Looking for your help with a Big Data roundtable
Summary: We're looking for a mix of executives, experts and academics to debunk big data and real-world implementations.
In October, we're looking to hold a big data panel to focus on the real-world implications for corporate IT. I'm looking to crowdsource a few ideas.
The overview of the panel roughly goes like this:
The big data and analytics panel will focus on how real world applications deliver value and how drowning in data has become the biggest issue in the enterprise. How do you filter and use that data to better understand customers, partners and employees?
That's the boilerplate, but here's where you come in.
- What panelists and big data experts would you like to see?
- What detours on the big data topic are worth exploring?
- And what are your must have threads on the big data and analytics topic?
We're looking for a mix of executives, experts and academics to debunk big data and real-world implementations. The potential panelists will riff for an hour or so on big data. The panel will also be livestreamed with a meetup afterward.
Thanks for your help and I look forward to hearing your suggestions via the talkbacks below, email or Twitter @ldignan. Cheers.
Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.
Talkback
Thank goodness...
As for threads of information, I would like to see a nice comparison of 'big data' promises vs. reality in terms of what problems people have actually solved. Most of the implementations I've seen and worked with have to do with HA and redundancy in systems. On the other side of the coin is the 'big data' analysis (with technologies like hadoop) to take advantage of parallel processing techniques. These still require real people to perform analysis, 'big data' doesn't mean there is now some magic button that does all the work for you. Right now, I see only very large institutions that can afford the type of hardware and people needed to run these types of functions.
Also, if we could start killing the term 'big data'... it makes people think that it is somehow fundamentally different from any other type of data... it???s not... there???s just more of it. In another few years, with the rate of data acquisition, 'big data' will just be regular data anyway.
Hmmm...
I think you really need to talk to the data mining experts at http://www.kdnuggets.com/ which is the most popular portal on the web for data mining, analytics, machine learning, BI, etc. The fellow who runs it, Gregory Piatetsky, is one of the top experts in the field, founded a conference on the subject, etc. His profile is at
http://www.kdnuggets.com/gps.html
He's been an invaluable presenter on some other panels I've seen and definitely a great resource for what you seem to be looking for.
OSS 'Big Data'
How about debunking . .
Or better yet, debunk the idea that ZDNet cares about something more than buzzwords and panaceas? That's an easy one.
ZDNet's obsession of panaceas is beyond belief. "Cloud" computing is a panacea for all ills. So is BYOD. So are social networks. So is big data. So are tablets and other mobile devices.
Frankly, I'm sick of the obsession with panaceas. For every article like this one that wants to look at the drawbacks of a new idea, there are 100 articles that poo-poo any serious analysis of any drawbacks whatsoever. ZDNet in general prefers to ridicule rather than seriously study things, and prefers panaceas over reasonable solutions.
Frankly, I'm tired of the panaceas. More articles like this, please.
Here are some things I'd like to see studied:
-The benefits. I'd still like to know what the actual, tangible benefits are. Right now, I'm not really seeing them. "We [i]might[/i] stumble on something that helps us" doesn't really sound like you're going to get a good ROI.
-The systems that process "big data" have to be pretty big. That really sounds like a whole lot of overhead, which ties back into ROI.
-The overhead of computing systems in general for businesses. I look at a lot of large businesses, and I do wonder if the overhead is really less than any loss of sales they'd get without it.
I'm not an expert, but "big data" seems to amount to looking for a needle in a haystack - when nobody said the needle was there, and in fact might not exist. You're just searching the haystack because it happens to be there.
Which projects succeed?
Alternatively they're incremental and build on functionality from other projects... so they're deemed to succeed regardless simply because they work.
Big projects that succeed are also linked to certain people and certain software.
This is theme should be explored in more detail to quantify what those qualities are.
:-)
Someone from Google or Facebook..?
Murtaza
Big Data and the "cloud"
Big data has big issues, including privacy.
If I had anyone I'd list as most important it would be you and the other editors here and at smart planet etc.
Big Data can be hyped all day long, but it can't steer free from common sense.
Start with a definition, end with a plan
I have seen "big data" used to mean: lots of data, unstructured data, data that needs to be constantly recategorized, data that I want but don't know how to gather or organize.
Perhaps if the panel starts with a clear definition of what Big Data means, that will lead to a useful discussion on how it is used. Once you know what kind of Big Data you have, you can then move to understanding its application.
Depending on your definition, Big Data can have very different applications for different sizes of company.
I look forward to seeing more detailed analysis on this topic.
It's not all or nothing
My takeaway is that BigData tends to be viewed as an all or nothing approach where to really have a project in the space, you need to be dealing with technologies typically associated with it (NoSQL, Hadoop, etc.) or you're not doing it. Reality is somewhere in the middle -- people are looking at real problems with real apps and real databases and are instead asking how they are going to scale it out with the resources that they have. It may not be BigData with a capital B, but it is introducing the question of how to address scale at levels well past what the original architects of the solution intended.
I'm loathe to ask it, but BigData needs to go through that painful cycle that cloud did where definitions need to get dragged through the mud because like the folks that were building highly automated virtual compute farms finally were called clouds by the purists, I think the people that are hacking together solutions to get things done with datasets larger than what their creators intended should get the credit for working bigdata. There is a good discussion here.
There is also the business question of value. Lots of cool visuals but I still get the occasional CIO that bluntly states that to date, none of the lines of businesses he supports can come up with a use case that isn't answerable using current solutions in place. Smart technical people can dream up use cases, but businesses are in-turn pressed to make use of the result.
Taking BigData and going vertical with application-specific data and vertical-orientation to business is something that I think is also inevitable. It is difficult to extract value out of highly generic systems. Room for discussion here... I'm sure there are some people that are building tools that would beg to differ with me.
The list goes on...