Remarks by Bob Muglia, President, Server and Tools Business for Microsoft
PASS Summit Unite 2009
Nov. 3, 2009
PARTICIPANT: It’s my extreme pleasure to introduce this morning’s first keynote speaker. His name is Bob Muglia. Bob is the president of the Server and Tools business at Microsoft. Bob is responsible for infrastructure, developer software, which includes Windows Server, SQL Server, Visual Studio, virtualization products, System Center management products, and Forefront products, and even other products. Bob is one of the most responsible people in this building right now. And we’re so glad that he’s going to join us. We’re excited to have him here with us today. Please join me in welcoming Bob Muglia. (Music, applause.)
BOB MUGLIA: Well, good morning. It’s really great to be here. And it’s fantastic to talk to all of you today. What I’m going to do today is I’m going to start by going way, way back. Going back to the really beginning days of SQL Server at Microsoft.
It was on January 13th, 1988, that I was inside the Marriot Marquis Hotel in New York at Broadway. I’m sure many of you have been at this big hotel there. I was in an elevator headed down to go to the announcement that day of the Microsoft Ashton-Tate SQL Server. I had been with Microsoft for nine days at that point, I joined early January. And this was the official announcement of the product that I was responsible for. I was the program manager, the only technical person on SQL Server at Microsoft on January 13th, 1988. And I got into this elevator and we’re going down a few floors and the doors open and Bill Gates got in the elevator.
I had never met Bill before. So, I stuck out my hand and I introduced myself and I said, “Good morning, Bill, I’m Bob Muglia, the program manager for the product you’re announcing today, SQL Server.” And he looked awkwardly at me. And we rode down to the green room where we were preparing for — he was preparing for his talk.
Now, you might not have guessed it, but back then Bill was actually pretty nervous doing this talk. He kept talking about how Ed Esber, who was the CEO and president of Ashton-Tate, had been practicing his speech for weeks, and Bill had barely looked at his coming out there. Fortunately, Steve Ballmer was backstage with Bill, and in classic Steve fashion, he was cheerleading, jumping up and down, and giving Bill energy.
Naturally, Bill did a fantastic job, and actually Ed did a fantastic job and it was a great announcement that day. And that afternoon, I was involved in doing some demonstrations for the press and for industry analysts where I was demo’ing SQL Server on OS/2 for the first time. It was the first public demonstration of the product. And I was showing off the then-brand new client-server programming model that SQL Server was very much bringing to market and introducing.
It was an amazing time back then. I mean, things were pretty much different. Let me ask you a question. This is that product. This is it. It’s the Ashton-Tate Microsoft SQL Server. Now, it’s hard for me to see out in the audience, but did anyone work on this product, this OS/2 product? I see some hands going up. There are a few folks that remember. Yeah, how about that?
We’ve come a long way, baby, since this product. (Laughter.) It’s come an awful long way. You know, you’ll be happy to know that this box says that it includes both 5 ¼- and 3 ½-inch floppy disks inside it. (Applause.) And I personally laid the software down on every one of those floppy disks when this product shipped.
So I think about what this product meant to the industry and to Microsoft, and it was in so many ways a dawning of a new era. It was the beginning of so much of what’s come since then. And you look at where we are today, and it’s hard to imagine the humble beginnings back there. But I look back very fondly to that and look back to all of the work that everyone in the industry has done, our partners, all of you, together with the folks at Microsoft who’ve helped to bring SQL Server forward all of these years.
And we’ve come such a far distance. I mean, if you look back then, what could we solve? What level of business problems could we solve? Well, as we introduced this product to market, Microsoft was a desktop company. We sold products like Word and at the time MS-DOS and OS/2. We weren’t in the business space at all, and the number of business problems we could solve at that time I could precisely measure at zero percent. So, there was no real business problems, all business systems ran on Unix systems or on mainframes or VAX computers back then.
