Remarks by Steve Ballmer, CEO, Microsoft Corporation
Microsoft Management Summit
Las Vegas, Nevada
April 20, 2005
ANNOUNCER: Ladies and gentlemen, please welcome the Chief Executive Officer for the Microsoft Corporation, Steve Ballmer. (Applause.)
STEVE BALLMER: It’s a little bit of cognitive dissonance, I don’t know if you noticed it, but we had the very quiet, gentle piano music for the video and then there was a little bit of bump and thump when I came out on stage. I guess we’re trying to mix the old and the new all together here for Microsoft.
It’s a real honor and privilege for me to have a chance to be here with you today. I’m particularly excited to be here because, for those of you who attended last year, I was a no-show. I don’t often no-show; in fact, I can only remember once in the 25 years I’ve been at Microsoft that I no-showed for a customer event and that was the MMS last year. And I really feel an urgent need to talk a little bit about that to sort of get me into the discussion for today.
This is I guess our fourth MMS. How many people have been here all four years, anybody? How many people have been here three years? Two? One, first MMS? OK, good, good; good to see some folks staying with us, new people coming.
I would say MMS has had kind of a gradual evolution. Four years ago, I think it’s fair to say depending upon your point of view and the level of optimism you bring to life, we were in pretty tough shape; the products weren’t that great, they were very v.1, shall we say, and we were getting a lot of feedback from folks about how v.1 they were. And certainly that first MMS, the second MMS we had a lot of that kind of feedback. And if you look at the feedback forums from the MMS two years ago, there was still a lot of, hey, come on, what’s going on with SMS, how about MOM, blah, blah, blah, blah, blah.
Last year’s MMS, we turned a corner. The only major complaint we got in the feedback forums last year was, “Where’s Ballmer? Was he ducking all the feedback on SMS and MOM?”
Well, it turns out last year at this conference I had to meet with Commissioner Monti of the European Union and learn subsequently that we were going to owe them $575 million, so I had a good excuse I thought, but never good enough for the MMS crowd.
So I am particularly excited to be here this year, I know we have a great set of products, the feedback has been wonderful from all of you, not that we don’t still have plenty to do with the products. I know we took care of the problem that said Ballmer didn’t show up this year, I’m here, and hopefully the combination of great products, great mixing amongst yourselves, some good presentations will really leave you feeling very confident about the investments that you’ve made in our management products and where we’re going in terms of manageability, quality of management and total cost of operations.
I know Kirill [Tatarinov] had a chance to talk to you yesterday, and I had a chance to kind of review what he said and read the transcript, talk to him about it, and I think he did a really good job in general of summarizing broadly our strategy, where we’re going and what we’re doing.
I’m going to show you a few new things today, I’m going to talk to you a little bit at the high level though about the importance and commitment we have towards management and what we think that brings to you and the organizations that you represent.
We talk a lot inside Microsoft with our customers and with our partners about our mission. And mission statements are fuzzy, you know, every organization, probably every one of us has ever been in, has had a mission statement. And most mission statements people ignore most of the time, at least that’s my experience. Our company has always had a mission statement that people took very seriously. It used to be a computer on every desk and in every home and then we expanded it to this: Enabling people and businesses throughout the world to realize their full potential.
I think that’s the mission not only for our company, but really for the information technology industry. We all in this room, whether we are on the vendor side or we are customers and users of IT, implementers of IT solutions, managers of that, our mission in life is to enable the businesses and people that we serve to fully accomplish their potential. Your businesses are pushing on you every day how do you enable us to do things we didn’t think we could do before, how do you enable us to do that at low cost so we continue to have resources to put into other IT activities or non-IT activities.
That’s what we’re all about, we’re about enabling the businesses that we serve and the individuals that we serve to do a better job. And for me, for our company and I think for everything in the IT industry that should be a galvanizing mission.
You can’t write a mission like this down for most other things in life. You can’t write it down for the auto industry, a great business but it’s not about helping people achieve full potential.
Even as we think about what we’re trying to do in management, this mission shines through: How do we give you the capabilities to set up, manage and configure the environment that lets you as individuals realize your potential.
It is really hard to serve the IT needs of small, medium or large corporations, and how do we enable you to do more with less, to get better insight into what’s going on operationally, to deliver better quality of service and SLAs, to deliver the services that you provide at lower cost. Every day there are folks at Microsoft who come in and just think about that, and think about that and think about that, and we have a real dedication and a real commitment to improving in this aspect.
A lot of people like to remind us that we grew up on the desktop and there are still people who will say to me, “Are you serious about the enterprise, are you serious about this stuff?” What that usually means is are you serious about being our partner in the way we manage operations in the datacenter and to the desktop in large scale environments, from hundreds on up to many, many tens of thousands of desktops and servers; do we at Microsoft understand that we need different technology foundations and most importantly are we committed, are we committed to be excellent in enterprise management, are we committed to give you a platform and a set of management tools that delivers the lowest total cost of ownership in the industry bar none, are we committed.
And if you get nothing else out of my speech today, I want to make sure of our commitment to being an enterprise class supplier of the lowest cost TCO platform from Windows on up our stack, and providing what you need for enterprise management tools that work at the kind of scale, that work with the kind of interoperability and heterogeneity that’s required in the organizations that you serve.
I’m going to talk today also about a major area of increased investment for us in virtualization, because I think as we think about the future of cost in management, improvements in our virtualization technologies, enhancements in our virtualization technologies will be a key enabler both of our enterprise management suite and of driving cost of ownership out of the platforms from Microsoft that you take care of.
So we are very committed on this dimension and we’ve come a long way. I think back to the late 1980s. We just had a very important milestone, we just shipped the 64-bit version of Windows client and Windows server. People ask how quickly will 64-bit take off. I could get onto that, but I want to put it into a little context. It was just 16 years ago that we started work on Windows NT. Sixteen years ago we began a journey to be very serious about enterprise computing, starting with the core foundation, the kernel, the platform. We’ve extended on that investment, we’ve improved the operating system, we’ve come a long way in terms of facility. And really in the last five years is where I think we’ve made the big advances in terms of management, particularly at the enterprise level.
