To install click the Add extension button. That's it.

The source code for the WIKI 2 extension is being checked by specialists of the Mozilla Foundation, Google, and Apple. You could also do it yourself at any point in time.

Kelly Slayton
Congratulations on this excellent venture… what a great idea!
Alexander Grigorievskiy
I use WIKI 2 every day and almost forgot how the original Wikipedia looks like.
Live Statistics
English Articles
Improved in 24 Hours
Added in 24 Hours
Show all languages
What we do. Every page goes through several hundred of perfecting techniques; in live mode. Quite the same Wikipedia. Just better.

Mendel Rosenblum

From Wikipedia, the free encyclopedia

Mendel Rosenblum
Mendel rosenblum.jpg
Rosenblum at VMworld Europe 2008
Alma materUniversity of Virginia
University of California, Berkeley
OccupationAcademic, businessman
Spouse(s)Diane Greene

Mendel Rosenblum (born 1962) is a professor of Computer Science at Stanford University and one of the co-founders of VMware.

YouTube Encyclopedic

  • 1/3
    14 010
  • ✪ Blitzscaling 15: Diane Greene on Scaling Products and Culture At VMware
  • ✪ Banyan Company Introduction
  • ✪ David J. Kuck receives 2011 Computer Pioneer Award


- I'm really excited here, to spend time with Diane Greene, who's a friend, but I also consider her as a former boss and mentor of mine in my career. I was lucky enough to intern for VMware in 2003, was in grad school, it was this small technology company of a couple hundred employees that no one really got, and then 2004, people were like, you're at business school, why are you going to this company? The smartest people I've ever met worked at VMware. The advice I always give folks that ask me for job or career advice, is find the smartest people, and go work with them because, you're going to learn a ton. Diane Greene was founder of a couple companies, but most famously, at least most recently, she was founder and CEO of VMware, which she co-founded with her husband Mendel, who's a professor here at Stanford, in 1998 with three other grad students, Ed, Ellen and Scott. Diane was talking about scale-up or blitzscaling, she was CEO of the company from those first five individuals in a small office in town and country, to over 7,000 employees, $2 billion plus in revenue, plus an IPO, it's an amazing journey from going from zero to two billion and to 7,000 employees. To top that, Diane is co-founder and CEO of another company, which we might talk a little bit about what she's learned the first time she scaled a company to this time. She's on the board of Google, she's on the board of Intuit, Kahn Academy, and a small technical college in Massachusetts called MIT. (laughter) Anyway, why don't we all welcome Diane Greene. Thank you. (applause) - So Jerry made VMware fun, so when we asked me, I was like, oh this will be really fun, you guys should have fun with Jerry. He loves to be teased. (laughter) - Okay, that's not on the syllabus tonight, but thank you. Well, speaking of fun and culture, why don't we start with that, around people. I know there's questions, and students, feel free to raise your hand and ask questions throughout the conversation, because we can steer the topics which way you want. Let's talk about culture when you scale up. VMware was doubling in head count, pretty much every year if not every other quarter, it just increasing a bunch of people. How do you keep hiring, and both hire the right people and keeping the culture strong from those early days? - Actually, the hardest, most difficult hiring was in the very beginning, because, we had this idea that was, we hadn't exactly refined our elevator pitch, and convincing people to come was a big deal. Each person, it's really interesting, 'cause right now I have a start up, and each person is a big deal, and then all of a sudden you're hiring a hundred people a month, and it's easier. (she laughs) So you can remember that it will get easier, you can never drop your standards, obviously, and that's why it gets easier, is if you create a strong culture and really good people, when people come in and it speaks to them, it's sort of fun, it's got the right look for them, we were competing with Google when were building VMware, and I just can't quite believe I'm on the board right now, because I did not like them. (laughter) - There's a story about people searched for VMware on Google, right? - Oh yeah, my husband, Mendell, who's a CS professor here, if you searched his name, it said, "Want a job at Google?" We were like, you can't do that. (laughter) We started out friends, and we used to go to each others parties, and then we realized we were competing for talent and we never talked to each other, and then we got really annoyed with each other, and actually when I, I'm digressing, but, when I joined the board of Google, Larry and Serge said, one of the reasons we hired Diane, is VMware was the only company we lost people to. When they lost people, it was generally because, and this says a lot about our culture, we found that people that had a big hobby, like they raced bicycles, or they like to go backpacking sometimes, or did different things besides just work, or if they were a little older and just starting a family, they would tend to want to join us, and if they really wanted a full 24 by seven experience at work, they went to Google, 'cause Google was running a 24 by seven service. That was, kind of, how it split out. - [Jerry] Interesting. - We got a lot done in our work days, but we didn't have nearly as long hours as Google. - One of the things you talk about is this autoimmune system when hiring or bringing people in, you say, they'd be great at hiring or great at firing or both. Talk about this concept of a cultural or autoimmune system at VMware. - Yeah, actually, my new company has it too, it just happens magically, and what it is is you get a critical mass of really good people, and when you bring someone in that just isn't, people aren't liking their check-ins, or if they're an engineer, or they're just not keeping up with the conversation, they kind of self-elect to leave. That is the most effective thing you can have to keep the quality of people. You have to watch and help someone leave if they're the wrong person, but I can think of a few examples. We hired a performance engineer, and two weeks later he left, he was like, I'm in the wrong place for me. I remember hiring a VP of Engineering at one point, and three weeks into his tenure we had an off-site, and I thought he was fine, he didn't say much, but right after the off-site he came in my office and sat down, he looked really shaken. I went, "What's wrong?" and he said, "I really couldn't keep up, I really didn't know what was going on." And I'm like "What are you, "oh, I saw there's so much to learn, don't worry, you'll get up to speed." and he goes, "No, no, I mean it, this is ..." And then I realized he wasn't going to make it, 'cause he just didn't have the confidence. I actually think he was smart enough, but he just, and so I said, "Okay." - [Jerry] That's interesting. When you scale that fast, from hundred to thousand to ten thousand, there's always this tension between promoting from within, versus hiring folks from the outside that either have seniority or domain experience. How do you balance that, and how do you think about that? - Whenever I went to hire the first of something, I really went for as much talent as we could possibly find, and I didn't, maybe it was my naivety, but I did not go for someone that had been a VP, senior person, I just went for the, so, for instance, Jason, who ran professional services, or Jeff who ran SE's, One of them, actually, who was a Stanford grad ran the track team. - [Jerry] Yup, Jason admired a cross country team here. - Yeah, but, he had been at two companies running professional services, I think he had been a first line manager, they weren't all that big, and I met him, and this guy was just so driven, and so sharp, and just the nicest person in the world. He came in as