From the Blogosphere
NoSQL and the Enterprise Data Hub | @BigDataExpo #BigData #Hadoop #IoT
By: Jim Scott
Oct. 3, 2015 01:00 PM
If you’re in business, you have data. And if you’re like a lot of businesses, you have a lot of data. And it’s not only coming from your customers, it’s coming from other business units, partners, in-house applications, the cloud, hardware logs, etc. And that data could help you be better at your business, if only you had the right solution to access it in ways that deliver quantifiable value.
One solution is to build an enterprise data hub (EDH) through which all your data flows for processing. Many IT professionals turn to Apache Hadoop as the core component of an EDH, but other technologies can be complementary. For example, a NoSQL database can play an important role in an EDH to help manage the complexities of processing and storing structured data for your organization.
What Can You Store in NoSQL?
To show how NoSQL can be helpful, we can describe a simple example. Suppose you have one data set that describes your inventory of books and another data set that describes your inventory of DVDs. Attributes that describe books (e.g., author, publisher, page count) are different than attributes for DVD movies (e.g., director, cast, length), which make data modeling harder in an RDBMS. You either have to have separate product tables, one for each product type, or you have to create a product table with lots of columns, many of which will be unused. On the other hand, NoSQL databases can vary the attributes on a per-row basis, meaning that each database record (representing a given product) can have its own set of columns.
Other EDH Requirements
Other security mechanisms like audit logs on data access let you investigate whether there are any suspicious activities on your data, like heavy data reads late at night. Audit logs are also important for demonstrating regulatory compliance, which is critical for highly-regulated industries such as healthcare, financial services, and telecommunications.
Other complementary security controls that are not necessarily part of your NoSQL solution, but can be added via third-party tools, should be considered as well. Encryption-at-rest, for example, can ensure your data is protected should your physical storage devices be stolen. Data masking and format-preserving encryption technologies lower the cost of securing your EDH by essentially obscuring the sensitive parts of your data.
High performance is another requirement of a successful EDH. Having so much data in one place is no good if it takes too long to access it. NoSQL allows you to take the workload off some of your overburdened RDBMSs and store even more data, while avoiding the overhead. Your datasets are only going to continue to grow, and that’s why you should invest now in technology that’s going to scale with the size of the data.
In addition, analysis of aggregated behavior lets you understand customer preferences. The premise is that customers with similar purchasing histories will buy similar items in the future, and using that information can help you steer customers to those expected purchases. Sometimes the behavior analysis suggests red flags because a pattern is revealed that leads to customer churn. If customers are exhibiting behavior that is consistent with other customers who have left, you definitely want to know that as soon as possible, not after the customer has gone to a different vendor.
Finally, in some industries, preventing fraud is top of mind, and understanding anomalous user behavior will indicate potential for fraudulent activity. Identifying fraud typically takes huge volumes of data for analysis, and collecting data from numerous sources into an EDH is one way to address this problem.
One example of a technology that could fit your needs is MapR-DB , available as part of the MapR Distribution including Hadoop. This product provides an enterprise-grade NoSQL database integrated into Apache Hadoop, so you can process structured and unstructured data in the same platform. You will likely have a wide variety of data types to store and analyze in your EDH, which is why such an integrated solution is an architecture worth pursuing.
Latest Cloud Developer Stories
Subscribe to the World's Most Powerful Newsletters
Subscribe to Our Rss Feeds & Get Your SYS-CON News Live!
SYS-CON Featured Whitepapers
Most Read This Week