Native Apps At The Client & Cloud

Srinivasan Sundara Rajan

Subscribe to Srinivasan Sundara Rajan: eMailAlertsEmail Alerts
Get Srinivasan Sundara Rajan: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn


Related Topics: Virtualization Magazine, SOA & WOA Magazine, Microsoft Developer, Big Data on Ulitzer, Hybrid Cloud

Article

Hybrid Big Data Solutions

Use Cases for Microsoft PolyBase

Microsoft recently announced the PolyBase technology as part of SQL Server 2012 Parallel Data Warehouse solution. PolyBase is a breakthrough new technology on the data processing engine in SQL Server 2012 Parallel Data Warehouse designed as the simplest way to combine non-relational data and traditional relational data in your analysis.

PolyBase is part of an overall Microsoft "Big Data" solution that already includes HDInsights (a 100% Apache Hadoop compatible distribution for Windows Server and Windows Azure), Microsoft Business Intelligence, and SQL Server 2012 Parallel Data Warehouse. PolyBase is built into SQL Server 2012 Parallel Data Warehouse to bring "Big Data" benefits within the power of a traditional data warehouse.

Much like the success and adoption of the Hybrid Cloud Delivery Model whereby enterprises found it comfortable to run their workloads partly on-premise and partly in the cloud, this new set of technologies like PolyBase can be termed "Hybrid BigData Delivery."

This concept of bringing the power of Big Data analysis as part of traditional relational databases will likely have a huge impact on the enterprise for the following reasons.

  • In terms of knowledge management of their IT department, enterprises cannot afford to spend time learning new technologies on Big Data whereas they have huge investments on staff with skill on traditional RDBMS.
  • More than staff skills, enterprise use cases that can benefit from Big Data can't really ignore Relational Database processing. This is mainly because enterprise have built all their key decision criteria as relational database data over a period of time and without utilizing them any meaningful insights from massive quantities of unstructured data generated from unconventional sources cannot be realized.

The following use cases explain why "Hybrid Big Data" solutions like PolyBase are likely have a larger impact on the enterprises than the Pure Play Big Data solutions.

Voice Of Customer & Product Quality Analysis
Understanding the issues of your customer and ensure that they are fixed in the product implementations is of the utmost importance for the organizations and the following are components that fit in to the Hybrid Big Data Model.

  • Big Data Components:
    • Call Logs from various call center transactions
    • Event Data from Social Networks
  • Traditional RDBMS Components:
    • Integration with Product master / Bill Of Material
    • Integration with Plants
    • Integration with Supplier Data
    • Integration with various Engineering Processes

From the above segregation of data sources Hybrid Big Data solutions like PolyBase can best attend this use case.

Campaign Effectiveness & Advertisement Targeting
Campaign analysis enables organizations to gauge the success of various campaigns by measuring campaign costs, leads generated and leads converted to customers. One of the major drivers for campaign effectiveness is the advertisement targeting engine, which selects the best recommendation for a given user. This solution is again best served with a Hybrid Big Data solution.

  • Big Data Components:
    • Click Stream analytics information to understand customer behavior
    • Event data from social networks
  • Traditional RDBMS Components:
    • Historical scores on existing customers
    • Predictive mining models built as part of OLAP engines
    • Integration with legacy systems of marketing & finance
    • Integration with risk analysis

ITIL Event Analysis & Predictive Analytics
Information generated from various server logs along with other real time information like network utilization are key to predicting the behavior of IT systems and predict the likelihood of any failures in the near feature. This kind of analysis is key to the availability of large enterprise IT systems. Again this is a good candidate for Hybrid Big Data analysis.

  • Big Data Components:
    • Massive amounts of log files generated from various servers
  • Traditional RDBMS Components:
    • Known error databases
    • System topologies and mapping
    • Event correlation engines whose algorithms are best implemented using traditional RDBMS stored procedures

Fraud Detection Analytics
Financial organization which deals with large amounts of real time events, need to have a robust event correlation mechanisms to ensure that there are no fraudulent transactions.

  • Big Data Components:
    • Event data generated from various channels
    • Social media data like customer location information
    • Image processing data
  • Traditional RDBMS Components:
    • Validation metadata for transactions
    • Customer master

Summary
As evident major use cases in the enterprise will continue to use large amounts of relational data already captured and accurately maintained; however, enterprises do need the power and flexibility of Big Data processing solutions to handle the unstructured real-time data coming from various new sources.

This can be best achieved with Hybrid Big Data technologies like Microsoft PolyBase, and more case studies on this front will emerge sooner.

More Stories By Srinivasan Sundara Rajan

Highly passionate about utilizing Digital Technologies to enable next generation enterprise. Believes in enterprise transformation through the Natives (Cloud Native & Mobile Native).