Stronger data center management and automation technologies are on tap from IBM’s software group, which is headed by Senior Vice President and group executive Steve Mills. This week at IBM's Pulse conference, Mills sat down with Network World Senior Editor Denise Dubie to discuss IBM’s software strategy and his views on other hot network topics. Here is an excerpt of the interview. (Listen to the full interview.)

On the complexity of managing IT:

The desire for simplicity is one of those aspects of human nature in which everyone says why can’t this be simpler, why can’t this be easier. That is a universal lament. But you’d rather sacrifice simplicity for completeness. If your job depends on it, you’d like it to be complete first and simple second. It’s only through completeness that you can then get the simplification. If you leave things out, if you are incomplete, you are in danger of bad things happening and not knowing about it.

On embedding intelligence into the network:

Having intelligence in the network is not a bad thing, but shoving all your applications simultaneously on top of the network structure doesn’t make a lot of sense. Now you are asking the network to do two things: you are asking it to manage itself effectively, and you are asking it to actually control the applications themselves. The applications are more about the rules of the business than the rules of the network. It’s better design to separate the network and the applications, let the network do its job and the applications do their job.

On Cisco’s VFrame Data Center:

VFrame is focused on network-related provisioning. It is not an application-provisioning model. We can take a server, and we can install an operating system, infrastructure software and applications. We can make all those things installable in an automated way and actually have them made flexible in the sense that we can actually change what is installed on that server. It is a much more elaborate set of capabilities than VFrame, which is focused on network configuration and network equipment-related installation activity. I would view what we do and what VFrame does as fairly complementary actually.

On IBM’s automation capabilities vs. those of competitors:

The others have made their move long after we made our move. It is not a matter of when IBM is going to do this. They’ve woken up to this; we have been doing it for a long time. This whole area of how do you provision, how do you schedule and how do you do it across all devices. HP would like to focus on HP servers. That’s interesting, but there is lot more in the data center than HP servers. Products like BladeLogic and Opsware have a more narrow focus as to what they do than what we are doing in terms of bare-metal provisioning, and provisioning up and down the stack along with scheduling for any systems environment. We do scheduling for mainframe, Unix-, Wintel- and Lintel-based environments. We have a much wider portfolio of capability to deal with the challenge of work scheduling and run-book administration.