I was serious, I know four years ago at the first MMS we were struggling and you were telling us how we were struggling at management. I’ll meet with a number of you while I’m here and I’m sure I’ll continue to hear about things that need to improve but our team today stands very tall because we get lots of feedback about our tools for enterprise management being amongst the best, if not in many cases the absolute best available in the marketplace today.
So we’ve come a long way, we’re committed and we’re committed to go much, much further down this path and I’m going to talk about some new things in terms of our enterprise management, our commitment to lowest total cost of ownership and what we’re doing with virtualization as a key enabler.
Three Buckets in Enterprise Management
In enterprise management I want to center on three buckets. First is DSI. And if I was to characterize what I think Kirill had a chance to talk to you about yesterday he really talked about how far we’ve come in DSI and what’s going on and I’m going to add just a little bit to that.
No. 2, I want to talk about intelligence because at the scale of the enterprise it’s not enough to just give you good tools to manage and maintain an SLA. We’ve got to give you good tools to help you get intelligence about things that are systematically working and not working, to be able to really understand in a large environment not just the incident of the day but to be able to go back and say, hmm, are we having problems statistically that are higher in some locations with some equipment, with some vendors, with somebody’s code. You need that data and we need to provide you not just operational tools but the ability in some senses to step back and get business intelligence on your operations and your IT infrastructure.
And certainly I want to talk about interoperability. We grew up focused in on Windows, managing Windows, taking care of Windows. Today I want to mark essentially a step forward where you’ll see that our dedication now is to providing you the kinds of tools that you need to manage heterogeneity in your datacenter and we’ll show you a little bit of that here in just a couple of minutes.
Dynamic Systems Initiative
First on DSI, the Dynamic Systems Initiative: The Dynamic Systems Initiative was really all about expressing knowledge in models across the full app lifecycle from development on through to operations. Kirill had a chance to talk about MOM 2005 and the work that we’ve done to capture that kind of model information in Management Packs. And I think we’ve come a long way, you see that every day in your operations and you had a chance to hear about it some more yesterday from Kirill.
The next step is to really put SDM, the System Design Model at the core, to make it the backbone of what we do and we have a major milestone this year that cannot be understated. We’re within now several months of shipping Visual Studio 2005 where we really deliver to all software developers the ability to encapsulate management model information into their applications from the point of creation, so the ability to sit there in the development tool, write the code and build in the management model at the same time becomes available. And there really is no reason why new code created after we ship Visual Studio 2005 shouldn’t all rise to the kind of manageability that we’re able to show today with a number of these products with the MOM Management Packs.
Kirill also had a chance to talk about the second wave of System Center and how we’re going to do more to consume SDM models in both MOM and SMS, which I think is an important step forward.
When we announced DSI a couple of years ago it was a vision and frankly analysts were saying to us it feels a little bit fluffy. Today it’s a very concrete architecture for our management technologies that’s very well implemented in our System Center product suite and later on this year with Visual Studio 2005.
So I think today it’s fair to say DSI is very, very real.
I want to have Francois Ajenstat from Microsoft come up on stage. We’re going to demonstrate a little bit to you what I mean by intelligence. Kirill had a chance to talk about System Center reporting yesterday. It is very important to us, as I said, for us to put tools at your hands for you to be able to analyze what’s going on overall in the datacenter. There are so many alerts, there is much going on, there’s so much data to be processed, but for all that data that needs to be processed every day to keep systems up and running and SLAs going you need, if you’re going to take cost out, the ability to step back and ask big questions about what’s going on in your operation and we want to show you some of the ways you’ll be able to use our new SQL Server 2005 product with the System Center reporting stuff to get better intelligence about what’s going on in your operational infrastructure that should help you improve quality and reduce costs.
Please welcome Francois. (Applause.)
FRANCOIS AJENSTAT: Thanks, Steve.
Well, today you collect large volumes of information about your systems and applications and gathering all of that in your databases so you can actually report using the MOM Management Pack on that information. However, it does require an operations manager or a DBA to author and share those reports with the rest of the organization. With SQL Server 2005 we want to empower a broader set of users in IT to get access to that data so they can get better intelligence or better operational insights on that data.
So in this case on the demo we’re going to be looking at our management environment and here are all the reports that are available for my various applications. With SQL Server 2005 we’ve added a new capability called Report Builder, which enables me to author a report on the fly without having to have a lot of knowledge on the back-end database.
STEVE BALLMER: So you’re now pointing the new SQL Server Reporting Services at data that comes out of System Center?
FRANCOIS AJENSTAT: Correct.
So we’re going to connect to our System Center database and this is a very simple environment for an IT user to create the report. So this is going to be a report about the various events that occurred in our systems. So let’s look at our event types, just drag and drop the information on our screen and let’s look at it by the various categories of events as well as the source of those events and finally the number of events that occurred.
There you go. And, of course, we don’t want to get all the data so let’s add a small filter so we can only get a subset of that data. So let’s look at the events that occurred in, say, 2004 and the month of August, so month number eight.
STEVE BALLMER: Is that the month I was having trouble getting information for our 10k filing?
FRANCOIS AJENSTAT: That’s it, that’s why it’s picking it up.
STEVE BALLMER: I should have asked.
FRANCOIS AJENSTAT: So you can actually use this to author the report, you don’t have to go back to IT to get that report authored for you.
STEVE BALLMER: Check up on the guy.
FRANCOIS AJENSTAT: Uh-huh, pretty good.
STEVE BALLMER: That’s easy. Go ahead.
FRANCOIS AJENSTAT: Now we can actually go ahead and run that report.
Now, traditionally you’d actually have to go into Visual Studio to author that report and you have to understand on what the database schema is and bring that to the screen. And you saw how easy it was to just drag and drop the elements I wanted.
So here are all the events that were generated on our system and normally whenever you have a report, Steve, you actually want to get additional information, why is that number that number. These are actually interactive reports so I can actually drill into the report. So let’s look at our setup events and Report Builder will automatically create a new report on the fly and pass that context to that new report so now I’m only looking at those setup events. And there’s all the information about the events that occurred in August 2004.
