Events, news, technologies and products about cyber security
Tuesday, February 11, 2014
[forbes] Big Data: 5 Reasons Why Hadoop Is Ready for Enterprise Prime Time
As 2014 gets into full swing, Hadoop is increasingly being used for applications that are integral to daily business operations. No longer is Hadoop viewed by some organizations as just a platform for big data proof-of-concept applications. IT leaders should be developing a strategy for production-ready infrastructure, now, so they are ready to leverage the emerging technical advances that make Hadoop more capable of supporting business-critical big data applications.
Inspired by Google’s MapReduce, Hadoop has earned its stripes as a platform supporting large-scale computing for Yahoo, Facebook and other Internet giants. However, until now it hasn’t been universally considered mainstream technology, in part because of v1 technical constraints and the scarce technical skillsets Hadoop requires.
But in my view, Hadoop is ready for production, particularly with the release of v2 and when delivered via a managed services model. Here’s why:
With technical enhancements in v2, Hadoop has transitioned from a batch-processing model to a multi-use platform for crunching data in many more ways, making it ready for real-time big data applications.
Explosive data growth is challenging organizations to scale big data operations at the speed of business after initial use cases prove successful. When Hadoop is delivered through a managed services model, organizations accelerate delivery without worrying about large capital investments, lack of in-house skills, data center constraints or implementation success.
Emergence of a big data best practices stack, encompassing an infrastructure layer, a data layer and an insight layer, ensures delivery of big data applications in the context of an organization’s overall information architecture. (I will have more to say about this emerging stack soon.)
Complete big data managed services that include market-leading Hadoop platforms from suppliers like Cloudera and MapR, network services, IaaS and proven ITIL-based approaches meet the most stringent enterprise needs.
A groundswell of open source contributions, combined with technical innovations in commercial distributions, ensures a Hadoop enhancement roadmap that will enable an ever-expanding list of use cases. Industry noise about proprietary enhancements is meaningless when innovations deliver 100 percent API compliance with HDFS.
In fact, we recently commissioned research by Forrester Consulting that found 65 percent of surveyed enterprises are already using, or planning to use, Hadoop for one or more proof-of-concepts. But only 16 percent were using it in production. I believe that if we retake that survey at the end of 2014, we’ll find a significant shift toward production applications, fueled by the availability of big data managed services.For these reasons, I’m absolutely convinced Hadoop is ready for enterprise prime time. IT leaders are increasingly realizing that the best way to meet their organizations’ demands for big data solutions is through a managed services approach.
Milan Vaclavik is senior director and solution lead for big data at CenturyLink Technology Solutions. Visit the CenturyLink Technology Solutions Blog for more of his posts on big data, and join him at this week’s Strata conference for a discussion on how Hadoop and big data market dynamics fit into the broader IT services landscape. ThinkGig has a Strata sessions preview.