes evaluating and understanding several technical and business aspects – What are the most common NoSQL stores? How are their architectures similar or different? How do you choose the correct NoSQL database for specific use cases? etc.

In order to cover some of these topics, Inferology is now offering a one-day training course titled “Understanding the NoSQL Landscape”. This survey course is targeted towards both technical and non-technical people who want to understand the emerging world of Big Data. The class is divided into the following six sessions – Intro to NoSQL, Hadoop Introduction, Column Family Face-off: HBase vs. Cassandra, MongoDB overview, Neo4J overview, New Horizons. The students will be given a ~100 page slide deck which can be used as reference material after the course. A PDF printout of the specific examples demoed in the 6 live demos will also be given to students. Interested students can re-run the demos in their own time after the course using the PDFs.

Inferology is offering this course on July 29, 2013 at StampedeCon, a first national Big Data conference of it’s kind, held every year in St Louis, MO. Providing education for both novice IT professionals and Big Data veterans is at the heart of this conference. Several interesting success stories from companies like Walmart, Ford, Riot Games, Cerner, Monsanto and others will be presented here.


Infer-ology /inˈfəräləjē/ - The Art and Science of Inferring What Matters.