And over time, what’s happened is the industry has matured, our partners have come so far, the hardware has come so far, and of course SQL Server has come so far during this period. And we keep driving this forward together with all of you to solve a broader and broader set of industry challenges and solve more and more business problems, more and more business opportunities.
You know, we’re now at a point where I think it’s really fair to say that almost any business problem can be solved by SQL Server and the applications that you and the people that work for you and that you work with build. You know, I think it’s fair to think that we can solve 99 percent of all problems.
But there are still some that, honestly, if we sit here honestly that SQL Server isn’t appropriate to solve, really large data warehouses are a classic example of that, or extremely large scale-up applications where today’s limitations limit our ability to solve those last really high-end problems.
Well, one of the things we’ll talk about at this conference and today and as we go forward is all of the investments that we’re making together with you to ensure that we will be able to solve all of the business problems that you have, 100 percent of all of the business problems.
And so we’re going to show some things. Ted Kummert’s going to come out here and show some things that talk about ways that SQL Server is evolving and changing, really focusing on solving the problems that are important to you and are left to be solved.
But what I wanted to do was talk about how in the very short term with the combination of Windows Server 2008 R2 and the upcoming release next spring of SQL Server 2008 R2, we’re going to really help to solve the highest-end, scale-up business problems, business applications that you have.
So what we want to do is show you a server rack we have on stage. This is a set of IBM servers. It’s a single server that is running SQL Server. It’s running both the pre-release of SQL Server 2008 R2 as well as Windows Server 2008 R2. And it’s quite a bit larger than any machine that previously used to be able to run SQL Server. The capacity is quite a bit beyond that.
So if we can switch to the demo screens, I’ll show you that in fact what we’re doing is we’re running this with not just 64 different processors, but we’re running it with 192 processors. So, we’ve increased the scale here dramatically.
And here’s an example of a work load that is running on 64 of those processors, the previous limit of what Windows and SQL Server could do, and it’s pretty much maxing out the CPU. But with this new set of releases, what we can do is we can expand that capacity and go to 128 processors and as we go to 128 processors, what happens is they all start to take up the load and they begin to run this business application very effectively with quite a bit of load available.
And then, of course, over time it’s possible that the application work load for the business might increase, and there might be a need to put more capacity in the system. So, in this case, you see the capacity — that the CPU is going up as the load is increasing, and we’re now once again to the point where the system at 128 processors is really at full capacity, but the head room that we can provide with SQL Server is quite a bit beyond this. So, on this machine, we can go to 192 processors. And as we do that, the work load once again comes back down. Pretty good, huh? 192 processors. (Applause.)
Now, what has really impressed me about this work is the architectural limits are quite a bit higher. They’re actually 1024 processors, but realistically, scaling up a database application above 64 different cores or threads is a very complicated thing to do. And combining the work that the Windows Server team has done with some major new changes in the kernel, removing some core locks. There was a lock that was removed called the dispatch lock that very much freed up capacity and removed contention in the kernel. Together with a lot of great work that the SQL Server team is doing, we’re seeing capacity increases on the order of 1.6 as you add incremental processors, which is quite good for the industry average.
I mean, if you look at the best systems that are running on Unix, for example, typically, that’s about the kind of capacity increase you get, or maybe even a little less. So, we feel very, very good that the combination of SQL Server together with Windows Server is world-class in terms of its scale-up, and very clearly this solves a significant amount of the scalability issues for very, very large, high-end business systems. And we’ll talk through this conference about other solutions we’re providing to both improve manageability of systems and to provide other solutions for high-end, scale-up and scale-out systems. Particularly, we’ll talk about what we’re doing around data warehousing, which I think is very, very exciting.
So that’s where SQL Server has come to. And the results for that are pretty solid. We’re seeing now that as we go ahead and implement these systems and run benchmarks against these systems, we’re getting some fantastic benchmark results, in fact, some industry-leading benchmark results with SQL Server. So, the combination of a IA-64 and SQL Server has now set, and we’re announcing this this morning, a world-record TPC-E number of 2,112 transactions per second, pretty amazing result for SQL Server to set that world record.
