Knorr: It seems like you've made a huge investment in the Oracle Cloud: $5 billion a year for seven years. How could you ever expect to get that back?
Hurd: Let me tell you a little about the industry, because I think we're at a point that's going to be a pretty dramatic change. First, the IT industry now is $1.8 trillion, something like that. A little less than half of that is enterprise. Inside the enterprise, a little more than half of that is services. The other half is product. I think this is going to change dramatically.
I think IT budgets will be 2 percent or 3 percent growth. I think you've got a very complicated infrastructure. What do you think is bigger, the R&D budget of Oracle ($5 billion) plus the R&D of IBM ($5.4 billion) -- add them together for $10.4 billion -- or the IT budget of one of the top three banks in the U.S.?
Knorr: I have no idea.
Hurd: It's the IT budget of one of the three banks by over $3 billion. At its lunatic worst, this is nuts. The complexity ... this bank has over 8,000 applications, 70,000 servers, 72 petabytes of storage, tens of thousands of humans coming together to cobble a server together with OS, with database, with middleware. All of which adds, at the end of the day, no economic value-add to the bank.
So this whole issue of the provisioning of IT, I think when you hear us talk about engineered systems and the cloud, I think they are phenomena that come out of the same issue: Eliminate complexity. Simplify IT.
I think what's going to happen is a standardization of IT that's going to be pushing the work to us. You, the customer, say: Stop, you do it! You engineer this stuff together.
So when you look at the world's apps, 25 percent of the world's apps are standard, packaged. Of the 75 percent that are not, half of those, less than half, started out as standard, but got messed around with by some customer so that the guy who is the originator of the app -- us guys -- can't recognize it. And the rest are home-grown.
Imagine if the apps that are "standard" include sales, accounting, supply chain, manufacturing, HR, across fundamentally every process. We think this is a renaissance that is occurring and it's driven by an economic dynamic -- we've got low IT budgets for one thing -- on top of that we have lots of appetite to do new and innovative things. So we've got to get money out, we've got to simplify IT to do new innovation. How do you simplify? You push the work back into the industry.
Knorr: There are a few components to what you're saying. One, if you're going to offer some integrated ecosystem, you've got to have best of breed across all those application categories, which is a pretty tall order. The other thing is that if you're using virtual machines rather than multitenancy, how much does the customer have to work on setting up and maintaining those environments? How much are you really taking out of the equation?
Abhay Parasnis: While it's certainly not a trivial task to bring that broad a suite of best-of-breed applications, there are three or four fundamental shifts that we believe are happening. One is at the hardware level where virtualization is at the core of everything, so it lets you decouple the workload from metal and create highly parallel scale-out architectures rather than scale out. Once you have applied that design point, one set of integrations becomes a lot easier, because you no longer have single points of failure.
The next one is service orientation, where you're inherently designing systems to be loosely coupled, but integrated for the user.
Knorr: It's pre-integrated?
Parasnis: No, it's designed from the get-go to be composed together. And the last piece is actually quite important, which is to enable this new class of innovative experience -- whether it's mobility, socially enabled apps. A lot of pressure on IT is about serving the edge, if you will. So everything in our cloud -- though as you noted it's a very difficult task for anyone else to match -- we actually feel extremely good about the foundational bets the company made six years ago. We happen to be extremely on target.
Architecturally and core productwise, we actually feel that the portfolio is exceptional, but they are engineered to be delivered together. That's actually very difficult to do post-facto.