 Live from the Moscone Convention Center in San Francisco, California, it's the Q at Oracle Open World 2014 brought to you by headline sponsor Cisco Systems with support from NetApp. And now here are your hosts, Stu Miniman and Jeff Frick. Hi everyone, welcome back. You're watching the Cube. We're at Oracle Open World 2014. We've been here for three days of wall-to-wall coverage. We're here in day three. We're excited to be here. It's a lot of buzz, a lot of energy and actually our next guest is going to provide a little historical perspective. But I'm joining this segment up on my co-host. Hi, I'm Stu Miniman with wikibond.org and pleased to have not only a practitioner but a Cube alum, Steven Zay. Welcome back to the Cube. Thank you very much. So you're SVP of IT at SEI, a financial services company. Tell us a little bit about it. It's been two years since last time we talked to you. Those are audience that aren't familiar. What's your role? What does SEI do? Well, SEI is in the financial services space and we provide technology and investment products to intermediaries, banks and advisors around the world, primarily in the UK and the US. And we've been embarking for the past 10 years on a project of very grand proportion for our firm to build a new platform to serve that set of constituents. And it's on all new technology and it has global reach and is intended to serve clients of all types and sizes around the world, as I mentioned. We spent well over $750 million on this platform and it's been a venture that we've started since 2003. The linkage here is that we built this platform around the notion of serving clients and intermediaries in a global perspective and the platform has to support different types and different sizes of firms. And we invoked the use of UCS as a journey off of the Spark platform a few years ago and it has proven to create a great deal of value and our ability to support that constituency on this new platform. So wow, if you started this in 2003, that predates Cisco UCS, it predates all our cloud discussion. Flash was something that hadn't made its way back in the enterprise. So can you walk us through a little bit, what were you facing and how do you build a platform that can grow with some of these new technology adoptions? All right, probably the most practical case we have is in the UK we have clients who use our platform but also consume securities and investment products from around the world including the US. So although this is a brand new platform, it still has a batch element to it and that batch element, since it depends on US pricing as well as European pricing, the window to run that process is much more narrow for our UK clients than for our US clients. And we were challenged a couple of three years ago with doing something beyond just the development agenda to optimize the batch runtime of those jobs and support those clients. So we ran into the classic scenario of the constraint of IO versus CPU and we did some things on the IO side and needed to scale more on the CPU side and we were on the Spark platform and we again looked at UCS as a way to dramatically improve our price performance, migrating from Spark to X86 and it provided a huge boost and now we've created a bottleneck on the IO side, which as we talked before the show is a classic conundrum that all of us have struggled with for the past 40 years. Yeah, yeah, Steven, you always say you can never eliminate bottlenecks, you just move them to another place and there's always got to be some part of the system. So how do you go through that architectural decision point as to, you know, where the resources get fit and where it makes sense to make change? You sound like a true performance practitioner yourself. So I'm sure you realize that it's not a perfect science, right? You just try to move that bottleneck and continue to move up that scale and the UCS side has allowed us to really go forward from a performance on the CPU side. Now we're beginning to look at data memory techniques on the IO side, a lot of the engineer appliance from Oracle and from EMC to sort of begin to strike that balance again beyond where we are now. Our constraint is, we're IO constraint with, you know, a CPU infrastructure now that runs 30, 40%. So we've got some work to do on the IO side. All right. So I'm sure you have lots of things that you've learned in this project. Where do you still see the vendors needing to do some work to help move things forward for you? Well, I think, and I think this platform is very helpful. And I think what's going on in the engineer appliance space is very similar is sort of a throwback to the 70s where a firm brings together all the components so that you're not creating one off permutations of an infrastructure on your own. So what Cisco is doing with UCS and sort of an ecosystem between network, sand and the server and what Oracle and other firms are doing on the storage side is continuing to consolidate all these components. So you're really managing the ecosystem, not managing individual components and trying to engineer this yourself. Talk about that from, you said you've been working on this platform a long time, you know, as Stu said, the tools available to you and the processes have changed dramatically. Absolutely. So how are you incorporating those? How are those kind of changing your product roadmap? And I wonder if you can speak to kind of productivity or the way that you're able to deliver value to your customers quicker, more efficiently, better, faster. How have you guys evolved over this? Well, I think, again, not to reiterate the theme, but as firms, the infrastructure providers provide more consolidation of the complexity of these infrastructure components. We're focusing more on the business side. We're focusing more on how deliver value to our clients from a higher level from the technology than worrying about engineering infrastructure. And looking at balancing the systems, looking at more of a macro approach as opposed to a component-by-component approach. And we're not engineering one-off permutations, as I mentioned earlier. We are constructing systems that are comprised of engineered elements that our providers are giving us. And from an application development side, do you find it's more kind of unleashing the power that you wish you had before with these new architectures? Or is it really opening up new territory, new fields, new places for your development to go? Again, I guess what I would say is we move up a layer now. We're managing these environments in a different way than we did in the past. We're focusing more on the value of the platform. Our technologies are able to express more energy on the value of the platform in our clients as opposed to again being engineers at the component level. So Stephen, one of the biggest changes we've seen over the last 10 years is how we think about our data. So data used to be a challenge. And of course, lots of people are looking at things like analytics and other ways to turn that data into more valuable information, new revenue streams. Can you talk to us? Is this something that you're seeing in your organization to try to take advantage of all the information that you have? We're beginning to. I mean, the honest truth of this platform, we're still focused on kind of core services. We do have a data warehouse and the data there