Jonas, chief scientist of the IBM Entity Analytics group and an IBM Distinguished Engineer, doesn’t fear large data sets. He’s working on systems that work better and more efficiently, the more data there is to crunch. “Big data, new physics,” he calls it.

“It’s really about big data in context,” he said. “On this journey, with context, you end up with having higher-quality predictions because both your false positives and your false negatives are declining.”

Whereas data managers for years have worried about data cleansing, data hygenics and data sterilization, Jonas says that more data — and more information about that data — helps define patterns in data that otherwise would not be found. “Your bad data becomes your friend. It turns out you don’t want to overly clean your data.”

The next generation (dubbed G2) of business intelligence systems Jonas is working on will be able to evaluate new observations vs. previous ones in real time, he reported. It also will be able to handle “abstract entities” and “exotic features,” and become tolerant of “uncertainty or disagreement.” In other words, G2 will be able to “learn.” And all in a response time of less than 200 milliseconds.

Is there any limit to our ability to understand the world around us through data? That’s the big question.

One of the first technology news stories I ever worked on in my career was titled something like “Mainframe at your server.” The phrase recalls the bygone days of client/server, when hardware was king. Today’s adage, however, is “mainframe at your service.”

The February edition of the Enterprise CIO Decisions ezine on the topic of shared services is an important one, as the promises of virtualization, cloud computing and centralized IT become fully realized in enterprise computing. Yet IT executives don’t have a set definition of what shared services are. For some, the concept is all about pooling resources through virtualization; for others, shared services are a means to an end, that end being cost savings and efficiency.

But in reality, shared services are about moving from the server to the service. “It’s a fundamental shift,” says Jake Hughes, chief technical architect at Seattle Children’s Hospital, in the ezine. “Instead of HR saying, ‘That is my server and that is my storage,’ it is their service and they have no idea what’s on the back end. It is no longer any one person’s or any one business unit’s storage because we may move that storage 10 times in one week, depending on the needs of the overall organization.”

Experts agree that the goal of a services-oriented IT environment is to make services readily and easily accessible to end users. That in itself is what will really revolutionize IT.

Like the intern who was hired to perform daily data backups to tape drives and mail them to the SMB’s disaster recovery location — an off-site vault. An IT manager decided to check up on the intern’s work after the intern left — and found that the tape drives had no data on them … zilch.

The architect who told me about this chuckled, partly because an intern was used to perform such an important part of his company’s disaster recovery plan, but also because, well, it was just bad quality control on the part of IT.

I look back on the summer jobs I had while in college and remember an example of bad quality control. I worked for an advertising agency that had me (an English major) making copies of floppy disks — its strategy for daily data backups. Did I know the importance of these mysterious black squares? No. Did I do a few things wrong? Oh yeah.

For example, a guide on how to make duplicate copies of your floppy disks says:

• Don’t allow them to come into contact with heat, dust, magnetic fields or electrical appliances.
• Do not keep all of your backup disks together in one place.
• Do not continually use one disk, as disks do wear out! One high-density disk can store a lot of text-related documents, but it is best to make several copies of your work on separate disks.
• It is best to use Windows Explorer or My Computer in Windows to copy files to floppy disks rather than application software such as your word processing or spreadsheet programs.

I didn’t know any of this, and neither did the person in charge of me — we used the same disk over and over.

I know that most SMBs have better quality control for data backups in place than these scenarios, and the technology has come a long way: disk-based backup options are dizzying, and there’s new cloud-based backups popping up all the time. Then again, a recent study by Enterprise Strategy Group found that on-site disk and tape are still the backup approaches of choice at most businesses.

Moving into September, SMB backup options is a topic that several experts will tackle on SearchCIOMidmarket.com, including how one CIO is sticking with tape drives as his primary backup plan, and why he’s not keen on cloud-based backup options.