Now, that is an overall performance record for TPC-E, and it is a price-performance record at the same time. So, the combination of that just demonstrates the value proposition that SQL Server provides, together with this very high-end head room for business, transactional OLTP-based systems.
We’re also seeing some very stunning and powerful results for data warehousing with TPC-H, where SQL Server is now setting the record TPC-H number on Windows at 102,778 queries per hour, so that is a Windows-based record, but again, fantastic result, and fantastic price-performance.
Another benchmark that we’re announcing today is one for one of our own applications, Dynamics, where we’re seeing really great results for very high-end — in this case it’s a CRM Dynamics application. Twenty-thousand users, simultaneous users, sub-second response time, really, really great results. Again, demonstrating that SQL Server provides the capabilities that you need, of course in an environment that you’re familiar with, and with the price-performance that really cannot be matched anywhere else in the industry. And we’re going to continue to drive these things forward to continue to produce these incredible results.
So when I look around and I think about what is happening broadly, it is a time when there are a lot of changes happening in IT. And it’s a time that has a substantive impact on the roles of data professionals within an organization.
Some macro trends that are happening that impact us all as data professionals are the influx of people coming into the workforce with a different set of expectations. If you look at younger people who are entering the workforce, they’re used to using Twitter and Facebook and communicating in a set of social interactions. They’re used to working with information, finding things on the Web and combining things to get the answers they need very quickly, much more so than workers who’ve been in the industry for a while.
And in this, that offers a tremendous opportunity for us as data professionals to empower those workers to do their jobs much more effectively than they do today, to give them tools, give them access to information that they can combine and bring together to solve business problems and to get the answers that they need.
What you’ll hear from Ted and throughout this conference is the work that Microsoft is doing across all of our products — Office, SharePoint, and SQL Server together — to unlock business data to that broad set of information workers that are able to and are capable of working with information and building solutions to do insight and analysis.
And I think we have a set of tools that are very, very unique in terms of the accessibility they’ll provide to enable this self service for users. So, it’s a great opportunity for us as we think about our roles as data professionals and what we’re trying to do. It isn’t only about working within the IT department to solve the needs of the business applications that IT is creating, it’s certainly still that, but it’s also the ability to take and provide the data in a form that people within our organizations can all use it. So there’s a huge opportunity.
Simultaneously, what’s happening is the move to very, very low-cost memory, whether it be solid-state flash or very large amounts of RAM inside servers and end computers, means that the ability to work with vast amounts of data and make that available on low-cost systems has been opened up in ways like never before. In fact, what we’ll be doing is unlocking amazing capabilities on the average laptop that people are carrying around with them. And if you take a look at what can be done with a low-cost X86 server, it’s pretty amazing what’s about to happen as all of this memory becomes available and also for many applications, the ability for high-speed flash disks become available.
What all of that means is that as a data professional, our roles begin to change. We are critical to the role of the business, we’re at the center of the business, and in so many ways, computing, which is so broadly available at a low cost, is moving to where the data is, and the location and protection and maintenance of the data becomes a central role for organizations while at the same time there’s this tension of wanting to control the data and ensure that the policies that are right for that are maintained. At the same time, what we want to do is empower our end users to work with that information.
Our job is to help you by providing tools that help you to do these things, but the opportunities that this provides means that your role will expand over time in importance within the organization because of the fact that the data is so critical and the ability to work with that data, both maintaining business policies and enabling new solutions, becomes all the more critical for organizations as we move forward.
So lots and lots of great opportunities. And there are many things that are driving this change. I talked about the workers, I talked about how the new hardware is coming and the things that are making those sets of changes. But the same token, we’re also going through a shift, a paradigm shift, in the way information is managed and business systems are managed overall.