Now you’ll notice that a number of these events actually have these XML tags while others don’t. Well, these are actually coming from the “Longhorn” clients that I manage in my environment. And with “Longhorn” I can actually drill down and get additional information. With “Longhorn” we’ll be providing you with more complete and useful event data so you can manage your “Longhorn” clients more easily. These events will be schematized using XML so you can more easily sort, search and filter on that data.
STEVE BALLMER: And this is the help model stuff that we’re building into Longhorn?
FRANCOIS AJENSTAT: Correct.
So if we go back to that report that we created, once we have that report the way that we like it we can then go ahead and save that to our server. And doing so will enable us to share that with others on our teams, we’ll be able to subscribe to that report, be notified when the data changes so I always get the latest and greatest information from this report that I just authored.
Now, of course, because there’s a broader set of users using this data they want to have the information delivered in new ways. Well, with SQL Server 2005 we’ve increased the integration with Windows SharePoint Services so now those reports can actually be delivered directly in a Web part in SharePoint. So here’s that new report that we created, we can click it and that same report we authored just a few minutes ago was rendered on the fly through the Web and, of course, that same interactivity, all the same capabilities that you have through the server is available directly here as we had in Report Builder.
So really what I wanted to show you is how easy it was to get operational insights on our data by leveraging the new capabilities in SQL Server 2005. In SQL Server 2005 we’ve really enhanced our business intelligence capabilities, enabling you to better integrate, analyze and report on data that you have both in your operational systems and across the organization.
SQL Server 2005 will be available in the second half of this year but it’s actually available today to download as a preview so you can test that out and get these capabilities in your organization today.
Thank you very much. Thanks, Steve.
STEVE BALLMER: Thanks, Francois. (Applause.)
I think you get something of a sense of why and how we think we can put better intelligence in front of you, how do you drill through the massive information, what tools do you have and what does that let you do to understand cost and quality of operations and then make that data very widely available.
In a sense I think I would say it’s kind of this notion of how do we help folks who manage IT infrastructure to realize their potential, giving you the knowledge-based tools to actually drill in, look in, understand and share that understanding through intranet portals with other folks in your organizations who need to have the information.
A major theme for us and one that you could see in a sense the powerful synergy between what we’re doing with our database, what we’re doing with the portal and the basic investment that we’re making in management tools and management infrastructure, one plus one really is three as these things come together.
The next thing I want to talk about is interoperability and heterogeneity, a very big point. I can say this many, many times, I say it many, many, many, many, many, many, many, many times, Microsoft is committed to interoperability, and I’ll still get people who will scratch their heads and say, “Are you really committed to interoperability?” And I’ll say, “Oh yes we are, look at all we’ve done. Look at our services for UNIX and how we’ve driven that ahead, look at the work we’ve done with IBM on standardizing Web services interfaces, we’re very committed to interoperability.” And people will say, “But are you really committed to interoperability?” And I’ll say look at the relationship we announced with Sun a year ago, we are really committed to interoperability. And still people will look at me in kind of a quizzical way.
I want to make sure today you understand that from a management perspective we are absolutely committed to interoperability. We’re committed to work in broad industry partnership to deliver great interoperability. You’ll see us in Windows Server release 2 later this year deliver integrated and improved services for connecting UNIX systems of all forms to Windows, tools that will allow us to federate identity through Active Directory across systems. You’ve seen our deep commitment to XML.
And today we want to show you something, the first time ever using some of the standardization work around XML Web services and particularly WS-Management to get stronger interoperability from a management perspective.
Very specifically we’ve worked closely with Sun — yes, Sun, the people we never used to work closely with, that same Sun — we’ve worked closely with Sun over the course of the last year.
Scott and I are going to do kind of a state of the union report on where we’ve gotten in 12 months here in a few weeks but we thought we’d show you a little bit of some of the good work that’s come about.
We’ve really worked closely on WS-Management in conjunction not only with Sun but a number of other players in the industry, AMD and Intel, BMC, Dell and many others.
And what we’re getting to today is a very strong form of interoperability where from within System Center you can see and manage and look at events and take action on events that come from UNIX systems, from switches, from a wide variety of non-Windows equipment.
We’ll always be I suspect first and best on Windows but we want to take advantage of the great management work other people are doing on their systems, the kind of great management work Sun is trying to do for their systems and make sure you can interoperate, collect and work on all that information from within System Center.
So it’s a much broader vision, it’s a much more heterogeneous management vision, it’s a much more enterprise oriented management vision than we have really pursued before but because of XML Web Services and interoperability we’re after it now.
So we thought we’d show you, together with some of our colleagues at Sun who are here in the audience today, we thought we’d show you just a little bit of that in a demonstration. Please welcome Bill Anderson from our product management team who’s going to give us a little bit of a demonstration of heterogeneous management. Bill. (Applause.)
BILL ANDERSON: Great. Thanks, Steve, appreciate it.
Absolutely agree with Steve on the criticality of interoperability and the work that we’ve done with Sun, Intel and an array of partners over the past year. Hopefully we’re going to enable a lot of those scenarios.
Now, WS-Management is also important not just from an interop standpoint but it enables a couple of other things. First of all, it allows us to better manage sites and sources outside of our corporate intranet. Using Web services as a protocol, it makes it a lot easier in those kind of business-to-business scenarios.
We can also manage things in an almost agent-less way. In fact, this stuff I’m going to walk through today all feeds into the MOM console, but it’s agent-less, we’re using WS Man remotely to be able to pull that information.
And then finally it allows us to actually view the hardware without the operating system present, so in some situations where we don’t have an OS, it’s a new machine, we can actually remotely go manage that.
So for the course of this walk through I’m actually going to be an IT professional but a little bit of a different twist, I’m an outsourcer. I’m actually managing services outside of my corporate intranet and we’re going to walk through a couple of ways that WS Management can actually bring you guys some benefit.
STEVE BALLMER: So basically all this is over the firewall?
BILL ANDERSON: All of this is over the firewall, HTTPS remoted across so we can do it in a secure way.
Now, if you take a look at the sets of screens you’ll see on the outside screens blackness, not an error, it’s supposed to look that way. We’ve got a Windows server that for some reason I don’t have up and running and I’m going to go troubleshoot that and turn that on.
