Monthly Archives: February 2014

The NoSQL Question

How the NoSQL technologies are shaping up in 2014?NoSQL Question

Few weeks before, I’ve provided an overview on NoSQL databases mentioning the key features, classifications and use cases. Coming back to this topic today, I’m summarizing the key trends that Pundits are noticing with respect to the NoSQL technologies in 2014 and beyond.

  • NoSQL is racing – The NoSQL database market is expected to grow at a compound annual growth rate of nearly 60% till 2017*. Major enterprises are already putting NoSQL datastores in key line of business applications.
  • It’s about use case  – NoSQL technology is filling the gap for many applications where transaction consistency concern is secondary to scalability, agility, variability and performance gains. There is no one-size-fits-all approach here. Each NoSQL database is fit to serve a particular business case.
  • Hybrid ecosystem – Now it is about managing all data, not only Big Data. Enterprise is looking for a data hub – a hybrid ecosystems where all data will be stored from everywhere. In this hybrid ecosystem both relational and non-relational data stores coexist, correlate and complement each other (called polyglot persistence).
  • Renewed focus on security – Customers are demanding tighter security and data protection. To address this, database vendors are offering features like Transparent data encryption (TDE), file/block level encryption, de-identification techniques for data masking, and process-based access control lists (ACLs).
  • Marrying with cloud – Pushing databases into the cloud is going to be mainstream. NoSQL databases are perfect fit for the cloud, relational databases are catching up too.
  • Bridge the gap with SQL – You cannot ignore SQL language! Many vendors are trying to bridge the gap between SQL and NoSQL code – offering ways to write SQL syntax on the top and giving cross-node scalability with support for SQL queries and even providing ACID capabilities. Example: FoundationDB.
  • Focus on data virtualization – Data Virtualization is expected to gain more importance in the hybrid ecosystem where several disparate, multi-layered data sources physically coexist.
  • Consolidations on the way – More M&A are expected to consolidate the NoSQL market.

What you are thinking? Let me know and I’ll add to the list.

Further reading / related articles:

  • Top Trends in Database Management – White Paper By Patric Gates (datAvail)- Feb 2013
  • Trends in Big Data a Forecast for 2014 –  CSC : http://goo.gl/lReYzb
  • 5 Big Business Intelligence Trends For 2014 – InformationWeek : http://goo.gl/kUzn5B

* source: Silicon Angle 2013

Advertisements

Interactive Financial eXchange (IFX) Overview

Few weeks before, I had published a post on the Interactive Financial eXchange (IFX) Standards.

Recently, I’ve taken a deep dive and covered the following topics with more details in some slides.

IFX Logo Blue 2012

  • What is IFX and why
  • IFX Forum overview
  • IFX Framework
  • IFX Object and Message structure

Please find and review the presentation at http://www.slideshare.net/PinakiGuha/ifx-overview.

I’m thankful to Judith Vanderkay from IFX Forum for allowing me to use the resources and diagrams from the ifxforum.org website and providing me some valuable pieces of information.

Please visit ifxforum.org for more information.

Cloudera launched Enterprise Data Hub: One place to store all your data!

Cloudera has announced the release of Cloudera Enterprise by calling it as the industry’s first complete Marketing and strategyenterprise data hub solution. Last year Cloudera had clearly mentioned it’s plan to go beyond a Hadoop-MapReduce solution and expanding it’s scope as a hub for all data in the organization.

As per Cloudera, this is the one single platform for all your data needs – ranging from text based search to batch processing, SQL based query service to near real-time, in-memory analytics. No need to purchase and manage several data systems –  bring everything into the hub and do anything you want!

The new data hub does have the following key capabilities.

  • Allow organization to consume, process, and analyze data in a single unified platform
  • Support several interfaces for easy distribution and integration
  • Facilitate variety of workloads – SQL-based query, search, batch and near real-time processing
  • Possess enterprise-grade security model, governance and data protection

Currently, Cloudera is offering the subscription service in three editions. Pick any as per your need.

  1. Cloudera Enterprise Basic Edition: the basic edition with the company’s generic Hadoop software and support.
  2. Cloudera Enterprise Flex Edition:  gives option to choose any one advanced component as per your specific business need.
  3. Cloudera Enterprise Data Hub Edition: the most feature-rich offering, which includes all the premium components to support all data management needs for the enterprise.

Will Cloudera Enterprise significantly increase the Hadoop adoption in 2014?  Some complains that, it may be too early for many companies to adopt such a data hub. Gartner’s Merv Admin has said  the enterprise data hub as “aspirational“. With lack of available talents to run Hadoop centric data systems, and less matured technology and integration standards; many companies are not ready for the Hadoop yet, leave aside an enterprise data hub.

However, Cloudera is countering that the data hub will be critical for companies future data strategy as  this solution will save enterprises from spreading data across different data centers and face the cost, complexity and integration related challenges. “With Cloudera Enterprise, today’s leading organizations put their data at the center of their operations, to increase business visibility and reduce costs, while successfully managing risk and compliance requirements”, says Cloudera.

Now it’s time for HaaS!

Data CloudSilicon Valley startup Altiscale has recently launched Altiscale Data Cloud offering Hadoop-as-a-Service (HaaS). As per Altiscale (read the full report here), this is a “purpose-built” Hadoop cloud – a cloud infrastructure optimized just for Hadoop.

While Apache Hadoop is the most popular platform for handling Big Data, deploying and managing this key technology has its own operational and infrastructural challenges. For example, the learning curve for gaining the right skill and knowledge for deploying and running Hadoop is steep. Many companies aspiring to adopt Hadoop, lack these operational skills. Apart from the operational issues, not many companies do have the infrastructure like Google or Yahoo to run Hadoop on their premises.

HaaS may be the right solution for these companies. By adopting Hadoop as a service, companies can really focus on their analytics, rather than dealing with the day-to-day operational challenges.

“People need to spend more time on the data science and not on the Hadoop infrastructure”  – says Raymie Stata, the co-founder and chief executive of Altiscale. Altiscale claims that their Hadoop cloud service does have a built-in “auto elasticity” feature to optimize resources based on the data need. This is achieved by dynamically adding and removing resources while offering storage from few Terabytes to Petabyte scale. Altiscale backs their service with the operational support on Hadoop and the jobs running on top of it.

The Altiscale Data Cloud is built on HDFS and YARN and accessed by a Linux based workbench with the pre-installed components like Hive, Pig and Oozie. As per Altiscale, the data cloud easily connects with the external systems like customer’s data center or any other cloud service.

As the technology landscape seems to be heavily influenced by the “3rd Platform”, the emerging platform built on mobile, cloud, big data analytics, and social networking, we still see slower adoption to Hadoop primarily due to the operational and infrastructural challenges. The HaaS market can hit the spot here. With a solid and secured HaaS, companies can focus their energy on gaining valuable insights rather than trying to tame the Hadoop elephant on their backyard. As public cloud adoption is predicted to go up, overcoming the current data security challenges, companies offering the Hadoop cloud services, may play a strong role on accelerating Hadoop adoptions while competing with the popular on-premise commercial Hadoop distributors.

Note: All the views mentioned in Pinaki’s personal blog are his own and NOT of his current and previous employers. Pinaki is NOT representing any of his organizations through this blog. This blog is just for sharing his reading, finding and learning in the Information Technology area.