Now, one of the shifts that’s driving this very much is virtualization and the advent of virtualization being a technology that enables easier management of data centers and higher flexibility in terms of ensuring high availability and making sure that systems are able to run at the capacity and utilization that’s appropriate for them.
You know, we are on a journey as an industry, the whole industry is on a journey where we will see the evolution of the way applications are written transformed over the next set of years as new opportunities, new technologies open up new opportunities for applications to be created that scale to meet the needs of what the business organization has.
And this journey is a set of steps. And the database is at the center of it, but an enabler that will help us along this path very much is virtualization, and we see a trend, of course, within organizations to virtualize many aspects of their data center. In many ways, the database is one of the last things to be virtualized and yet we are now reaching a point where as we step forward, through virtualization, we can manage databases more effectively than we could before, and we can help to reduce our management costs somewhat through this process.
So the other point that’s very important is that as virtualization becomes standardized in the way business applications are run, the database very much is placed at the center of those business applications because all of them, regardless of what virtual system they’re running on, need to have consistent and coherent access to the information that’s stored in the database.
So we’re on a journey with virtualization right now and it’s one that will help all of us as industry professionals to provide better service to our organizations. It’s a key component of that.
What I’d like to do now is show you a demo of some of the work that we’re doing with virtualization and Windows Server and Hyper-V and some of the impacts that will have on the database and the ability to virtualize the database. And so if I can invite EDWIN YUEN up, and EDWIN YUEN’s going to join us to show us this demo. EDWIN YUEN.
EDWIN YUEN: What we’re going to show today and we’re going to show the demonstration system is how a great new feature, Live Migration in Windows Server 2008 R2 Hyper-V, can really make SQL Server more highly available and load balanced.
So what we see here is Virtual Machine Manager 2008 R2, which is our virtualization management product, and we’re managing a cluster of Server 2008 R2 boxes. And on those virtual machines, we’re actually running one that’s running SQL Server.
So what we can do with this is we can actually go ahead and migrate it from one node to the other. And what I’m going to do is I’m actually going to launch this program that loads a stored procedure and creates a worker thread, so we’ll actually generate some loads.
BOB MUGLIA: So you’re putting some load in. One of the things that people have been concerned about with database applications is the effectiveness of the I/O subsystem in virtualization. We’ve done a lot of analysis of this with Hyper-V and have come to the conclusion that while there’s no ambiguity that virtualization puts some overhead on systems, there is a small amount of overhead associated with it, the capabilities that Hyper-V provides in terms of providing high-performance I/O are very, very strong. In fact, one of the things we looked at when we launched Hyper-V 18 months ago was our performance relative to VMware.
And we looked across the board at that overall, and frankly, ESX has always done a pretty good job, so we expected that we would fall short. When we did the analysis, we determined that Hyper-V was very comparable to VMware pretty much across the board. And one area where we’ve excelled relative to VMware, frankly, is I/O performance for running applications like SQL Server.
So it’s a very good environment to virtualize and get some of the benefits of consolidation. And, clearly, that’s an opportunity to reduce some costs.
EDWIN YUEN: Absolutely. So, what we’re going to do now is I’m actually going to go and hit the Migrate button. And what Virtual Machine Manager will do is automatically go out and look at all of the available nodes and analyze the available I/O, disk, and memory and pick the best available host. And we’ll hit Next and then we’ll hit Move. And what it’s going to do is it’s actually moving that virtual machine right now, and we can see the information.
And as you can see, the process still continues, there’s no delay, there’s no loss of connectivity to the SQL Server itself. And it’s much more than just having the live migration. We can actually link this to our performance resource optimization product, and then I can actually dynamically move these virtual machines based on load or memory or other factors.