As you look at the inside screens you’ve got my MOM topology view here, and, of course, the Steve cam here that’s going to be watching Steve and I as we play around with some of these fun hardware pieces over here. We’ve got a Sun rack here with a series of sun servers inside of it. We also have a Dell box running Windows over there that I’m actually running hosted MOM from. And so while Steve is gesturing and they’re photographing, these are the things that we’re going to be playing around with for the next couple of minutes.
So as I mentioned, I’ve got a problem, I’ve got a Windows server in one of my remote locations from my customers that doesn’t seem to be up and running. When I take a look at my MOM topology view I’ve got this Sun —
STEVE BALLMER: Where’s that Windows server?
BILL ANDERSON: It’s this Sun Win 1 right here.
STEVE BALLMER: That’s Windows Server running on this Sun box?
BILL ANDERSON: That’s correct.
STEVE BALLMER: Just thought I would ask. (Laughter.)
BILL ANDERSON: That’s a Sun piece of hardware running Windows. (Applause.) Yes, it’s great for us to be able to have this to be able to do this across all of these different processor platforms.
So we’ve got this problem. For some reason I don’t know why it isn’t up and running. I’ve been having some intermittent power problems. I think that’s probably it, but I’m going to use WS-Management to troubleshoot that.
I’m going to go open up my state view in MOM that’s actually going to give me a little bit more detail on my topology here. You’ll notice this middle box, Sun Win 1, shows me that, from a hardware standpoint, I’m actually healthy. That tells me that the service processor on the device is up and running, but from a software and an OS standpoint, it’s showing red. So the device is plugged in, it’s powered on; for some reason that particular device isn’t up and running.
Normally I would call a tech or give you a call if you’re on site, you’d go turn it on, but that’s kind of difficult. I’m going to first of all use WS-Management to turn that item on. All I have to do is go select, power on. If you view our outside screens in a few seconds we’ll actually start booting up and showing that that Windows server is now actually up and running in that remote datacenter.
Then normally I would just go ahead and do this, restore the service, forget the problem and go on about my business but this has been a recurring problem. So I’m going to dispatch a tech and I want to give that tech some data so he can fix the problem more easily. WS-Management is going to help us with that.
The first thing I’ll do is I’ll close this window out. I’m going to go ahead and actually run a remote hardware inventory from the service processor. Using WS-Management remotely it’s querying the service processor for the hardware data that’s there. This way I can pull back power supply data so when the tech gets there he knows which one he needs to have and replace.
STEVE BALLMER: So now you’re talking to the Sun system overall, we’re not talking particularly to Windows on a Sun system, we’re talking to the Sun system?
BILL ANDERSON: I’m talking to the service processor on that particular device, that’s correct.
STEVE BALLMER: Good heterogeneous, interoperable management.
BILL ANDERSON: Great heterogeneous, interoperable management. (Laughter.)
So you notice here I’ve got my numbers for my power supply, I can copy and paste that into an e-mail for the tech when he or she arrives. But for those of you who have worked in a rack environment you also know that the servers may move around and be in different places. Well, I want to be able to actually better enable him upon getting there to know which one to go fix. So this may seem kind of trivial but it’s super important. I’m going to go turn on my chassis identification light or my indicator light on my chassis. What this is going to do if you take a look at the top server in the rack mount it’s going to flash an orange indicator light so that when the technician gets on site he knows exactly which server he’s going to go fix and exactly which power supply he needs to replace.
So now at that point in time we’ve got this great interoperability story but you guys I’m sure are sitting out there going, “Sure, it’s on Windows; show me how you can do that with a non-Windows operating system.”
STEVE BALLMER: How about Solaris?
BILL ANDERSON: How about Solaris. So let’s go ahead and choose phase two of the demo. We’ll change the outside screens and what you should see is at this point in time we’ll be moving to a Solaris server that’s up and running here and another black screen.
So in this same location I actually have a series of Solaris servers that are there, a primary one that’s up and running a business application and then a standby one in case something goes wrong. Well, now historically at this same location it’s a small business for me so I’ve got a CEO kind of person there who likes to dabble in IT a little bit, every once in a while, goes mucking around in my datacenter. (Laughter.)
STEVE BALLMER: Not that this is a very representative case of CEOs who shouldn’t be allowed to dabble in IT, and I shouldn’t either, but go ahead. (Laughter.)
BILL ANDERSON: So here we go. So I’m going to ask for a little help on this one, Steve, and I want you to go pull one of the fans out of that top Sun server.
STEVE BALLMER: That’s the kind of thing CEOs do in IT, they break things. (Laughter, applause.) Really, I’ve got to ask this; you’re asking me to break a Sun server?
BILL ANDERSON: I’m asking you to pull a fan out of a Sun server. (Laughter.)
STEVE BALLMER: I bet a year ago I would have felt differently about this, but now it’s on a partnership. (Laughter.)
BILL ANDERSON: So he’s got fan 1 out.
STEVE BALLMER: Fan 1.
BILL ANDERSON: At this point in time I’m going to wait a couple of seconds and then I’m actually going to refresh my topology view.
STEVE BALLMER: This thing is starting to race like a jet airplane waiting to take off.
BILL ANDERSON: Absolutely. So I’m just going to go and let’s make sure we get this refreshed. Oh, I can hear it picking up. We’ll go encourage MOM to speed this up a little bit. Oh, look, we’ve got all kinds of errors are kicking off. Bad things are happening. We’ll refresh the topology view again and now I’ve got a warning alert. It’s not critical yet but it’s actually giving me a warning alert. If I go look in detail at this and drill in you can actually see we’ve got a warning alert here that’s indicated using WS Management that I’ve had a fan that’s actually gone out. Now, of course, it doesn’t know that Steve’s pulled it, it just knows that it currently isn’t working.
But I can still run. Steve, grab another fan.
STEVE BALLMER: You want me really to take a second fan out of a Sun system? Okay.
BILL ANDERSON: Absolutely. So Steve has now pulled the second fan out and we’ll give this a second to actually refresh while Steve is actually tossing them out. And this at this point in time is actually going to go critical on us. So what we would basically have is we would have a critical failure up here, I’ve got some other ones that are there, let’s go find a refresh for this one. It’s not giving us critical yet. Oh, there it is. We’ve got one. It’s actually an older one but it’s the same error that we would have gotten showing that fan 2 has actually been replaced. Remember, we’re doing this over the Internet so there is a little latency that’s involved.
