samedi 1 mars 2014

Big Data Isn't Just For The Enterprise These Days

A lire sur:

The idea that Big Data is only for the enterprise is ridiculous. Data can be easily created by many sources, not just the ones with the most money.
Brian Proffitt  February 28, 2014

Around this time of year, the concentration of hoopla around Big Data, data warehousing and business intelligence is so thick you could probably make a chart about it. (I'm sure someone already has.) A lot of that is due to the Big Data love-fest known as O’Reilly Strata.
Strata, of course, remains one of the most interesting events in the tech circuit (at least to me), and I'm sorry to have missed it earlier this month. Big Data may be a balloon filled with a lot of hot air, but it's still an important trend. We have, after all, a hell of a lot of data coming in and it seems like we should be doing something with it.

Who's Got Big Data? We've Got Big Data!

But who, exactly, is this "we"? If you listen to the press releases and announcements that came out of Strata this year, you might think that the enterprise is the only real customer for Big Data insights. And of course there's nothing particularly wrong with that particular assessment; enterprises should be using Big Data to adjust more quickly to changes in their markets.
Lately, however, that unremarkable insight seems to be shading into the notion that large corporations are the only type of business who stand to benefit from Big Data. This idea, to put it mildly, is BS.
The most obvious culprit this week is IBM’s Dexter Henderson, who did a conveniently timed guest blog over at InfoWorld that lays out the case with about as much subtlety as a brick. From the article’s teaser: “Henderson argues that enterprise analytics require specialized, high-performance iron.”
Sure, non-commodity hardware is often what you need to tackle Big Data problems. But it’s pretty clear that IBM is putting this idea out specifically to sell that self-same customized hardware. If you have the big iron to sell, it's only natural to want to drum up a business need. What's problematic is the implicit idea here that only enterprises stand to benefit from big data. Henderson doesn't say this explicitly, but since only an enterprise could afford the machines his company is selling, that's certainly his implication.
Other company announcements coming out of Strata were also heavily slanted toward enterprise uses. MapR and Hortonworks are just two examples.
Again, they are not wrong to say that big data would be useful for the enterprise. It certainly can be. But big data is not just for the enterprise. Given that Apache Hadoop can be run on commodity hardware and connected to all manner of databases and analytic tools to work with data, it seems to me that limiting the use of big data to just the enterprise is short-sighted.

A Toast To Start-Ups

The biggest hole in the argument that says big data belongs to the enterprise comes from the very existence of public cloud providers like Amazon Web Services, Rackspace and Google. In very different ways, these companies all provide "public cloud" services that make it possible for small companies and even individuals to obtain sophisticated computing and storage infrastructure to offer services quickly and relatively cheaply.
Sure, some people scoff at this "credit card IT." But the way these services are thriving suggests that a lot of people and companies that don't qualify as "enterprises" need them pretty badly. And if they're doing computing on that scale to start with, well, odds are that at least some of them are bumping up against Big Data problems in one way or another.
You don't have to look far to find examples of smaller-than-enterprise companies figuring out ways to implement Big Data within their business models.
Recorded Future, for instance, collects millions of online documents a day to provide real-time trend analysis for their customers. Coursera and edX provide aggregated catalogs of free and open online courses. Vino Volo is a chain of wine bars located in major airports that uses big data coming from their customer loyalty app, point-of-sale systems and social media to figure out inventory, menu items and overall customer sentiment.
These are not just rare examples, and the trend of smaller businesses using bigger data should continue, according to the University of Denver's Andrew Urbaczewski.
"As more businesses use ‘Big Data’ to boost profits, track trends and grow, the influx of analytics and 'Big Data' will just become ‘Data’. In other words, massive data sets will become common data that more businesses use to propel growth and will be fully integrated within everyday operations.” Urbaczewski said in a recent statement from the school's Daniels College of Business.

Don't Tread On Small

Enterprise business gets a disproportionately large amount of attention in the marketplace. This is not because there are so many enterprises—there aren’t—but because enterprises have a lot of green to spend. To your average IT vendor, one enterprise score could be worth dozens of small- to medium-sized business sales. So it makes sense to go after the low-hanging fruit.
But not, I would argue, to the exclusion of all else. Smaller businesses are easily able to take advantage of big data, and their continued need for public cloud services to store and analyze this data is pretty much solid proof of that.
Business and IT managers in the know will adopt big data tools if they are needed no matter what size their company is. But the continued slant of marketing may lead less-savvy business and IT decision-makers with the erroneous impression that big data is not something for them.
Like any other tool, Big data is not for everyone. But nor is it just a tool for a select few with money burning a hole in their pockets.
Image courtesy of O'Reilly

Aucun commentaire:

Enregistrer un commentaire