The business challenges and opportunities of today and tomorrow are simply too demanding, too complex, and too dynamic for traditional systems.
Think about the mission-critical technology initiatives your company’s either evaluating or already pursuing: Big Data, social business, mobile applications, private clouds, real-time analytics, superb real-time engagements and customer experiences. When most or all of those are up and running, will your enterprise architecture and core systems be able to deliver the performance all of those game-changing business tools require?
(Part 2 of a 4-Part Series: How Software on Silicon Will Redefine Enterprise Computing.)
And, as if that scenario’s not scary enough, jump forward 12 or 24 months from now: even if your current systems can handle all that, which is very unlikely, how will you be able to keep pace with the new requirements that’ll be in place by April 2015?
To help businesses in every industry avoid today’s performance gap that will become tomorrow’s performance trap, Oracle’s taken a page from NHL legend Wayne Gretzky, who once said while most players skate to where the puck is, he skates to where the puck is going.
To bridge that IT performance gap and give Oracle customers the ability to skate to where their business is headed rather than where it’s been in the past, Oracle has just introduced computers with world-record-breaking performance by fusing critical pieces of the formerly disparate worlds of software and hardware in ways that deliver cost-performance unlike anything the tech industry has seen. (For some of the details on that, please check out Part 1 of this series, called You Can Get A Chip Slower Than The New Sparc T5, You Just Have To Pay More.)
Yes, those are some big promises—but, in today’s highly demanding world, incremental improvements aren’t going to be able to meet the needs of businesses that have to move at the speed of their customers, capture and analyze and act upon massive volumes of data every hour of every day, engage with customers in the ways those customers demand, and dramatically change the way IT systems have been designed, funded, and run for the past few decades.
And in that context, here are 10 reasons why Oracle believes it can help businesses not only cope with but fully exploit the dynamic new business models and revenue models that are essential in the emerging world of the global consumer-driven economy. In so doing, Oracle’s bold new software on silicon approach will redefine enterprise computing: its economics, its performance standards, its expectations, and its value to the business.
The first five reasons come from comments made by Oracle CEO Larry Ellison at last month’s launch of Oracle’s new T5 and M5 systems—the fastest business computers in the world—and the second five come from executive vice-president John Fowler’s remarks at that same event.
1) Software on Silicon: Oracle’s breakthrough approach to extreme performance. “So the idea we have to not only maintain our leadership and keep on this doubling performance every year slope of the curve is to start moving a lot of features of our software—from our database, from Java, from middleware, and from applications—out of the software and into silicon. Out of software, and into hardware – the ultimate optimization,” said Ellison.
A short time later, he picked up the software-on-silicon theme once again: “We’ve done this for encryption/decryption. We’ve done this for scientific computing. But isn’t it interesting: we’ve never done it for database? Isn’t it interesting, we’ve never done it for Java, where we’re going to do it? And as far as I know, we’re the only ones that are doing it, and we think this is going to give us the ability to double performance again, and again, and again.”
And that’s only possible, Ellison said, because through its acquisition of Sun Microsystems three years ago—an acquisition that Ellison recently said was the most strategic and profitable in the company’s history—Oracle had the engineering talent and the technology to do what no hardware company and no software company could do: “We are a software company that also does silicon. And we’re a silicon company that also does software. And maybe you’ve heard this before: engineering hardware and software to work together yields up huge benefits in performance, in reliability, lower cost, lower power consumption, all of those things.”
2) Oracle has displaced IBM as the cost-performance leader in mid-range computing. “And this is the one that I think the people who follow the microprocessor wars will be most surprised by. For years, IBM has been number one in microprocessor integer performance [a key speed indicator]. But that’s no longer the case. The T5 is much faster, and even in integer performance. That’s kind of shocking. Look at the size of the T5 machine, look at the results, look at the cost difference, look at the advantage—it’s almost a factor of 10—in raw integer performance for servers based on the new T5.”
3) Your business can run faster when all of your software runs far faster. “But as of now – as of now, this is the fastest single server in the world,” Ellison said. “It’s been tested on the industry standard TPC-C benchmark. It’s fascinating. It’s not very big, but there’s been no single server that’s ever run the Oracle database faster than this, on an official benchmark. We also tested the T5-8 for Java, and it’s the fastest Java server in the world. With respect to enterprise, it’s the fastest middleware server in the world.”
And the same holds true, Ellison said, for applications, which are the backbone of your company’s processes and behaviors:
“For your applications, it’s the fastest computer for SAP applications as well," Ellison said. "
We set 17 world records with this microprocessor and systems based on this microprocessor.”
4) The new high-end Oracle M5 handles the world’s most-demanding workloads. “Again, it’s chock-full of the reliability and scalability and fault-isolation features you would expect to find in a mainframe computer…. It’s a combination of extreme scalability, where you could then subdivide the machine into lots and lots of VMs. It’s great for your most demanding applications, it’s great to build big cloud centers and then partition, and virtualize, and parcel out, and chop up the capacity into lots and lots of virtual machines.”