So now what’s happened though is this is critical, this means that either the server could go down and we could damage the hardware, it may actually cause a fire so Steve might be in some kind of jeopardy, so what I want to do is be able to remotely solve the problem for you as opposed to having to coach you through putting those pieces back in.
Now, fortunately Steve does this enough I’m used to handling this. (Laughter.) So I’ve got an automated task inside of MOM that’s going to allow me to do this failover to backup scenario. It’s going to allow me to actually bring the primary one down and bring the backup up so that we don’t incur any further problem.
There we go. At this point in time you actually should see this one over here on your left shutting down, the second one coming back up; for those of you who are close the airplane landed, so the rack is actually shut down, the fans have stopped.
So by using things like WS-Management we can not only manage the Windows platform but we can interop and manage non-Windows platforms with a consistency and a low operational cost that really makes this solve your enterprise needs.
STEVE BALLMER: Thanks, Bill; nice job. (Applause.)
Because I’m a good guy I’m going to put the fans back in and let things go.
That kind of heterogeneous interoperability not only are we committed to but the work we’re doing together with Sun is actually very instrumental in facilitating and enabling the scenario you saw there, agreeing on standards, making sure we get those well implemented across Windows systems, Solaris systems and many, many others.
Commitment to Lowering TCO
I want to shift now from enterprise management, DSI, intelligence and interoperability, to lowest total cost of ownership platforms and what we’re trying to do in Windows, in SQL Server, in Visual Studio to really drive — we are absolutely committed to lowest total cost of ownership in our platform.
There’s four areas I want to touch on that reinforce over the next few years that commitment. I want to talk about what we’re doing in “Longhorn,” I want to show you a little bit of what we’ve got coming for deployment, I want to talk about security because I know security has been a major cost driver for folks from an operations perspective over the last several years and then I want to show you where we’re going with virtualization and what that means both for lowest total cost of ownership as well as for enterprise management.
First let me just do kind of a recap on “Longhorn.” “Longhorn” has been a long time in the making, a longer time probably than we had originally intended but what we’re going to have when we ship “Longhorn” next year is I think really quite a phenomenal release. It’s got six key pillars. I think what most people know on a historical basis doesn’t really reflect what we will ship next year for “Longhorn.”
“Longhorn” Client Management
The new “Longhorn,” the one we’ll ship next year, is really a platform for the next 10 years. What we’ve done in terms of being a good, intelligent, edge platform to the Internet will ensure that we retain an environment where there’s intelligence not only at the center of the Internet but also in the PCs at the edge.
We’ve built 64-bit and a variety of other capabilities in, which will be important for the platform for the next ten years.
There are improvements in a wide variety of experiences at the client level for people at work, at home and mobile devices on the road, searching, finding information being one of the primary enhancements and improvements that we’ve made inside the “Longhorn” client.
From a management and operations standpoint we’ve designed improvements to make sure Windows really just works, not a lot of complexity and hassles, continuing to reduce that.
At the server side we’ve put a lot of effort into componentization of Windows so that you can deploy Windows Servers in a way that allows them to do just the job you want them to do and doesn’t leave a lot of exposed surface area to manage, to secure, et cetera.
We’ve improved safety and security in “Longhorn” and we’ve done a lot of work to make “Longhorn” easier to deploy and manage.
Those are the six big pillar of improvement and progress in the “Longhorn” release, which will come out next year, “Longhorn” client first followed by a “Longhorn” server within about six months.
From a client management perspective I think it’s important for me to emphasize some of the improvements that we’re making. We talked a little bit in the intelligence section about the additions we’re making in “Longhorn” to the model-based information that’s included in “Longhorn,” help models built in for a wide variety of the capabilities in the “Longhorn” client system so that “Longhorn” clients are easier for you to manage from System Center, smarter event data.
There’s a lot of work that we’re doing to make “Longhorn” more cost effective to deploy and at our WinHEC conference next week we’ll be unveiling some of those innovations.
We’ve reduced the number of reboots that are required for update, up to 70 percent fewer reboots, and you’ll see us include Network Access Protection technology so that you can quarantine clients that are on and off your secure domain, so clients that are coming in from across the Internet, people who are mobile and are plugging in on both sides of your firewall, you want to be able to quarantine those people before you give them access because if you do that you’re going to improve the performance and operations of your environment, and we’re building technologies into “Longhorn” client to facilitate that kind of Network Access Protection and quarantining.
Network Access Protection Support
Today I want to announce that we will support in “Longhorn” Network Access Protection technologies that are consistent with the standard spec that the Trusted Computing Group has come out with. We are not actually members of Trusted Computing Group, some of the companies on the page here are: Dell, Hewlett-Packard, MacAfee, Juniper, Symantec, VeriSign, Sygate and others. But our Network Access Protection technologies built into “Longhorn” will support — will support the spec that comes from the Trusted Computing Group.
I think that’s very important. You’ll need to have consistency in how you quarantine systems inside your network consistent across the way the network infrastructure works, the clients work, a wide variety of the software works. So this is I think a very important aspect: Standardization here should lead to lower cost and improved operation in all of your environments.
From a security standpoint, moving to the next area, we’re making a variety of investments. I talked about Network Access Protection. We’re improving our identity services around Active Directory. Active Directory may be amongst the most important infrastructure that most of our customers are implementing today. When people first put in Active Directory it was mostly about implementing Exchange. Now Active Directory is a vital technology for management, for security, for a wide range of important scenarios and we’re driving all of our customers to be aggressive in utilizing Active Directory to the fullest.
We’re improving our updating services. In Windows Server SP 1 we’ve already had over 500,000 servers updates downloaded off of our Windows Update site.
And if you take a look at the acquisitions and investments we’re making, we bought Sybari to bring better security to the server and e-mail environment. We’re pushing hard in the antivirus area. We’ve got new technologies that we’re bringing to market for spam in Exchange. You’ll see new technologies for us in phishing. We recently have announced technologies for Windows in spyware. And so across the board we’re trying to help you implement a more secure, which means a higher quality and lower cost environment.