an individual contributor and he scaled that organization up to, it got up to about 1200 people. - Well over a thousand, not just the US, world wide. - And not only did he scale it up without a hitch, I had thought we would have to run professional services cash neutral. He ran it as a profit center. He was like, no way, I'm going to be a profit, not on my watch, and that guy was amazing. It was just wonderful to have him grow up with the company, because he understood the company, he knew how it worked, he knew everybody at the company, and he knew the nuts and bolts of it from being an individual contributor. The same thing, we hired this guy who had been a pilot. - [Jerry] Navy. - In the Navy, always hire a military guy if you can, they are the best, they are so disciplined, especially in anything approaching sales, I really love military people, and never had one of them go wrong. Anyhow, he came in as an individual contributor doing sales engineering, and he built that team out to quite large. - Yeah, he's still there. - And he's still there. - I remember when we could all fit in one lunch room, and every Wednesday you would stand up, in a room smaller than this, and would tell everyone what happened that week, what's happening next week. Obviously, that communication, that transparency was great, but then, you're a public company, the company's several thousand, how do you maintain that communication at that scale? - Right, It does get hard. One little thing I was going to mention, that I personally loved about, and I'm sure none of you have a problem with this, but I was very shy, and very self-conscious, and got nervous in front of people when I started VMware, but when you give a talk every week in front of a growing company, before you know it, you're talking to a few thousand people, but since it happened so gradually, you're not nervous anymore, and that's how I conquered my fear of being in front of people. In terms of, how do you keep the communication up, as we got bigger, we sent out, there's much better mechanisms today, we sent emails out to everybody. We did have big all-hands, but the other thing that was really good, that I thought worked very well at VMware was every Monday we had a staff meeting. I had a requirement that everybody on my staff, I said, you're welcome to do it Friday, but by 9:00 Sunday night, you've got to get me, and it's not a status report, what it is, is a write-up of anything going on in your group that everybody else should know about, or anything that you need to coordinate with your peers, take advantage of this meeting. They had to send it to me every Sunday night by nine, and then at 9:00 I would put it all together and then I would write mine and highlight what I thought was important, or anything I wanted to tell people about. Then I would give that out to everybody, and in that meeting I also always invited the founders, particularly Ed Bugnion always came, Scott came to a lot of them, and they were individual contributors in the engineering organization, and I always thought that contributed hugely to the culture, because everybody knew that these two engineers were in every staff meeting, so there were no secrets. They could just go ask these guys, hey, what's going on? They would just tell them, so it was a way to get the information out. The other thing was, that this staff report went out to my staff, and then they could selectively share it with their staff. The funniest thing happened, because I never asked anybody to, but everybody started making their staff, by Friday, get them the same write-up, and I hear people still do it over there. - I still remember Diane, basically, I was a young product manager, calling in or emailing about some customer interaction I had on a really beta product, so Diane read everything that was going on in the company. - Yeah, it's invaluable. - At that stage, it was a great way to understand- - Really dense, quickly. It's so efficient, there's nothing more efficient than that, yeah? - [Black sweatshirt] I have a question, are there ways you plan on doing it differently with the new company as it scales up, or roughly similar? - It's roughly similar, although ... - The question is, with the new company, are you planning to do this communication same or different with new tools? - It's really interesting, my new company has a lot of very different kinds of people. It has designers, which are, how many people here are designers? I've never worked with designers before, I have incredible respect for them, but, they're challenging for me. (laughter) I wish I could do what they can do, so I'm very at their mercy, and they do incredible things. So you have designers, who are not in the habit of doing something like that, and then you have all the other people who don't mind writing this up. I've been using it to kind of get everybody to know how to communicate with each other, we're not really using it to run the company, because we're small enough we don't need it, but it's been invaluable as a way to get everybody to kinda speak the same language, and care about the same things, because we're all so different. It's been an interesting purpose, it's been super valuable. The other thing that thing is so good for, is when you to bigger, is when you bring in a new, I brought in a new CFO and a new head of sales, and I said, here, go read all these, and you'll know exactly what's been going on. Nowadays, people go, here, go search Slack and read all the interesting channels, and you'll know what's going on. - I want to move from people to talk about the products, go from more of the product side. VMware started with selling Workstation via download, and then later this thing we call shrink wrap software, Frye's, where people actually buy CD's, and then added GSX Server, then ESX Server. One of the questions I always wonder, how do you know when to add new products, new product lines, because a lot of companies start out, they have a hit product, they ride it, and sometimes they don't come up with a second act or third act, second, third product, so how do you know when, and do you, how'd you guys create this culture of finding, experimenting new products, starting the desk top business, for example. - Well, we actually wanted, our vision was to build the server product, and we're like, nobody's going to put our, nobody's going to remove their operating system and put our "bare bones" hypervisor operating system on their hardware, and run all their mission critical workloads, this is not going to happen. What can we do, and how can we test this software? We wanted to do the bare-metal hypervisor, but we went, that's just not going to fly, we have to be much less disruptive. Then we're, sort of, Linux was just happening, and we had this idea, how about if we did it as a tool to run Windows applications with Linux, and we can get the heart of our system was this very difficult piece of code called a binary translator, for the x86 architecture, it's not so hard anymore because of the hardware assist, but, anyhow, we can really get the bugs out of that. So we said, let's build this tool, and we'll build it on the desktop, but then we had this terrible problem of all the devices, and we couldn't possibly write a device driver for all the, because, we were basically the OS. Mendel actually had the idea, oh, we'll just have Windows there, which has all the devices, and we'll pretend like Windows is running, and then we'll shove it out of the way, and use it's device drivers but we'll take over the system, and then we'll be a tool to run Windows with Linux and we'll see if people like this, and we'll get all the bugs out of our system. Linux was just happening, and everybody thought we were geniuses because Linus Torvalds got put on the cover of Time, and they're like, how did you know, and we didn't, we got really lucky. A bunch of companies tried to buy us as a Linux tool back then. - The question around that is, if the first product market fit was around Workstation, just like