BOB MUGLIA: So in this case, there were transactions running consistently, the transactions didn’t skip a beat at all, they kept running, Live Migration does that seamlessly behind the scenes. And this opens up a set of opportunities to simplify management. And as I said, also to provide cost savings by consolidating pre-existing systems that may be under-utilized on modern hardware.
EDWIN YUEN: Absolutely.
BOB MUGLIA: Great.
EDWIN YUEN: Thank you very much, Bob. (Applause.)
BOB MUGLIA: So that’s our technology that’s available today and it’s an opportunity to take advantage of it. One of the things that we’ve also been excited about is as we work with our customers, you know, we see the ability for customers to save a great deal of money over the competition by working with and just using the facilities in Windows Server 2008 R2 as well as the management tools that Microsoft provides. We have a very comprehensive set of solutions that simplify the administration and at the same time provide a much lower cost than the alternative. And it’s something that I think is very important in terms of the way we think about the evolution of the data center, the role of the database within that data center and then the long-term implications as we move to the cloud.
Now, over the next few years, what we will see is applications being rethought to take the next step and think about how they can run more effectively in a cloud environment. And it’s worthwhile for me taking a second to define the word “cloud” because, in fact, it’s a word that’s used pretty broadly in the industry and it’s misused a lot; it’s used to mean different things.
And if you take a look at what Gartner or Forester or any of these folks have to say, their definition is all relatively similar. The cloud is about providing IT as a service to organizations either within a company or across different companies. It is about providing a pool of computing resources that all operate together, effectively as a single computer. You can think of it as tens, hundreds, or thousands of computers all connected together to provide this computing and business resource for the organization.
And one of the things that clouds enable is the creation of elastic or scale-out applications, applications that respond to the needs of the business by utilizing more resources as they’re needed, then giving them up when they’re not needed anymore.
We had a classic example of this that Microsoft was running inside our own cloud, Windows Azure, that we used internally for the company just recently. We’re just completing our annual Giving Campaign where all the employees at Microsoft can go into that application and provide donations to the charitable organization of their choice.
And this is a classic application where for most of the year, it’s not being used at all, for a few months of the year it’s used relatively lightly, and in about a 36-, 48-hour period, it’s used incredibly heavily by a large number of applications as they all meet the deadline and actually sign up to give.
And in this case, what that application did is it scaled from essentially just a few computers that were running the application up to about 24 computers total to handle the full load that everything was required for that one period, and then scaled right back down. That kind of elastic computing capability is something that I sometimes think of as the fifth generation programming model.
If I kind of go back in time, I mentioned client-server earlier, the first programming model I would call the monolithic or mainframe programming model. That, of course, was created in the ’60s and ’70s with terminals and mainframe computers. In the late ’80s, with SQL Server, we introduced, the industry introduced the client-server programming model. In 1995, we saw the creation of the Web application programming model, and in roughly 2000, we saw the creation of SOA or Web services-based programming model.
Well, now what we’re seeing is an evolution of all of those, because each one of them builds on the other, to create this scale-out programming model, this next generation. And that’s what clouds will enable.
At first, what we will see is clouds building these things, companies building these things within their own cloud environment within their own organization, and that will be important because companies will want to maintain control of that data. So private clouds will be a very, very important thing. And one of our focuses as we move forward with SQL Server and our management tools and Windows Server is to enable companies to build their own private clouds.
But we’ll also see public clouds being created where you have many, many different companies working together on a broad set of shared resources. And those public clouds will be created by companies like Microsoft with our products like Windows Azure, our services like Windows Azure. They will also be created by many hosters, which will build their own public clouds for use within organizations to get computing and data resources available.
The key from our perspective is that we see this as one of the natural transitions in the industry. It very much does affect the data professional, and we’ll talk about that in the context of the evolution of our database, SQL Server, as we’ve learned from and have implemented SQL Azure in our own public cloud.