At the same time we’re working on our fundamentals, tools that lead to higher quality products with better engineering excellence and with great systems to respond when there are security problems, communicating with you, giving you the patches and fixes and help you need to keep your environment functional, operational and up and running.
Deployment: Wall of Fire
I’d like to turn now to deployments. So I talked about “Longhorn,” talked about security, I want to talk about deployment, but before we get to virtualization. Deployment is one of the areas of greatest cost and hassle from an operations perspective in the world. When users decide that they want to implement some new application, deployment can actually be the critical lead time factor for people rolling out new applications and new capabilities.
One of the major areas that Kirill and our management team have been investing in is in this area of deployment, not only for applications but deployment for the operating system itself.
And so I’ll have Bill Anderson come back on stage with me and we’ll show you something we call the Wall of Fire. And this wall here — come on up, Bill. (Applause.) Take a look here as those pretty little window screens to my left, to all of your right. This is Bill Anderson’s Wall of Fire, and I’m going to let you explain what he’s going to do and hopefully by the time I’m done finishing my speech what I’ll be able to show you he accomplished during the course of the rest of this talk. Bill.
BILL ANDERSON: Wish me luck, huh?
STEVE BALLMER: There you go, baby, Wall of Fire, it’s either that wall burns or you do. (Laughter.)
BILL ANDERSON: This is what we know in Microsoft as being five feet from termination, so everybody wish me luck. (Laughter.) I feel like I’m David Copperfield and I’m going to do an illusion.
Now, you saw the wall as you all walked in and I wanted to make sure that you understand what it’s going to do. Well, first of all, it’s kind of a cool marketing ploy, right, nice Windows flag showing there so you kind of see where you are, but the second thing is we’re going to go really prove what Steve was talking about, the ability to reduce the cost of ownership specifically in the area of deploying Windows-based PCs.
Now, in our case against that security message XP SP 2 is equally as important so what we’re going to do as part of the Wall of Fire — we thought about pyrotechnics and other things but we skipped that — we’re going to basically over the period of about ten minutes we’re going to use the SMS technologies and we’re going to rebuild this wall. It’s currently at Windows 2000 and XP Service Pack 1 and we’re going to move the whole wall to XP SP 2.
Now, for those of you who’ve been using —
STEVE BALLMER: A hundred machines while I’m speaking.
BILL ANDERSON: A hundred machines while you’re speaking, approximately nine minutes from the time I kick this off.
And I’ll walk you through how we actually do it. So those of you who have used SMS for a while you know SMS has always been a good preparatory tool and a good tool to deploy software like Office. You’ve got asset management, inventory, metering to be able to prepare. You could go roll out software in an upgrade fashion but we never went that last step, the ability to remotely re-image a device and still manage all my critical data and critical user information during the process.
Well, we announced the OS deployment feature pack last fall. To tell you how popular it’s been we’ve had over 50,000 downloads in the past six months of people downloading this free add-on for SMS 2003 to be able to do this exact scenario.
Now, one of the things I want to show you in the demo is just how closely integrated it is with SMS. I’m going to actually take this one over here, Steve.
The first thing that people ask for us, hey, look, I know SMS is a tool, make sure it’s integrated very tightly in the user interface and uses the same kind of constructs that I use today.
So if you take a look at the demo window I have this concept of an image packages node with programs, distribution points and access accounts, just like any other piece of software you would deploy within SMS.
Under program properties it’s going to look very similar. However, for an operating system deployment there are oftentimes a lot of customizations that you may have to do to be able to effectively roll out operating systems across your enterprise. We want to make sure that you no longer have to be script jockeys or have doctorates of “scriptology”, we wanted to automate as much of this in the user interface as possible. So something, for example, like licensing, we’ve automated it by being able to make it very easy for you to put your select ID or your PID in there so you can automate that process and not have to nest it buried in some script somewhere.
Now, for many of you, you may still have some very advanced needs because you’re large, complex enterprises. No problem, we’ve actually built the entire process in a very modular way. So if you find that the pieces we’ve automated just aren’t quite enough, you can actually take each one of the modules and you can go customize each of the pieces that happen. For example, what we’re going to be doing here and what I’ve done in this setup is I’m using the User State Migration Tool, which ships in Windows, as a tool to be able to accurately capture the bitmaps, capture the data and then be able to restore those post-provisioning.
Now, just to kind of set the expectation on this, most of the demos that you guys will normally see in a conference like this, there’s an array of people back there clicking madly behind, if something goes wrong we can flip UIs. I’m without a net on this one, guys, so I’m going to need you to wish me a little bit of luck because there’s no backup. What you’re going to see over on this wall is exactly what you would see as an enterprise administrator doing this particular rollout.
So now, as Steve would normally say, blah, blah, blah, blah, blah, shut up with the UI and let’s kick the wall off, so I’m going to do that.
STEVE BALLMER: Now, the only user state that we have on these machines is this bitmap.
BILL ANDERSON: Yeah, I don’t want to set expectations.
STEVE BALLMER: Just for perspective.
BILL ANDERSON: Yeah, we don’t have a gig of MP3 data or anything else that’s stored out there. We also don’t have an array of applications, we’ve tried to keep the stack short so we can fit this in the constraints of a demo. So what you would see in your normal environment is you would see this to be 20 to 30 to 40 minutes instead of the 9 or so that we would normally do this in.
Now, as an SMS administrator you would schedule this in off hours, you would say at 10:00 or midnight be able to kick this off so users weren’t interrupted. Well, we aren’t accurate enough to be able to choreograph and know that at 9:25 or something like that we needed this to automate, so I’ve actually scheduled this, it’s kicked off but it’s waiting for what we call custom action.
So what I’m going to do is ask for Steve’s help one more time. Please wish me luck and don’t fire me if this goes bad. OK. (Laughter.)
So I’m just going to go execute this script and we have activity. So what’s done is the custom action has been invoked, the OS deployment process has been kicked off.
Now, a couple of other things. You notice the machines that jumped out fast; those are the XP machines. XP performs faster, runs state backup quicker, does shorter reboot time and you’ll actually see when the window finishes up that those machines will finish early.