Windows on Linux and Linux on Windows, and, you're right, you could have gotten stuck, and people could have-- - There was another company that tried to go into the server first. They weren't doing the technology we were doing, but they were kind of a container company. There were two container companies that tried to go into the server business and neither of them made it. - When did you know to revisit the original vision to go to the server? - We were building both products from day one. What happened, was sort of funny, because the desktop business did pretty well, we got that thing to 100 million in revenue, we were pretty happy about that because we thought we might have to give it away, we weren't positive, because Linux users ... (laughter) We figured we made more money on Linux than any other software company. - [Jerry] It's probably true. - Even though Michael Dell, this is a very funny anecdote because he's trying to buy VMware, he was quoted in CNN, he invested in us, am I jumping around too much? - [Jerry] No, this is great. - Okay, he invested in us, and he was quoted, he said, "I'm investing in a lot of Linux companies, "because I want to equalize the playing field against "Microsoft, I'm going to lose money on all of them, "VMware, Linuxcare ..." (Diane laughs) You know how many billions of dollars he's paying for VMware right now? - [Jerry] Yeah. (Diane laughs) - It's funny. (laughter) 'Cause when you're the CEO of a company, you don't really like seeing Michael Dell quoted as saying he's going to lose money on his investment in you. - [Jerry] I don't think the audience realizes this, with Workstation $100 million, VMware was cash flow positive-- - We were cash flow positive from day one. - Day one, right, which is amazing. - But we ran the company at break even, we just ran it at break even. What happened was, we were going to bring out this hypervisor, this bare-metal way to do virtualization, does anybody want an explanation of virtualization, since we're talking about it so much? - [Voiceover] Sure. (laughter) - So, basically, it's a layer of software that sits in between the hardware and the operating system, and it kind of masquerades as hardware. What it does is multiplex the hardware so you can run multiple virtual machines at the same time, and you can run any operating system you want, as long as it's binary compatible with the chip set, in that virtual machine. That had benefits on the PC because you could run Linux and Windows together, or DOS and Windows 95, and then on the server, it solved all kinds of scalability problems. We mostly benefited from Microsoft's bad operating system that you could only run one application on it. Database vendors and whatnot would only let you, if you wanted support, you could only run their software on the operating system, so these servers were vastly underutilized. What we let you do is run one per virtual machine, and then all of a sudden you got the utilization up. What happened was, we wanted to bring out the server product, ESX Server. - [Jerry] Yeah. - You know what the name is? - [Jerry] Yes, I do. - We went to bring this thing out, and we launched it and nothing happened. Nobody was buying it. Meanwhile, we had also had this idea, we could take our desktop product and repurpose it as a server product, we barely did anything, we just said, it's a server product. - [Jerry] Just renamed it. - We called it GSX Server, and people liked that, and they were buying that, but we just couldn't get them to buy ESX Server. Then we invented the preferred hardware vendor program, and we told all the hardware vendors that everybody was signing up to the preferred hardware vendor program, and here was the deadline, and you had to have someone willing to support this, and you had someone willing to take it to market, and we had to get all these people, so we went to HP and Compaq, they weren't merged yet, and Dell, and IBM, they all got worried and signed up for it, for the preferred hardware vendor program. We relaunched GSX Server, nobody knew we had already launched it once. That worked out that time. - Let's talk about that. Early, for ESX Server, using these partners, IBM, HP and Dell, Compaq, were key channel partners. That helped scale, I'm remiss for not explaining virtualization in the beginning of class. VMware started in 98, for the first 11 years it was the fastest growing enterprise software company in the history of tech. Faster than Oracle, Microsoft, Veritas, in terms of revenue growth, it was based upon the technology that Diane, Mendel and the founders built, but part of that super-scale was a great product, to find the right channel, the right go to market. Maybe talk a little bit about the relationship with IBM, HP and Dell and the server folks, how did you use them, how do you stay neutral between all these parties? - It was great, we basically were pretty channel neutral. We had every channel, we had telesales, direct sales, we had resellers, we value added resellers, but our focus was really on the hardware vendors. I took this attitude that we were going to be Switzerland, which, you don't see in too many companies. It worked extraordinarily well for us, but for some reason, people, they get a deal and they're willing to give it exclusive or preferential rights when they do their partnerships. You see it all the time, but I said no, we need to run on all these hardware platforms. I had a algorithm, I said, customers first, what's good for the customers is good for the partners, and good for us. Partners second, what's good for the partners is also good for us, and we came last. That worked so well, so if you use that algorithm, you don't want to do preferential treatment to a hardware vendor, because, who knows what hardware your customers want to run on. I wanted to be Switzerland, so I told all the hardware vendors, we're going to treat you all exactly the same, here's the terms, if you take us into an account, we promise not to tell anybody else about the account, and we'll work with you, and we'll put dedicated people on your hardware or on your storage if you'll support that with us. If you develop something proprietary with us, we'll keep it proprietary for you, and if we develop it, we'll share it across all the hardware vendors. We just had this whole sort of dance worked out that was very clear to everybody, In fact, we actually wrote it down, eventually, as rules of engagement and sent it to all of them. I told our sales people they'd be fired if they ever violated it, I only had to fire one salesperson in the whole history of VMware for that. We did this neutral thing, and they would argue with us, but I think they liked it, because they knew they could trust us. - Most famously, even after EMC bought VMware, you doubled down on that neutrality, almost to the fact that all the partners, - Well it got me fired, eventually, but, (Diane laughs) - Customers first, right, so to your credit, you did the right thing for the customers in the early days saying, no, we're going to actually trade every other storage company equally, if not better, because it's right for the customers. - Right, and it worked. They were just a humongous channel for us, and we had that added complication that a lot of our customers were running Microsoft, I'll just finish this, and then, were running Microsoft operating systems, and Microsoft wouldn't let us support Microsoft running in our virtual machines, and enterprise customers aren't going to run software that's unsupported, so then the hardware vendors became extraordinarily important to us, because they all had support contracts with Microsoft. They didn't mind having some leverage over Microsoft, so we got them all to do the front-line support for Windows when it ran on their hardware, in our virtual machines, so we had IBM, HP, and Dell giving the support contract to our customers, if the customers wanted that, so we didn't get hurt at all by Microsoft's tactics, there. - It's amazing, right, do you think about what Oracle tried to do