Volume vs. everything else
When it comes to Big Data, too many companies have an "unhealthy focus" on the amount, or volume, of data they have, rather than the variety, veracity and velocity of it, added Bamforth.
Cloudera's view on the Big Data era
Mark J. Lewis, senior director of marketing at Cloudera, also spoke at Big Data or Bust. Post-event. Lewis joined IT Pro for a quick chat about some of the key themes touched upon during the roundtable discussion.
IT and business: Meeting in the middle on Big Data I think that's been a long-term challenge, actually. IT and the business. Is IT another cost centre,like catering, is it really a part of the core business?
Those who are in IT know they are, but it's very difficult for the business to recognise that because it's doing it's own thing. I think that is about to change.
The Big data skills gap There is definitely a skills gap when it comes to big data, and what typically happens when you have a skills gap, is the experience you already have [within your business] can be built upon.
The only danger is if people assume their existing knowledge is all that's needed. It's useful as a baseline, but then they need to extend it further too.
How Hadoop and co. can help make sense of Big Data Big Data is not really about the data itself, but the amount of times you want to manipulate that data, and ask different questions of that data [which is something Hadoop helps to facilitate].
When we started out we used to talk about Hadoop, and we started talking to customers to understand what we needed, we started adding this new framework around it because we realised they needed more.
"Variety [of data] is critical and really badly understood in many ways by not understanding what's out there, what's available or what to do with it."
On the topic of how to manage the information overload, David Postle, head of global services and customers at Actian, said Apache Hadoop is often cited as a "cure all" for Big Data problems.
"Let's use Hadoop for everything, seems to be the common answer and certainly 18 months ago we were finding that a lot of IT folks were presuming Hadoop was the answer to all their problems and headed down that path, before running into other complications."
This was a topic fleshed out by Mark J. Lewis, senior director of marketing for EMEA at Big Data management experts Cloudera, who outlined the work his company is doing with the support of Intel to make Hadoop-based data management easier for the enterprise.
Lewis provided further insight on the Cloudera Enterprise Data Hub reference architecture, which combines Hadoop with a series of other data and system management tools so data from a variety of sources can be processed on a single, scalable platform. The end result is a significant reduction in processing time.
He backed Postle's point that Hadoop alone is no longer enough for the weighty Big Data demands of the enterprise these days, which is why it makes sense to package it with other tools.
"Eighteen months ago it was all you needed and all that was available, and as long as you had an ability to write data and shove it in there... but people need to extract data in different ways, they need to upload it in lots of different ways [now] and they want to be able to do a lot more with it," he explained.
