When we talk to customers about putting together a big data management system of some kind, there’s a common thread to the issues they are worried about. Not everybody has the same problems, or weights them identically, but there’s a huge overlap. And these three issues cover most problems with most organizations.
First, it’s about skills, and there are two angles here. There are the skills that organizations already have, which they’d like to reuse. If most of your analysts are using SQL, for example, that’s a pretty big resource which you’d like to be able to keep on using. And then there are the skills that organizations don’t have, often related to new technologies like Hadoop. Acquiring new skills takes time and money, both of which are in short supply.
Second, it’s about integration. Specifically, how do you integrate that new data (and data management tools) into your existing environment? How do you query data when it’s distributed across multiple different sources? How do you get your existing applications or BI reporting tools to access this new data, without a big re-write.
Thirdly it’s about security. Hadoop security is no longer the oxymoron it was, but security solutions there are incomplete and not fully mature, though lots of good work is being done. What is there often hard to setup and configure, so even when the tools are there, they are not always being used correctly (back to the skills argument again).
Skills. Integration. Security. Are those the problems you see as you look to use Hadoop and/or NoSQL in your organization?