in terms of support with Microsoft in the early days, having friends early, like the OEMs was critical. - That was huge. - Chris had a question? - [Chris] You said something about, if they took you into an account, then you wouldn't tell anyone else about it. - So, for instance, if... - The question is, clarifying when Diane said, if a partner took VMware into a customer account, the statement, we would not tell the other partners, the other programs. - So if our sales guy went in with an IBM person to an IBM customer, and that sales guy happened to be best buddies with the HP sales guy, they couldn't tell the HP sales guy about that deal. That was very important. - [Voiceover] Why did they care? What? - [Jerry] HP won't try to come in and strike the deal. - If they know someone's in the market to buy some virtualization, we had this wonderful thing, if you bought virtualization you ended up buying a ton of hardware. So, oh, that customer's going to buy a bunch of VMware, I want to sell the hardware. - In terms of thinking about this strategy using partners for the folks learning, for every dollar of VMware sales, generated $8-$10 of storage, servers, and add-on services around it. They could give VMware away for free. - I think with that HDA stuff it got even higher. - It got even higher, to like 12 or 14, they could give VMware for free and make all the margin back on the storage and the servers. It was a hugely profitable partnership for these companies. Let's talk about the sales, then, organization, scaling that out from single individuals like Jeff doing the first SE, and then adding more sales reps, and more territories. One question startups always have when hit the scale up, is, do they hire sales too early or too slow? Typically, they're an expense, if they don't make their number it doesn't look great, so how do you think about when to, pour the gas, if you will, on that aspect of a business? - This was pretty new territory for us, and we're like, a sales person, we've got to go find a sales person, we didn't even really know what they looked like. I remember one of our first was this amazing woman who was a stand up comic, she made a really good sales person. I can't remember her name, but anyhow, we went out and found two or three different sales people. They each did different kinds, one worked with channels and one got on the phone, and one was, I was learning all of this terminology, one was, what they call, a direct sales person, they went out and talked to the customers. We tried one of each, and we said, okay, what can you do? We really paid a lot of attention, and we had to figure out whether they were bad sales people, or whether we didn't have the right market, because we weren't really sure which it was, and that's the hard thing in a start up. With trying every approach, we started seeing some traction in places, so we had a sense we had a product fit in the market, so then we started, well I started, getting introductions to every VP of Sales I could to hear about what they thought. They all had a completely different way we should set up our sales, but it was really interesting to hear it. I went out and studied sales. That was how we developed, we realized our CFO at the time, Tom Durowitz, just a wonderful, wonderful person, he and I and this other guy, Jeff Burn, we came up with this model from studying all these different sales organizations all over the place. They actually teach this model at the Stanford Business School that we came up with. I think Kirk Bowman teaches it. We had this really technical product, it was really hard to deploy, and so we said, we need someone on the phone, kind of drumming up business, and kind of explaining virtualization, we need a sales engineer to go in there and find out how it's going to run on their hardware, and if they have compatabilities with their storage network, and so on, and then we need a direct sales person to close the deal, so we're going to make these three person sales team, and we're going to give them all one number. Now, the direct sales people always make more money than anybody, and the SE's make the least amount of additional comp, so they didn't all get the same amount, but whatever they got was tied to the same number and how that team did, and that worked like a charm. The other thing we did that worked like a charm that was a little different from what a lot of people did, was, we said, we want to make you as productive as possible, so if you can go sell with our partners, like IBM, Dell and HP, we'll give you the same commission for that sale, even though we don't make as much money 'cause we have to give some money to the hardware vendor. We're going to make you compensation neutral, regardless of what channel you sell through, and we gave them a geographic territory. And boy, the smart sales guys, they just went all out, educating the channel, helping the channel, and they got this really leveraged market where they were making millions of dollars, and I loved paying them the millions of dollars, because it meant VMware was getting even more millions of dollars. It was really effective. - The fascinating thing, there are two lessons, one, having this triumph of the inside sales rep, the SE, and the direct sales, all on the same team, comps at the same number, number one. Number two, a lot of folks don't realize, when you had this indirect channel, like a reseller or in HP, even HP sells the deal, you still get paid, even- - [Diane] As if you direct sale. - [Jerry] So your incentive was to get as many partners out there on your side, like you said, self-selecting, the smart sales executives realized, if I could just train a bunch of my channel partners to basically do my work for me, even better. VMwares able to kind of ride that leverage, if you will. - Yeah, it was great, and then we had no politics, because everybody was in it together, and everybody was getting rewarded. - During this growth, one thing, going back to the product side. - My advice is, I think it pays to be generous in sales. - Let's talk about that, on the back of the people side, let's talk about compensation-- - But you can't force sales, if you're not getting traction, hiring more sales people doesn't help. You've got to get the model right, and you've got to get the market fit right. - [Jerry] How do you think about compensation, with the new company and the old company, cash equity, salaries, it'll be kind of curious. - It's different today. (Diane laughs) You can make more money at Google and Facebook and all these unicorns, well, for the time being the unicorns. (laughter) Some of them are real, very real. You can make more money there than you can if you come late stage to a start up. It didn't used to be the case. Now the startups have to take huge amounts of money to pay, if you pay startup wages, these people are going to come out behind where they would have come out had they gone to one of the bigger companies, because the salaries are so high. - So you're finding, with the new company, you have to increase the salaries considerably. - I haven't had to, I have this incredible team, but I actually did a little study on it, and I realized this isn't fair, so I took a bunch of money and I started paying people a bunch of money, because it wasn't right. - That's great. I want to move back to the product side, one of the things, when we talked about how you keep scaling up, you found this product market fader inside the proverbial tornado, how do you enable your R&D teams to keep coming disruptive types of technology, like VMotion or live migration, or HA or Fault Tolerance. The core virtualization was such a magical piece of technology, you think, like, hey, we're done, but the VMware product teams didn't stop, the engineering team created live migration, they created HA, Fault Tolerance, Resource Management. - [Diane] Virtual Center. - Even at 7,000 people, along that scale, how do you encourage and foster that culture of startups within VMware at the time? - We just had