It says something about how this is evolving that IBM doesn't think it can reach that 2017 opportunity without AT&T as a network partner. IBM wants to be the "high-end" of the cloud market, not the cheap generic but the big-time brand the Fortunate 500 and big government call on first, as it is today for enterprise computing. It already has solid networked security, thanks to its purchase of Internet Security Systems in Atlanta a decade ago. Now it's going to connect ISS to AT&T's global network operations, monitoring sections of that network as it monitors client systems now. I visited ISS's offices a few years ago. Their NASA-like network control centers are staffed 24/7. Behind them are teams of security experts, backed by genius programmers who can take apart a virus faster than you get dressed for work. It's an office building but it has the feel of a university lab and the workstyle is modeled on that. For customers it's a systems approach, sold as a service on top of everything else IBM does. I think it's well worth the money. Latching that to actual data lines, teamed to compatible clouds in many places around the world, gives IBM that high end of the cloud market, a high end for which demand has not yet evolved. But the only way to win in any technology trend is to get ahead of it, so when they talk about three "cloud networks" five years from now, they'll probably be talking about Amazon, Google and IBM. This deal is not big now, but it points to cloud's future, and how it will in time replace enterprise computing with something that can feel, to a corporate user, just as my own networked desktop feels to me. At the time of publication, the author was long IBM and GOOG. Follow @DanaBlankenhornThis article is commentary by an independent contributor, separate from TheStreet's regular news coverage.