Now, the second thing is, as Steve alluded to, backup up the bitmap is an important piece for us. Why? It’s how you know we’ve succeeded. So as long as that wall pattern is intact in about nine or ten minutes or so you’ll know that we’ve successfully rebuilt, re-imaged those systems, restored the state and data. The only difference you’re going to see, because it’s XP SP 2, System Center is going to pop up right away to let you know that it actually has XP SP 2.
But the goal is really to give you the tools and the processes so that you can deploy hundreds or thousands of workstations in short period of times at low operational costs, so watch the wall, wish me luck and I’m going to get out of here. Thanks. Thanks, Steve. (Applause.)
STEVE BALLMER: See you, Bill.
It says here I have up to 25 minutes left. That better be looking like it did to start by the time I’m finished. The first person who notices it please call it to my attention because I’m going to move on and talk a little bit about virtualization.
Increased Virtualization Capabilities
Virtualization, I think, is an area of intense interest and activity throughout our industry these days, certainly at Microsoft. All operating systems have essentially been in the business to some degree in some way, shape and form a virtualization for all time. That’s how operating systems grew up, Windows grew up virtualizing the screen, the printers, et cetera, so virtualization in a sense is not a new concept.
But this notion of taking virtualization in the PC world, in the PC server world and really taking it to the next level, driving virtualization as a key technology to facilitate better compatibility, lower total cost of ownership and with the appropriate management tools really helping you operate a simple environment, that’s an area of intense focus, passion and interest certainly at Microsoft and I think a number of places in our industry these days.
Again in this area if I don’t get anything else across, I really want you to understand that this is an area where you should expect to see a large amount of very exciting and interesting innovation that should help you improve quality and cost of operation going forward.
We bought a company called Connectix a little over a year ago. That was part of us beefing up the arsenal of technologies that we had to apply to making virtualization a key part of our quality and cost proposition to you and we’re going to show you some of the product shipments and deliverables that are coming out of that acquisition during the course of my talk here today.
I want to start though with a little longer term perspective, the “Longhorn” timeframe, and I’m not saying “Longhorn” but about that timeframe there’s a number of things that we’re working on that are very important in virtualization. We’re going to take the virtualization format, the .vhd format, which we’re working today to get standardized across industry participants, we’re going to take that format and make it extensible, which we think is very important for all of you and for us in terms of where virtualization technologies can go in the future, open and extensible approach to virtualization we think is very important.
You’ll see us introduce hypervisor technologies around Windows. That is important. We have virtualization technologies today but really this notion of a smaller, thinner hypervisor and what that can mean I think is very important. We’re building on some of the work that we started on with NGSCB just in the security context and are broadening it out to really a very rich set of hypervisor technologies.
You’ll see us support key hardware technologies for virtualizations from Intel and from AMD and in System Center we will bring the capability to help you manage a set of virtual machines to make it easier to run virtual technologies as a way of improving datacenter consolidation, running multiple applications on a set of virtual machines, on a given server, et cetera.
So underlying core infrastructure, operating system infrastructure technology enhancements and management tool enhancements delivered in a very open and extensible way; that’s the target for our technology ambition in virtualization in the “Longhorn” timeframe.
Microsoft Virtual Server 2005
Today we have available Microsoft Virtual Server 2005. We’ll deliver SP 1 later in the year. It is in beta today. We have a Management Pack available for Virtual Server for MOM today. When we release Virtual Server SP 1 you’ll see a number of improvements. We’ve added support for non-Windows virtual machines being hosted on top of our Virtual Server product, including support for Linux. Remember what I said earlier about interoperability? We’re really believing that. We know folks are going to want to run Windows systems and Linux systems and other systems together on top of our Virtual Server and Windows. You’ll see support for that later in the year.
We’re dramatically improving our performance and you’ll see us support 64-bit hosts now that we have 64-bit support in the Windows Operating System. We’re licensing our VHD format broadly. You’ll see that in the fall, and there’s a large amount of support going into all of our Windows Server System products to support the Virtual Server in a very strong way.
Today I would tell you the following: If you are looking for a virtualization environment to improve cost and quality of any person in your environment or any set of people in your environment doing software development for tests, we have absolutely a blow-away product, and I encourage you to take a look at Virtual Server 2005.
For people looking to do data center consolidations for production applications, we have a very good product, but we also have a list as long as my arm of requests for enhancement, improvements, additional features, more performance, and we’re working, and we’re working, and we’re working, and we’re working and we’re working. This really is with the technologies I talked about before, one of the most significant areas of R&D investment for us, because we think that this is a core enabler to help us improve our enterprise management, and to help deliver to you the lowest TCO platform absolutely on the planet.
What I’m going to do now is ask Jeff Woolsey from our Virtual Machine Technology Group to come up on stage. Jeff is going to demonstrate some of the new capabilities in Virtual Server 2005 SP 1. Please welcome Jeff. (Applause.)
JEFF WOOLSEY: Hi, Steve.
STEVE BALLMER: Before you get started, I’ll now point out, it’s been seven minutes since he kicked off the demo. We’ll see how he does.
JEFF WOOLSEY: Excellent.
STEVE BALLMER: Not that I’m going to interrupt you when that thing finishes, but I may have a small celebratory dance.
JEFF WOOLSEY: No problem, we’re all looking forward to Bill keeping his job.
STEVE BALLMER: There you go.
JEFF WOOLSEY: As Steve mentioned, we have some major announcements around virtualization and virtualization management, and I want to demonstrate some of those deliverables that you can start using today. As you can see, I’m running the 64-bit version of Virtual Server SP 1 Beta, and I’m running it on Windows Server 2003 Enterprise X64 Edition. With the 64-bit host OS support, and additional performance improvements in Virtual Server SP 1, we have seen significant performance improvement. In fact, some of our early adopters have reported as much as a 50 percent decrease in host CPU usage running the same number of virtual machines and workload. This translates into higher virtual machine density, and lower hardware cost, making Virtual Server an even better platform for server virtualization and consolidation.
Let’s bring up the master status and show you all of the virtual machines we’re currently running. As you can see, I’m running a variety of operating systems, including Windows 2000 Server SP 4, 2003 SP 1 Enterprise, and as Steve has been mentioning with interoperability and heterogeneity, you may want to avert your eyes, Steve, this is Red Hat’s Enterprise Linux Advance Server 3.