exceptional, we really had exceptional people that pushed each other. One thing, culturally, you want, is, you want to have people that are self-driven. Even the person sitting at the reception desk, you want that to be a driven person, every single person in the company, no matter what their role, to be someone that likes to set the bar high for themselves, and so I think we had that at VMware, and then we expected it of everybody. I have had people say, they really enjoyed VMware, because they did more than they thought they could do. I think these high expectations are good, they say to do that with your kids, too, I don't know. I don't know if it was real or not, but I always said to our VPs, give me your schedule, but, put a little slack in it so people have some extra time to kinda play around. Don't schedule them to the max. Always leave a little slack in the schedule and I always coached them to do that. I don't know if we knew what we were doing in scheduling enough to really do that, but that was the philosophy, and then we did things, like, we had these poster sessions. We would have these little conferences within VMware, where people could write a little paper, or a little proposal about a new technology and present it, and then the top people would win something, and also get to go build it if we chose it. That was really helpful. Also, there was this, sort of, fun thing, like, I remember when the Mac moved to x86, and everybody's like, we've got to build VMware for the Mac, and so I went over, I was trying to talk to Steve Jobs about partnering, and he wouldn't have anything to do with it, he just thought, why don't you guys get out of my life. - [Jerry] Windows on my machine, no, never, right? - Somebody told me he liked me, but I don't know. Apple wouldn't give us the time of day. So, I told everybody, we cannot build this, because they don't want us to, it's illegal, but I'd walk around engineering and I'd see all these Macs, and it was kind of obvious what was going on, and I'd just go, what's that Mac there for? But I wouldn't say anything. So I made it really clear, that if you were showing initiative and doing good things, we like that, and you weren't going to get in trouble for that, ever. I didn't even have to do it, it was in the culture. It was just great people, we were lucky. - [Jerry] We were lucky, well, I think two things, one, it was always the carrot and the stick, if you were innovative, I think there was a culture of rewarding that, but there's also with some companies as they scale up, there was a fear of failure. If everything's going right, you fear to try new things, and I don't think we really had that at VMware in those early days, because we were trying a bunch of things, and people weren't worried. If it worked out, great, if it didn't, we had the slack in the schedule, so to speak. - When things went wrong, we were always really careful to talk about how to prevent it from happening again, but never to say this person screwed up. We never highlighted someone screwing up. We did highlight our mistakes, but we diversed it from who it was, we all took responsibility, somehow. - [Jerry] It always goes back to the people, maybe I'll segue back to that culture in hiring. I'd love to talk about diversity as hiring and growing both gender, age, and on, I know it's a topic you and I have talked about a lot. How'd you think about that when you were scaling up VMware, and with the new company, how do you think about it now? - [Diane] We had this attitude, let's hire everybody we can out of Stanford. (laughter) And MIT and Brown and Princeton, and a few other schools, we liked bringing in interns, we actually pioneered that intern program, actually. VJ, our office manager, back in 1999, we had an intern program, where we went out to the colleges and recruited people to come before they were graduating to come spend the summer with us if they were a junior. We made a big deal about it for them, and then if we liked them, we made them an offer. We got a lot of people that way, it's a little more competitive now. We tried as hard as we could to get all kinds of people, it was hard to find women. We had a fair number of women, but there wasn't a huge pipeline. I understand now it's the biggest major for women, in CS here. - [Jerry] That's great. - Yeah? - How are you doing the super early stage hiring with the new figures at 100% personal network this time, or are you using other? - So the question is, with Diane's new company, how is she doing hiring, the first set of hires. - [Diane] How am I recruiting? The beginning was personal network. - How big is, continually- - We're only 36 people. There's five former VMware principal engineers, (she laughs) and a former Senior VP. But, we also have, for instance, design, it took me a while, just like at VMware, I had to keep upgrading the VP of Sales, because I really didn't know what a great VP of Sales looked like, and it took a while, Kirk was a very good VP of Sales and so was Carl, but it took a while to get to them. One, I couldn't attract a great one, because we hadn't established ourselves, so as soon as I could upgrade I would, and it was a continual process. With my current company, I had the same problem in design, I finally had to read about ten books on design, and go hang out in a bunch of mission coffee shops. (laughter) And then I mined my network and I asked everybody for any great designer that they knew, and I finally found a fabulous head of design, someone that was running design at Walmart labs. That was just going out, very specifically, to a bunch of people saying, I need someone to run my design. I did bring in a recruiter, we have an in-house recruiter. But, she doesn't really tend to, I mean, I recently, this is my favorite hire, someone sent me a designer resume, and I was looking at it on LinkedIn, and on the side popped up a picture that this woman, I was like wow, she has so much energy and drive, I really like her. I looked at her, and she was a designer. So I sent her to my recruiter, I said, see if you can get her in, and it was purely based on the picture. (laughter) - [Jerry] Too bad Reeve's not here tonight. - I know, it was so amazing, so, we hired her. She's an amazing designer, and she told me that her husband is in advertising and he took the picture, so he's very proud of that picture. (laughter) Isn't that a great story? - [Jerry] That's an amazing story. I'd love to just, maybe spend more few seconds on a design, and this user-experience layer, VMware built this great platform, and now you and Mendel are spending a lot of time understanding HCI, human computer interface. - Well, also this layer, it's a mess. The layer above like this HTML, CSS, iOS, Android, and how you talk to the back end and the database, and update things, and separate the database from the front end, it's an area of the system that everybody thought, oh, that's ... real software computer scientists do systems, they don't do that upper layer, so not a lot of people have worked very hard on it, and then, HTML, of course, was never designed for what it's being used for, so that's a little bit difficult. It's incredibly difficult to test this stuff, and then you have to be able to rapidly iterate, because you can't get your design right, because humans are so hard to predict. Some people thing machine learning's going to do that design someday. So, anyhow, I think it's We're designing complex interfaces, where you have lots of roles and lots of workflows, and communications, and collaboration, and access controls, and to make that simple and easy to use, and rapidly iterate on that, and test it all, we've had to build a lot of stuff to support that. - It's interesting, we talk about founders and the first ex-hires early are people you know, you trust, and then you find folks that are purpose with that skill, like design, or user experience, and then bring them into the company. - There's the product person, and the user researcher, and the designer, and then, there's some