Live Blogging: Q&A with Steve Mills, IBM’s Hardware & Software Chief


IBM’s Steve Mills put forth a strong case that IBM’s deep expertise and integration prowess will differentiate the company from it’s major systems and storage competition. It’s early and the converged infrastructure game and IBM is late but deep and has a strong roadmap. As Mills said, this announcement is more about innovation than invention and packaging expertise and components to simplify IT operations is what this is all about. Mills emphasized that the major costs of IT are staff not infrastructure which is fundamental to IBM’s strategy. If IBM can demonstrate that it can cut labor costs, it will be able to charge for integration value in my view.

What follows are notes from the IBM analyst session, a lunchtime conversation with Mills and some of his main tent remarks.

Steve Mills – SVP & Group Executive, IBM Software and Systems

IT budgets are endlessly under stress

Cumulative effect of years of infrastructure investment has stifled innovation

This announcement is more about innovation than invention – to allow customers to get out from under the challenges they face, free up budget and speed up their operations.

32M servers deployed ww

2+ zettabytes

The tail of the IT commet is still very long and PureSystem fits into a heterogeneous world.

IBM is striving for the optimal combination of flexibility / openness and speed (via integration).

Channel – fits into the IBM Blue on Blue program…all IBM branding with opportunities for value added dealers and resellers to make incremental margin by selling integration.

My PoV: IBM really didn’t talk about the channel – maybe it’s implied but I was a bit surprised there wasn’t more emphasis here. Maybe it’s because IBM feels the Blue on Blue announcement earlier this year addressed that issue. Nonetheless, the channel angle on converged infrastructure is relevant and important. IBM has a strong brand, strong channel and I expect the pickup to be significant in the channel.

We live in a labor-intensive business. More money is spend on labor than capital and the focus of this announcement is cutting labor costs and freeing up money to do more projects and of course sell more IBM hardware and software.

My PoV: Mills is right. Nonetheless, services will remain the largest piece of the converged infrastructure pie, accounting for more than 45% of the TAM by 2017.

Mills Main Tent Presentation:

Since 1996, WW spending on servers, power cooling and mgt has increased from $100B to $247B (2013E). “We are in a labor-intensive business.”

-70% of the cost of server is on people, not on the server asset.

-If customers are to go in new directions IBM needs to address the real cost of computing

-need to break thru barriers of what resources have to do everyday.

-Reality – IT and business goals are not aligned – Business wants to go fast, IT isn’t keeping up

My PoV: The traditional enterprise is 5-10 years behind the Web giants and leading cloud service providers in terms of simplifying IT and running IT as a business. Converged infrastructure will definitely bring value and attack the problem, however I suspect it will only allow IT to barely keep up with data growth and complexity. To really ‘compete’ with cloud service providers and not get outsourced, CIOs must totally re-think their processes. Organizations must re-organize, re-train and go beyond agile development, into devops, service management, IT-as-a-Service delivery and so forth. Where IT can’t cost effectively compete it will be outsourced and while converged infrastructure is necessary it’s not a sufficient condition to thrive.

I had a great conversation with Mills at lunch about pricing. Mills is in charge of all pricing for his group and has final say over everything in that regard. His contention is that: a) buying in piece parts is a nightmare and while some customers want to do it the reality is it’s much more efficient to buy packaged infrastructure; 2) he flat out said IBM is not charging a premium for integration. He candidly admitted IBM is going after share of wallet and as such is not uplifting its pricing for these integrated systems. He said the main reason why customers won’t buy this approach is because they’re afraid of lock-in.

My PoV: IBM, HP, VCE and other vendor messaging suggests that with choice lock-in goes away. Meaning if you can choose OS, Hypervisor, etc. that this flexibility gives customers the freedom to avoid lock-in. Caveat emptor. Converged infrastructure is the “mother of all lock-ins” so be careful. But the value of integration will outweigh the risks in many cases, because if you don’t converge the infrastructure you won’t be able to keep up with data growth.

I pushed Mills on this concept of charging for value and he said that IBM is not charging an uplift for its integrated systems relative to buying in piece parts. That’s significant but in my view it won’t last—but customers may never see the uplift because over time vendors will quietly raise prices. There’s simply too much value in the integration and while at scale, costs will come close to piece parts, there’s incremental value that IBM and others will ultimately extract from customers. My view is that over time, piece part infrastructure will decline in popularity and converged infrastructure, in terms of both reference architectures and single SKU systems will predominate. But customers should expect to pay a premium for integration and choose carefully where to place specific types of infrastructure. It may very well make sense to deploy Exadata, for example, but over time you will pay a premium—so make sure the business case works.  


IBM SVP Steve Mills