Part one of a four-part Magic Wand Infrastructure series.
Part 1: What's your magic wand infrastructure?
Part 2: Waving away complexity
Part 3: Voila! New staff, new mindset
Part 4: Barriers to making magic
Pete Buonora, Enterprise Architect, BJ’s Wholesale Club
Tim Elkins, CIO, PrimeLending
Cynthia Stoddard, SVP and CIO, NetApp
Cliff Tamplin, former VP of Technology Support & Risk Management, Hyatt Hotels Corporation
INTRODUCTION: Most, if not all, IT problems are symptoms of complexity and the costs involved in managing legacy infrastructure. A 2013 Forrester survey of IT leaders at 3,700 companies found that they spent 72 percent of their budgets on replacing or expanding capacity and supporting ongoing operations and maintenance.
But what if you were handed a magic wand and could wave away the past? The Enterprisers Project assembled a group of seasoned IT leaders in an interactive exchange to find out. Here are highlights from the conversation.
The Enterprisers Project (TEP): Let’s start this conversation with the first things that come to mind, now that you’ve been handed an IT magic wand. What would you get rid of, what would you keep, and why?
Cliff Tamplin: I’d start by getting rid of all the legacy applications, because in most cases it’s the applications that drive the goofy infrastructure we have to maintain. Generally, the complexity you talk about is introduced by the variety of different systems we brought in over decades, and we still have systems around that require completely different infrastructure requirements to the new systems that are currently being deployed. That diversity and heterogeneity are the real killers.
Tim Elkins: If I were to start over, I’m not sure I can give you any system I want in house. Our philosophy is that we want to work on strategic things and not waste our time on anything that is not a differentiator. No one gets rewarded for hosting Exchange really well. I don’t have any interest in having staff maintain systems if it can be done equally well at a third party. Our core system here is actually hosted; we don’t have it on site. The great relief for us is not having to worry about disaster recovery or the updates or the refresh on the hardware or any of those types of things.
That being said, I may want to wave that magic wand you’ve handed me a bit differently. Where we run into problems is where we don’t have control over the applications to make updates and changes. We are dependent on that third party that may or may not move as quickly as we can. So if I’m thinking about the perfect structure, there are some core systems I may eventually want to host, but I also want to have full control to making changes to help it grow and strengthen.
Cynthia Stoddard: I agree with that. I’d love to have a magic wand to wipe out everything in our data center. I’d actually start by looking at the framework of where I’d want to be in the future and frame that around a services model. Because I agree with Tim: It would nice to put things outside the data center and not focus on it, but I think in some cases, because of the speed and how you want to differentiate yourself, you will end up still with some infrastructure and applications internally. But I’d start with the basics to make sure that IT was positioned to provide services, and provide those services out of a true cloud environment either private or public.
Making Software a Bit More Magical
TEP: When you said, Tim, that you don’t have control over the SaaS applications, was that referring to what Lee Congdon, the CIO of Red Hat, calls upgrade cadence?
Tim Elkins: Absolutely. And probably even more so than that, I’m talking about a very specific struggle we have with a hosted SaaS solution we have in place. The updates or upgrades to the solution and the base product roadmap are certainly one thing, but there are things in core systems that when you host them, you may want to do something different from other customers. And to be able to move quickly, and to have some control over that to move quickly has been really important. The flipside is, that application, if I’d hosted it internally, I would have a much larger control over the changes that I don’t have now.
TEP: Cliff, when you used that word “goofy” for infrastructure, does that also extend to the software customizations that often have to happen to meet the specific needs at a company?
Cliff Tamplin: Yes, although I was focusing more on an application that is kept in-house that requires very specific versions of UNIX, and it requires very specific processes. I guess I could think of two or three very specific applications that require that level of attention. That then spills into needing specialist skills to look after those elements. So it really is the whole gamut, which is where the desire Tim was expressing came from: If you can get out of the business and have someone else run the infrastructure for you, that obviously makes you more nimble because somebody else has that problem and enables you to focus on things that really add value.
Pete Buonora: I totally agree with that, especially around creating services to meet our immediate needs and also being able to have some customization capabilities in SaaS. We’re really trying to break these things down into the differentiating and non-differentiating, those that keep the lights on, those that help us to grow, and those transformative things that help us move into different markets. These needs break down into different delivery models, which means IT is becoming the broker for those services. This is very different from always choosing the traditional, in-house data center, whereas now we’re building up a new data center with services, starting on that public and private cloud capability for things that are differentiating, things that are transformative and give us that ultimate flexibility so we can go out to the cloud when we need to but also other areas where we may get more benefit from say hosting on an external platform as a service.
So one of the struggles is not just technologically but culturally, the way we’ve been running things through the years is drastically changing now, for example being a broker but somehow competing with the outsourced model where you have to deliver with the same cost structure and at the same pace. I think it remains to be seen how much stays in privately-owned data centers, and as security barriers and other areas improve in cloud services, there are times when it won’t make sense to even try to compete on that level.
In part two of our roundtable discussion, Cynthia Stoddard, Pete Buonora, Cliff Tamplin and Tim Elkins will discuss waving away complexity. Stay tuned.
What if you were handed a magic wand and could wave away the past? What would you change about your IT infrastructure? We'd love to hear about your experiences in the comments section.
Photo by Flickr user smbuckley23.