designers that are kind of product people, you guys all know this, some designers that can write ... - [Jerry] Code. - Can build the UI, and then there's front-end people that are sort of full stack people, and there's front-end people that only do front-end, there's front-end people that also do design, and you have all these different kinds of people. Then there's this whole process to bring it all together all this agile development is designed to try and help that, but it's not trivial, my husband's teaching a course in this, we've gotten so interested in how to do this better. He's teaching the course in the winter. - I want to compare notes with Mendel, we're looking at a bunch of stuff in that space, too. Question. - [Voiceover] This is not on the design topic, but that's super interesting. We look at a lot of companies Graylog and for people thinking about starting businesses here, there are many companies that deliver value to businesses, but only one or two where, to the CEO of the company, this is their strategic vendor, you can go to lots of businesses today, (drowned out) They make all their other decisions about technology based on being a VMware shop. A lot of that is the product, but did you guys ever aim to be the core decision-making power in an enterprise technology organization? - Let me rephrase the question. It's a great question. VMware's a platform, maybe, and a lot of large companies, now, make a decision if you work on this platform or not. In the early days, did VMer's aim to be this platform for IT that became such a standard, or not, and when do you just make the decision, or how do you fall into that? - Well, It was not all premeditated. I won't take credit for that, but we did believe right from the beginning, that this should run on every piece of hardware. We were so sure this was a better way to run systems, we said, look, this should really run everywhere. We did think of it as a platform from day one, that's part of why we were neutral with the hardware vendors, and just a lot of decisions we made about what to support and not support, were based on the fact that we thought it needed to be a platform to maximize it's value to people. The beauty of VMware was, it was useful on a single server, so someone could try it out, some Sys Admin would use it, just, stand alone, and it would spread virally through a company in little tiny pockets, and we could see where it was going, so we actually didn't want to go to the decision-maker right away, the CEO type person, because they always have this person called the purchasing agent, and their job, their only job, is to get as big a discount as possible. They're compensated on how much of a discount they get, so you kind of want to stay away from that guy, particularly at a place like General Electric, big huge company. We would let just kind of it spread, and our sales people would map out the organization and if they saw a hole where they weren't using it, they'd go in and try to get it there, and then we'd wait until it was all over the organization, and then we'd go in and say, hey, looks what's going on, would you like an enterprise license agreement, and we would monetize it that way, but we would wait. We didn't consciously say, we're going to be strategic to the CIO, what we consciously said was, this is going to run on every single piece of hardware. We didn't put two and two together to say, well, that's what's going to happen, but happily it did. - One of the results of- - There's a question. - In the back. - That's funny, I thought you were asleep back there, and then you woke up. (laughter) (student talking) - The question is, Diane's thoughts on virtualization now, containers and Docker. - He's the world expert on Docker, so I'll let him say a little bit, but I'll just say that we couldn't do a container at that point in time because we had to be completely non-disruptive. We had to make it like... nobody believed in virtualization, nobody believed in this kind of separation, so basically we actually built a tool, a P to V tool, nobody was going to build for a virtual machine, so we built a little tool that grabbed a physical machine and plopped it in a virtual machine by pushing a button. We had another tool we built that went out on your network and saw the utilization of all your servers, and we'd say, these are really good candidates for virtualization. People do things incrementally, they don't see the value, so VMware showed the value of putting things in containers, big time, but it's a fairly heavy-weight mechanism that, now that everybody appreciates the value and wants to do it, you can do these lighter-weight containers, and I'll let Jerry talk about what he, he's going to tell you Docker's going to own the world. There's also this Kubern - [Jerry] Kubernetes - Kubernetes. - It said full disclosure is, I left VMware in 2013, joined Graylock. My first investment was in Docker, so, I'm too recent for that. - It was a great investment, yeah. - Thank you. With tiers one I watched first-hand what Diane, Mendel and the rest of the company built at VMware for a decade. I saw the value at Docker and the containers were bringing as the next wave, the next evolution, so I think they're, I'm really excited, we talk separately about Docker, because I know the classes focus on learning from Diane, but I think both Diane, now the board of Google, that's very involved and it kind of gives the container ecosystem, we're both very bullish on Docker containers in that ecosystem. - She always wanted to do something in the lightest way possible way, but there wasn't a market for it before VMware. - Docker would not exist if it wasn't for VMware. - Yeah. - Question? (student talking) - Well, - The question is VMware stayed under the radar for many, many years, both in marketing and awareness and the culture was overdeliver, right, just be-- - Underpromise - Underpromise, overdeliver and the question was now you see a lot of startups out there that are called over the radar, that they overpromise and underdeliver. But, definitely there's a marketing around some companies out there, and so, - Yeah, so I think there's a lot of advantages to underpromise, overdeliver. The advantages are that you set your pace and you do things when they're ready how you want 'em and you can also build trust, because, you don't bother to deliver 'till it's, I mean, you don't try and deliver until it's ready. So, people come to trust your brand and trust the quality of your product. We were building a product that everybody was gonna run all their software on, so they had to trust it. We needed to be a trustworthy, that was an important part of our brand, was to be a trustworthy company. And it was just a lot less stressful and just a nicer way to operate, because if our engineers said oh, wait a minute, we gotta change something here, we could take the time to do it and we hadn't promised something to the market forcing us to take shortcuts. The advantages of making a lot of noise. Well, certainly if you have competition, it helps to make a lot of noise and sometimes you can, I mean, I won't name companies, but there are some companies that have done it that way and been super successful. Somebody once said to me "You know what, Diane, "if you took you and Mark Benny off and put the two of you together, you would really have something." (Diane laughs) - I'll invest in that. - He's extraordinary genius at marketing. - Yes, let's talk about marketing of VMware. In the early days, to the question, we weren't in the market very much at all and then all of a sudden, in 2004, we created VMworld, this conference. 