STEVE BALLMER: As much as that hurts my eyes, I know that’s an important capability for the virtual server technology for our customers.
JEFF WOOLSEY: We recognize the interoperability needs of our customers, and want to ensure that Virtual Server running on Windows Server 2003 is a great solution for that environment.
Next, I would like to switch gears and tell you what we’re doing to improve manageability around our Virtual Server. Well, some vendors have created completely separate tools to manage their virtual servers and virtual machines. We’ve listened to our customers who have told us very loudly and clearly that they expect to be able to manage their virtual machines using the same management tools that they use to manage their physical ones. And that’s exactly what we’ve build.
So, I brought up the MOM 2005 console so I can demonstrate the Virtual Server Management Pack. The Virtual Server Management Pack allows administrators to monitor the health and performance of their virtual server and their associated virtual machines. So, let’s take a look at the health status of our virtual server. By clicking on my virtual state view, I can get a list of all the virtual servers on my network. To drill down and see all of the virtual machines associated with this particular hardware server, I click on the virtual machine role and I can see all of the virtual machines that are actually running on this physical box.
But, because we’re leveraging the power of MOM, and our Management Pack uses that intelligence very well, I want to show you an even more detailed look at our Virtual Server. I’m going to drill down and get detailed information about the role my virtual server. AS you can see here, all of my virtual machines are displayed in this role view. I can see the instance name, the guest operating system version, I can see the location of all of my virtual hard disks, the amount of memory being used, the amount of disk space being used, even the network the MAC addresses for all of the virtual machines displayed in a single detailed role view within the MOM Management Pack.
And as you’ve seen earlier, you’ve seen MOM managing a Sun box, you’ve seen MOM managing Solaris, now you’ve seen MOM managing virtual machines and virtual server. And by working with our partners such as Sun and Vintella. Vintella is working on a Management Pack to provide the same types of performance, the same type of management for Linux virtual machines, as well.
STEVE BALLMER: Today we still don’t have much information about the Red Hat system, but we will that’s part of our absolute plan.
JEFF WOOLSEY: You’re right. Absolutely. You see the Linux virtual machine is lacking some of that detail, and we’re working with them to improve that level of integration.
The last thing I want to show is actually some new features and functionality that’s enabled by the virtual server Management Pack. One of the things that our customers have asked for is the ability to perform tasks on multiple virtual machines simultaneously. And with the virtual server management tasks it’s simply not a problem. Select the task you’d like to perform, such as save state virtual machine, walk through the wizard, select all of the virtual machines I’d like to save state and click finish. It’s that easy.
Steve, the team is really excited about this opportunity to deliver on this deeper commitment to virtualization, and virtualization management. And I hope you’ll find that Virtual Server Service Pack One, and the MOM Management Pack go a long way towards providing improvement in performance, interoperability, and manageability necessary to make Virtual Server running on Windows Server 2003 a great platform for server virtualization and consolidation.
Thanks for this opportunity, Steve.
STEVE BALLMER: Thanks, Jeff. (Applause.)
A show of hands, how many folks here have Virtual Server running in their environment? OK. I encourage people to really take a look. I think it will be an area of very dynamic investment, and a lot of value in some very important scenarios. I think if you take a look over on the all of fire, I think we are actually all done. What you have just popped up so we can see it on all screens. What you see is the security center registration on top of the bit maps that we migrated before.
So if we just pan out there for a second and kind of let your mind see behind the security center dialogue that is up on the screen, you can see the pattern of the Windows flag, which means the new version, the XP SP 2 update has been installed, the bit map has been restored, the personal information, customization, user state has been implemented, and in fact, we have updated over 100 machines in something under about 10-1/2 minutes, exactly as Bill said.
Congratulations, Bill. (Applause.)
Commitment to Enterprise Management, Lowest TCO, Quality Operations
I want to wrap up by just hitting again on some of the key points, and then saying what I think they mean for you. Three key things, we’re committed to doing enterprise management very, very, very well, and all that implies, with the scale, with the quality, with the cost, with the kind of heterogeneity and interoperability that you require. We’re driving, whether it’s deployment or operation, or development, or security, we’re driving forward to make Windows and the rest of the Microsoft platform absolutely to stay what it is today, the lowest total cost of ownership platform available for the IT data center in the world. And we understand that the measurement on you will not just be cost, it will also be quality, reliability, uptime, ability to meet SLAs.
I was joking with the team backstage, I wish my mom had named me Steven Louis Anthony Ballmer, then I’d be SLA Ballmer, but I’m just SA. It’s bad, but I was talking to our IT guy about his SLA commitment at the same time. It is very, very critical, we understand. We have the people, the talent, the resources really applied at the things that we’re hearing from you every day, every day, that are on your mind. We’ve shown how virtualization is a key technology that will allow us to build out new scenarios to help you improve quality of service, and cost in your environment.
I think it’s fair to say four years ago, five years ago, 10 years ago, there was very little deep knowledge and understanding of the things that are on your mind, the concerns, the challenges, the problems you face everyday. We’re involved with many of you on desktop deployment 10 years ago, and with a set of tools that weren’t very rich and very viable.
As I said earlier, we’ve come along way, but we’ve probably come the farthest in really being able to walk in your shoes, to understand your issues, and to make sure that not only do we have the technology that you need to drive the goals that you have, but that we also have the support infrastructure, the consulting, the architecture, the post-sales support to really make sure that you can effectively bet on Microsoft platforms in the data center and on the desktop, in the enterprises that you represent.
We’ll be back for MMS 5, and MMS 6, and MMS 7, and MMS 8, and 9, and 10, and 11, and I think some place in the course of that next few years we’ll all step back and say this MMS, the MMS in 2005 was really a turning point where it’s quite clear and quite visible that Microsoft is on path to be the leading management vendor, and provider of the lowest total cost of ownership platform in the world.
I thank you for your time today with me. I thank you for your attendance at the conference. But, most of all, I want to say thank you for being incredible users, and advocates, and givers of feedback on some of our most important product technology. Enjoy the rest of the conference. (Applause.)