is something that clearly has value to our clients, but they have not consumed the analytics side to a great degree yet. They're focusing more on this consolidated platform and adapting it for their business. In our industry, there's a lot of regional providers. There's not too many providers that provide a true global capability. And the one thing that we stress is we have a global platform that you can use for any region, anywhere in the world. And the consolidation that they're able to do around books and businesses on a single unified platform is really where they're spending a lot of their energy. They're not really exploiting the data yet. And we anticipate that they will, but they're not exploiting the data from an analytics perspective yet. Stephen, we talked a little bit about, we went on air about kind of the craziness in the buzz here at Oracle OPPO World 2014. And you were here in 2011, three years ago. And also you bring a perspective, you're not here in the valley. Sometimes those of us that live here, we're kind of spun up in the whole thing. We forget there's people that aren't necessarily here. So from your perspective, I wonder if you can talk about kind of the buzz at the show, what's the vibe, how are your meetings going, especially relative to what it was like in 2011. Well, we talked offline before the show started. It is amazing to me, well, I'm still a little more Oracle-centric maybe than a few of your guests. It's amazing how these engineered systems have become sort of the thrust of the show from Oracle's perspective. Here's a firm that went from software to hardware. Cisco is moving in some ways, cases the other direction, I guess. And it just amazes me as to how these engineered solutions, which are really doing a lot of the heavy lifting for clients have really become such a thrust for shows like this. It's a long way from where it used to be. We moved from the 70s, 80s, where it was all about consolidated systems to the 90s and early 2000s where there was all these distributed components and people were constructing their own. And now it's come back. And the idea is letting the firms do a lot of that engineering on your behalf. Yeah, so Stephen, do you find, is the service support that you get seamless? Because you've got obviously the Oracle application, you've got Cisco components, you've got storage components that aren't from Cisco. What do you see in your interactions and how they deliver to you and how they support it? I think there's certainly not singlessness, that's a word, between vendors, within vendors with this new sort of generation of looking at how we deliver infrastructure. Within a vendor it does bring a lot of simplicity when you have one firm to call and there's less finger pointing. But across firms we still have that issue. We still have the issue of EMC dealing with Cisco, dealing with Oracle. They always want to know, so what version of software do you have on the EMC side and the Cisco side when you have an Oracle bug and same with the other provider? So I think it's a cross vendor, you've still got a lot of issues, but within vendor it's a lot better than it used to be because they're consolidating so much of those pieces. All right, so you've been working on this project for a long time. I wonder if there's anything you could tell your peers as they approach something like this. What if you look back and say, geez, I wish I had done this differently or if I only knew. If I only knew. Well, it's really hard. And I'm sure history will tell all of us, as it always has, that very large technology projects are very challenging and a great boost of percentage of them in fact don't make it to market. We're happy to say that we've got real clients running on our platform, but I think it's one of those things where if you don't have a true sort of enterprise competency around large development, then you better get one when you venture on a project of this size and scale. We sort of learned as we went along and we spent a lot of time and energy sort of building up that capability. We were a development firm, but we weren't an enterprise development firm of the stature that we are now around the size of this platform. When did you make that shift do you think from kind of a development firm to what you describe as an enterprise development firm? I would say 2005-ish, about two years into it. We really started to ramp up and started to really think about architectural engineering in a much more enterprise perspective and in a much more enterprise fashion. And I think that was probably the turning point when we realized how big of a decision it actually is. And was it driven by the customers? Was it driven by you guys just kind of waking up that wow, this is a big opportunity or were there just some challenges you weren't able to overcome with the prior state? I'm not sure if you're talking about the genesis of the platform or the actual migration of the enterprise orientation. No, you figuring out that we're not a development company or we're actually an enterprise development shop and what that really means and how it's different. I think it was a broad-based realization. I think our CEO who actually was the one who envisioned this platform from the very beginning, he saw the need for a true global platform serving the industry. And he's been very close to the project. He was in the early days up on a whiteboard doing some design drawings as to what he wanted the user experience to be and the like. So he's been heavily engaged in the very early days. So that kind of executive sponsorship really created the discipline to stay on this thing for as long as we have. And there was some really difficult days in the beginning. And he also, I think, helped move us toward a enterprise orientation when the time was right. It was sort of a research project at his beginning and it just grew into a true development project as we moved along. So Stephen, who did you guys lean on as you were building this? Was it direct access to the vendors themselves? Was it a channel partner? Who kind of gave you the most help and education as you move forward? It's given the size of this thing. It was a broad set of firms. We had development partners and we had our infrastructure providers. I mean, I think we leaned on many, many providers to get us to where we are today. This thing is, you talk about ecosystem. It's a village. It's a village of infrastructure. It's a village of application software. It's a whole set of organizations within our firm that deliver and support the platform. So you can imagine a number of vendors that are engaged in trying to bring that together, both internal constituents and internal partners as well as outside partners. That's some great messages. That's some great advice for practitioners. Stephen, thanks for stopping by. Projects are hard. It takes a village and stick to it, right? I think those are the three key messages. So again, thanks for stopping by. We love to get people that are actually out there executing with these tools. We're watching theCUBE. You're watching theCUBE. We're here on the floor at Oracle Open World 2014. We're actually in the Cisco booth. We'll be here for the rest of the day, day three of wall-to-wall coverage. You're watching theCUBE. I'm Jeff Frick. Thanks for watching. We'll be back with our next guest after this short break.