1400 people the first year, then 3000 the next year. Now it's over 20,000 people and I remember setting up chairs in the first year. I set up chairs in the last year too. - I remember you the night before making my slides. - That's a separate comment, but, yes. I remember the night before making your slides at the first VMworld. But, you made that an industry event, right? Not a VMer event and at a company that's less than 1000 people at the time, that's pretty ambitious. - Well, again, I really believed this stuff belonged on every machine, I really believed that, and I also really believed for it to be valuable to people, it had to be very open. You had to let anybody participate and it had to be about virtualization, not about VMware, and so, it was totally open to anybody that wanted to come and it was just run with our partners in a very open way. So, I wanted it to focus on bringing virtualization to the world and we had the virtualization solution. Nobody was gonna compete with our virtualization solution, so it was a pretty safe thing to do from a company building stand-point and it just felt really, it was really fun to do it that way. - Yeah, well, I remember those slides. The message from day one was, we were creating this new industry and the rising tide would float everyone, partners in VMware included and it was amazing what happened. So, the 20,000 people later, to the platform question that he had vendors building entire startup ecosystem appeared around VMware, large companies, small companies. How do you think about-- - Yeah, I mean, over and over again, we did the generous thing and it always paid off, whether it was with our own people, with the partners, it just seemed to pay off. - Well, this ecosystem spread around VMware. One question I want to ask you is, some companies now when they're in the hyper growth phase, try to buy other companies big and small to get mass and VMware never did a big acquisition. Small tech ones here and there, but they never had to in those days and I'm curious to your philosophy around M&A for startups or just scale-ups as well as they're growing. When does that make sense when that doesn't make sense? - Well, I think VMware was a little bit unique in that we were creating this new industry, so there wasn't really anything to buy. As we got further along in there started being companies to buy and I don't think we did a very good job at those acquisitions, in all honesty, 'cause we didn't have experience, we didn't know how to do it. Although we knew how to manage our own, I thought we knocked that one out of the park, but, anyhow, yeah, so there's a company out there, Domo, and that started with an acquisition and I think he did several acquisitions and I think he's doing really well. So, number one, if you're a startup and you've never done an acquisition, I think it's a really bad idea, because they're super hard to make, I mean, very experienced companies blow it all of the time. - Most of the time, right? - Most, probably 95% of the time, they don't work out so well and so, you really have to know what you're doing, and then, you have to know what will work, what kind of a fit. The guy running Domo, I forget his name- - [Jerry] Josh James. - He's just very talented at it. - Josh, yeah, Josh is. And just speaking of growing, you said early you should avoid M&A if you don't know what you're doing, because often times they don't work out and if a big company can't be successful as a startup or scale-up, you should not think you're either. - It's a big distraction, 'cause they're very time-consuming and the smaller you are, the bigger a distration-- - Culture, people - Yeah, yeah. - How do you think about organizations and reconciling the two? One of the comments, I was just thinking about marketing. How we scaled in sales in conjunction, you said something about upgrading VPs of sales or your execs. When did you know it was time to upgrade? Was it because they missed a metric or performance or you felt like you weren't managing folks well, what kind of sign-- - There was always something different. - Yeah. And that's just sales, it could be engineering, could be anyone in your exec team. How do you think about upgrading people? - Well, I can give some specific examples. So, our first sales guy, who ended up staying at the company, he came to me when we were doing 10 million in revenue and we were planning and he said, "Diane, there's just not that much here. You're gonna be lucky to get this to 20 million," and that was when I knew it was time to upgrade my sales guy. I'm like, no, I know we can do much, much better than this. So, I kept him as the sales guy, but I went and got a new VP of sales and the next one, I caught lying. I had a VP of marketing that I walked out the door, because Intel sent me a mail he had sent 'em and he had sent them all our confidential stuff trying to get them to buy the company and so, I came to his desk and said, "Come on, "we're going to your car," and that was that. And then, so really, anything can happen in a company. It's always amazing, the kinds of things that go on, but, if someone is not being treated like a leader, if you have someone that nobody's listening to, that's a pretty sure sign that you don't the right leader. I always had a really open door. Our first VP of engineering, people were just getting more and more frustrated, saying he won't make a decision, he won't make a decision, and the old maxim is so true. When you fire someone, you never go ah, I wish I waited longer. You always go why didn't I do that sooner? - Bad news never gets better with time. - The thing I always tell people is, when you fire someone, it doesn't have to, well, unless they sent all your IP over to somebody at another company and then it's personal, (Laughter) but, otherwise, it's not personal. I always talk about the fact that you're gonna be happier somewhere else. I hired you for a reason, I thought you were talented. I still believe you're talented, you're just not a good fit here, let me help you find another role and I've remained friends with a lot of people I fired. - I'm just wondering if you could talk about some of the other things you've done differently, intentionally with the new company, do you feel like you are the CEO of just the company in general? - The question is things that Diane has constantly done differently with her new company. - Well, one thing, it's such a different climate now I did this radical thing which I never would have, I would have told someone they were an idiot to do this. It's so hard to hire and the traffic between here and San Francisco is really intense and my initial group had some people that lived up in San Francisco and out in San Ramon and some people that lived down in San Jose and Cupertino and whatnot. So, we made a decision to get two offices from day one and then we said two days a week were, one office is in Los Altos and one's in downtown San Francisco and we said two days a week we're in Los Altos and two days a week we're in San Francisco and when you're small, you can get an office big enough to hold everybody, and one day a week everybody can be wherever they want to be. - Two days in Los Altos, two days in SF-- - Recently, we've been letting people do one day, 'cause we know each other so well, but, yeah, so that was a pretty wild thing to do, but it took all the friction outta hiring. It was really valuable. - [Voiceover] So my question is actually more about, now that you are on the board of Google, what is sort of the one thing you see how in Google's culture is so great, I wish I had that at VMware and vice-versa, you feel like this is something we need more, so then, I wish, Google is doing that. - The question is, Diane's a board member at Google, what's a cultural thing that Google does really well that she wished she had a VMware and vice-versa, what's a cultural thing VMware had that she wished Google has more of. - Well, I think Google was bolder. They knew not to sell and I wish VMware had had that, although I didn't want to sell, but it was just, Google has, and it's gotten even better. I mean, there's just a real boldness. Incredible quality, incredible people, but boldness about Google and really culturally, we're gonna do the right thing, it's real. We want to help the world and I really like that. We felt that way at VMware, but it wasn't quite as strong culturally as it is at Google. Something