So we’re learning a lot through this process, but we’re learning a lot and opening up capabilities that will be very useful to you in your organization, both simplifying your job and providing new opportunities for the data that’s so critical to be managed more effectively. The key from Microsoft’s perspective is regardless of where you want to implement your applications, whether you want to use a traditional data center, whether you’re implementing virtualization, you have a private or dedicated cloud that you begin to implement, or you’re looking at taking advantage of some of the public cloud capabilities, we will be focused on providing a coherent programming environment and management environment across all of those.
So the skill sets that you have today with SQL Server will be highly leveraged going forward. When we show you SQL Azure, it will look familiar to you and the skill sets that you have today will be very helpful for you and will put you in a very good position as these next-generation clouds begin to be introduced.
And, I mean, that’s one of the differences where Microsoft can provide a unique situation for the industry and for our customers. Unlike other organizations, Microsoft has a very large install base of committed users and systems that run Windows Server and SQL Server, and at the same time, we’re investing hundreds and hundreds of millions of dollars in building very large data centers that have amazing amounts of capacity, capacity which we literally measure in megawatts. A typical data center has between 35- and 100,000 servers in it. So, these very massive data center servers.
It turns out that in order to understand how to build software to run effectively in these large data centers, it’s really important we run it ourselves. It’s an important thing because we learn how to do that. We are learning things today with SQL Server Azure which are being directly applied and will help with SQL Server in the coming releases. In fact, some of the administrative features like the DAC that exists in SQL Server 2008 R2 is something that we’re working to build in a compatible way between SQL Azure and SQL Server.
So learnings from our cloud experience are coming back to the work in the products we’re delivering on-premises. And of course the knowledge and all of the feedback and all of the partners and all of the history and information you have is critical as the industry moves forward into the cloud. So it is a virtuous circle that connects these things very much together in a cohesive way.
And, again, our focus will be to enable you and your organizations to work with and build business solutions in the way that makes sense for you at the pace that makes sense for you, leveraging the skill sets that you have all along the way. So whether you wish to build your next-generation applications in your own data center, to work with the many thousands of hosting partners that exist out there, or work with Microsoft in our own data center with Windows Azure, those choices will all be available, and they will be available in a way that is consistent and, again, leveraging that very important skill set you have.
So as we move forward, I think it’s a very interesting time. It’s a time when there is a whole set of things that are happening in terms of transformation of IT. And in the process of how IT is being transformed, your role is also transformed.
The thing that I think is important here when you think about the high-performance servers with lots and lots of memory, when you think about end users that have new capabilities, new abilities and new demands, when you think about the advent of the cloud and what does that mean, I think the key is that the skill sets and the knowledge that you have are critical skill sets for the future because the information that you manage, the data that you manage, is at the center of the companies and the organizations that you work for.
And the opportunities that will exist for you as you move forward is to take the skill sets that you have and then to be able to leverage that skill set to help your organizations, help your end users, help your customers to solve new problems, problems that were very difficult or maybe impossible to solve before.
And I want you to know that we’ve been working on this for a long time. I mean, I go back to the beginning with this product and it really started a long time ago in a very humble and simple way. But through your feedback, we’ve been able to really help you to solve problems and to bring the industry forward.
We very much appreciate that. What I think you’re going to hear over the next couple of days is how Microsoft is very much here for you and there’s a whole lot of great things that are coming your way, but it’s only possible because of the great work that you do and it’s fantastic to have an opportunity to share that with you and be here today.
Now, with that, what I’d like to do is introduce Ted Kummert. Ted is the senior vice president responsible for SQL Server. Ted actually now has a bit broader responsibility. He’s responsible for our overall Business Platforms Division, which means the business applications, Application Server, the whole infrastructure around that. And Ted has done a fantastic job of driving SQL forward with the SQL Server 2008 release and he has a fantastic road map and an execution ability to drive forward with the releases that are coming next year and into the future. He’s the man that’s making it happen for all of you and it’s great to have him here today. Welcome Ted Kummert. (Applause, music.)