Big Data as Core, Big Data as Context, and Big Data as Buzzword Bingo

It’s neither particularly newsworthy nor insightful to suggest that ‘Big Data’ gets everywhere these days, but two recent items reminded me of the gulf between credible execution of a big data play and the more questionable tacking of the big data meme onto an otherwise useful product.

Christmas is coming. Which means skating, and pantomimes (Captain Jack! And the Krankies!), and surprisingly expensive daughter shops, and pie with chicken and banana. But in amongst that lot, the weekend’s email and RSS brought news of

an ideal solution to store, manage and archive big data

and a

service built specifically for Fortune 1000 enterprises who want to rapidly explore how big data technology can unlock revenue from their data.

(both with my emphasis)

Infochimps has been around since 2009, and I’ve been following them with interest. CTO and Co-Founder Flip Kromer and I recorded podcasts in 2009 and early 2012, and we continue to meet up from time to time. From humble beginnings, the company grew to become one of a handful of credible Data Market offerings, before moving on to contribute key pieces of code to projects such as VMware’s Serengeti. Earlier this year, Infochimps’ broader ambitions began to become public as the Infochimps Platform rolled out. In August, the Platform gained streaming capabilities that helped propel it beyond any early reliance upon Hadoop. Then, this month, things got really interesting with the arrival of the Infochimps Enterprise Cloud. As Alex Williams reported for TechCrunch on Monday,

Infochimps data scientists and engineers developed the platform so they could collect lots of data and perform complex analytics along the way. A customer can pull in data from CRM systems and any of the other app silos where data pools then combine it with the data from Facebook, Twitter, and other services. The data flows into Infochimps’ data-delivery service and is cleaned up along the way. Data gets enriched, as needed, with other pieces of information such as demographic data.

The service works with any kind of database. Infochimps can implement any combination, including relational for SQL-like queries, and NoSQL for Hadoop jobs and big data storage. Analysis tools on the back-end provide the capability to create visuals and reports.

The company is setting itself some bold targets, seeking to speed up system deployments, making it easier for existing staff to do new things with data they already own, and freeing users to deploy a wide range of big data tools beyond the default of the cuddly elephant. And they’re targeting this directly at the Fortune 1000; companies with huge IT operations, demanding requirements, and an expectation of support, service and quality, all day, every day. For a small company of around 30 employees, which raised $1.55 million back in 2010 and hasn’t reported an investment since, that’s a big ask.

If even a fraction of what the Enterprise Cloud promises is available today, or demonstrably around the corner, then that team of 30 must be spending most of their time fending off a swarm of investors and acquirers. A nice problem to have, but a problem all the same.

I look forward to seeing real examples of the uses to which enterprise customers begin putting the Enterprise Cloud. I’ll also be watching with interest for rumours of acquisition or investment, both of which are bound to come.

The other piece of news also came from an established company. This time, consumer and small business backup provider Genie9. The company has a new backup product out, called Zoolz, and is making much of the integral “Cold Storage™ Technology” (Ugh!) that gives users reasonably straightforward access to Amazon’s very cheap Glacier storage service.

Personally, I achieve my backup and archival needs through a combination of DropBox, Google Drive, Spanning Backup, a Time Capsule and Arq (complete with its own non-™ hooks into Glacier). But that’s me. A one man band, with a particular set of devices and workflows, and it’s an arrangement that has grown up rather organically.

Zoolz makes perfect sense as a backup solution, and from a brief play with the tool it appears intuitive, capable, and affordable. The Glacier integration is also good, for those things you want to keep, but which you don’t need to access regularly. I have no problem with the tool at all, but what did (and does) bemuse me was the emphasis upon its role in meeting big data requirements.

Zoolz is designed with big data support in mind and will be a game changer to help companies move all their data to the cloud in a secure and fast way that is cheaper than tapes and traditional solutions.

Huh?

The web site devotes a whole page to the big data capabilities of Zoolz, but I’m singularly unconvinced. The whole point about big data, surely, is that you work with it? You pour it into very capable tools that allow you to hold it in (or close to) memory, and you chop and change it in a variety of ways whilst seeking insight? You don’t park it 3-5 hours away in an Amazon cold storage facility and think “job done,” just because Zoolz offers “photo preview” !

Zoolz (through Glacier) offers a place to park large volumes of data that you no longer wish to work with, but it does nothing at all to help people ingest, process, analyse or understand big data. Moving large volumes of data around is slow and expensive. Processes to work with data are often scripted or otherwise automated, and tied into workflows that make sense within the context of the analytic tools (like Hadoop, say) to be used. It’s wholly unclear that Zoolz’s pretty UI and consumer/small business workflows make any sense in that context whatsoever.

Personally, Genie9, I would be proud of what I’ve made in Zoolz. But I’d drop the ‘big data’ stuff. It doesn’t fit.

Bingo card image by Flickr user Sara