As an example, Ellison cited a “giant New York-based financial data company” that tested the M5 against its existing IBM systems. “They gave us enormous amounts of data, and they use both Sun machines and they use IBM machines, and they brought the M5 in, and they were stunned. They were stunned with the performance. They had an application, they moved it to the M5, here are the results: higher transaction rates, fewer cores used, much better response time. And the thing that perhaps they liked the most, it was 20 percent the cost! So, you can go faster, but only if you’re willing to pay 80 percent less,” Ellison quipped.
5) Oracle’s new systems set new standards for price as well as performance. Noting that a single Oracle T5 server came close to matching the performance of a 3-server cluster of IBM P780 servers, Ellison said, “And which do you think cost less? Take a guess. In terms of cost-performance advantage, it’s enormous.” Ellison noted some additional cost advantages as well: “And I submit to you, because the T5’s a much simpler machine, the inherent reliability of a simpler machine is also going to be better. Obviously, T5 consumes less power and less floor space and so on.”
(Here we switch over from Ellison’s comments to those made by Oracle EVP John Fowler.)
6) By specializing in both software and silicon, Oracle can unlock unprecedented value. “Historically, servers have been designed by worrying about networking and storage and the operating system and putting those together as a product,” Fowler said in describing Oracle’s unique approach. “It’s not like people designing servers didn’t know about applications, but it was an arm’s-length approach to applications. And here at Oracle, the reason we have all of these world records is because it’s not arm’s-length, it’s arms linked: between the database team, the Java team, the middleware team, and the apps teams. And those influence every aspect of our design. How do we build the right IO structure? How do we build the right coherency interface? What features do we put into different parts of the stack, and different parts of the silicon? And again, since we do all of these things here, and there’s engineers here from every one of those groups, we’re able to make the changes in the design wherever we see the need.”
7) Larry’s Law: Double the speed and power each generation. “What Oracle wanted us to do when we [Sun] came here is, you know, I call it Larry’s Law,” Fowler said. “Larry said, ‘Why don’t you, instead of doing 30% or 50% improvement each generation, why don’t you just double every time?’ And I said, ‘Larry, that’s kind of hard, you know.’ But we’ve had a great success so far with doing this. And as Larry talked about the innovations we can do in silicon, we’re now actually seeing all kinds of things we can do to get massive step functions in performance. And it’s easy to go talk about these things, but it’s a lot harder to go execute on them. And what we’ve been able to do here with intensive collaboration across different engineering teams, and then investments in silicon, is to really go execute now what is a drumbeat of silicon coming out.”
8) Deliver high-end performance and reliability at unprecedented low cost. “At the high end, what we really wanted to do is change the economics. We wanted to actually step the high end up, to allow you to tackle a different class of applications, and a different level of capability, but attempt to do that at a cost-performance point that still makes the high end a very attractive enterprise platform,” Fowler said. “And mission-critical reliability—you can see how at the high end, we now measure all the salient factors in terabytes. Gigabytes are actually not a useful measure anymore, whether it’s scale or bandwidth or performance—it’s just not a useful measure. And that’s the type of environment these things are designed for! So the class of problems that people want to tackle, or consolidation they want to go take care of, you can do all of them in M5 with a common architecture.”
9) Redefining the economics of mid-range computing as well. “We said to the team, ‘Okay, guys, let’s take a cabinet-level system that costs millions of dollars, and let’s go figure out how to build that an entirely new cost point,” Fowler said. “I want at least 100 cores, a thousand threads, terabytes of memory, hundreds of gigabytes of IO bandwidth, and make it scalable and efficient enough that you could run any scale of applications, small or big, on that. Of course,” Fowler joked, “that’s not a very tall order at all! And of course, in a very short period of time, voila, we have the T5-8. And this is the flagship product that Larry talked about, which for us is the best server in the world.”
10) Scaling the Solaris operating system to deliver not just power but also huge value to customers. “If you look at the innovations we’ve put into Oracle Solaris for running these high-end Oracle systems, one of the things we’ve also done is make sure to retain binary compatibility,” said Fowler. “Every customer so far has been able to simply take these systems and run their existing applications, whether it’s the application you’ve compiled, or it’s an ISV application. This is one of the things that’s very important for us: we do not want customers to have to port and fiddle to take care of their applications.”
RECOMMENDED READING:
You Can Get A Chip Slower Than The New Sparc T5, You Just Have To Pay More
Why Oracle CEO Larry Ellison Is So Bullish on Sun Hardware
10 Reasons Why CEOs Don't Understand Their Customers
Oracle's Secret Sauce: Why Exadata Is Rocking the Tech Industry
The Top 10 Strategic CIO Issues For 2013
Oracle Cloud Success Triggers Oracle-Derangement Syndrome
Larry Ellison Doesn't Get the Cloud: The Dumbest Idea of 2013
Oracle Cloud Tops 10,000 Customers and 25 Million Users
Dear CIO: Is The Time Bomb In Your IT Budget About To Explode?
The Strategic CIO Agenda 2013: Strategies, Priorities, and Career-Killers
The Biggest Big-Data Opportunities: How to Choose the Right One
Steve Jobs, The $60 Light Bulb, And The Future Of Technology