I think we did better at VMware, although we weren't at the same scale, so I think it's a little harder at 65,000 people, so, I'm not sure I could have done it at 65,000 if we'd gotten VMware that far. But, I feel like, actually I'm not gonna say. Sorry, I can say in private, yeah? - [Voiceover] What were the best and worst pieces of advice you've ever received? - The best piece of advice I ever recieved was you can never overcommunicate. - [Voiceover] And the worst? (Laughter) - I was at this startup that I had a major disagreement, a little startup that did streaming video, and I had a major disagreement with the other founders. I wanted to build software, they wanted to build hardware and so, I left and one of the board members said to me, he goes, "You shouldn't leave. This is the biggest deal "you'll ever be part of." (Laughter) I always remembered that. - [Voiceover] You talked about taking steps and sequentializing things and obviously learning more, for a lot of us, we undergraduates or students or just in the earliest days of starting the company. Talk about your intention in those earliest days starting VMware, like, back when you were an undergraduate or right before you founded the company. - Well, the sad truth is, I was pretty old when I founded VMware, I wasn't an undergraduate. I'm so flattered that you think I'm that young. - The question is, steps that Diane was thinking about before she founded VMware. - Okay, so, VMware, but I will tell you, VMware was actually my third startup and so, actually my husband Mendel had invented this technology and he was working on it with his grad student, or invented this idea and they were building the technology with his grad students and the first thing that happened was they put out a paper and all these people at Microsoft, like Bill G. and Nathan M. and so on, he got a letter from somebody. It was under blind review, somebody at University of Washington wanted to know if all these people could read it, they'd be very interested to read it. And, I go, "Wow, you should get a patent file, "before you do anything else." It was just in the graduate school, in his research group. Anyhow, as it went along, I said I really think we should take this to market, this is really neat, and I start talking with my husband Mendel about all the different ways it could be used, and I'm like, this should really get taken to market. So, we had the grad students over and, I won't go into the details. We had a discussion with them about whether they wanted to do it. They had some other ideas, but in the end they decided to do it and about that time, I found out I was pregnant with our second kid. So, I said okay, I've been involved in two startups. I know how to do it, 'cause both had had good outcomes, so I'll help get this off the ground, but I've got a four year old and I'm about to have another baby. This isn't gonna work for me, but let me get the company going and then we'll bring in a CEO. And we had this vision of what we were gonna do and that it should be, virtualization should be on all hardware and that we would come in with, you know, the desktop, the whole thing I've described here. And because one of my previous startups had been a little dysfunctional, one of the first things I did when we started, I said, okay, we have to write our vision and our mission and everybody needs to sign it and we're not doing anything else 'till we do that, 'cause I had learned a hard lesson and we did that. And we started building the company and then I had my baby and we were in this office, this isn't really answering your question, but I'll just finish this little story, but we were in this office where you could open the windows, so I just brought my baby in, 'cause she could get fresh air and not much was going on. We didn't have any customers. I was just negotiating a license with Phoenix for a BIOS that we needed. That was all I was doing really and hiring people and, did I answer your question? Okay. You had a question there, what's hard about scaling? One of the hardest things about scaling VMware ironically, the thing I remember as the hardest was space, having enough space for everybody and anticipating it, 'cause when people get overcrowded, they don't work very well and it was during the bubble and rents went up to 10$ a square foot per month. And that's what put Donna Dubinsky out of business. She signed a long-term lease at that rate and that's why she had to sell Palm and it was agony, 'cause we were, it was just constant agony having enough space. - And also that balance in committing to a long-term lease at the time. - Oh, yeah, you didn't know what to do and we didnt' do it. And luckily the bubble burst and we were so lucky. - Even then, the hyper growth, and companies are facing this now, and how to plan. Your space is also physical space as culture, right, in terms of two offices for a new startup, like, where people are and you want a space that people can collaborate, work together, or choose spaces. And I remember, we went through office defrag a lot, where we kept reshuffling. We had multiple offices and we kept redoing it every six months and I remember one office move, I opened a box and it was not my stuff, my office mate's from two moves ago (Diane laughs) and he had left the company and I'm like oh, I have all of Jeff's stuff. So, maybe just, we're running out of time, just let me, just last thoughts. In terms of scaling up besides space, if you give a piece of advice to students or entrepreneurs about, the one or two things to be thoughtful about, that would be helpful. - Shoot, I forgot to think about that. You know, everything I can think of is so darn trite, about concentrating on doing something you're really interested in and could be excited about, no matter what you do. One thing I have thought about lately, which is that you can be an entrepreneur anywhere. You don't have to start a company to be an entrepreneur and I think it's a little problematic. I do talk at a lot of schools when I travel and I'll ask everybody how many are gonna start a company and everybody'll raise their hand and guess what problem that creates? There's nobody to hire, because everybody's starting a company. So, I think it's wonderful to build something, but you can build incredible, I mean, I was an entrepreneur before I started companies. The only reason I left the big companies to start my own companies, was 'cause they didn't have great companies back then and the companies were so horrible they would just not do new, they weren't that horrible. Tandem was a good company, but they wouldn't do new things, they were so afraid. They wouldn't jump off that elephant, they'd ride it right over the cliff. So, I guess maybe, part of my advice and I don't know if you think this way or not, which is, if you have an idea and you want to build something, figure out the best environment to do that in. And it might be starting a company and it might be doing it as part of something bigger and what's really important is who you're with and what you're building. I think some of my most satisfying things I did were not, I mean, VMware was incredibly satisfying and fun and I wouldn't trade it for the world, but I did a lot of other things I really enjoyed that were at bigger companies, yeah. - Thank you very much. (Applause)

Early life

Mendel Rosenblum was born in 1962. He graduated from the University of Virginia, where he received a bachelor of arts degree in Math. While at UVA, he was a member of Phi Sigma Kappa. He went on to earn a Ph.D. in computer science from the University of California, Berkeley.


Rosenblum is a professor of computer science at Stanford University.[1] His research group developed SimOS.[2]

Rosenblum is a co-founder of VMware.[3] He served as its chief scientist until his resignation on September 10, 2008, shortly after his wife Diane Greene was terminated as the company's CEO.[3]

Since 2008, Rosenblum is a Fellow of the Association for Computing Machinery[4] "for contributions to reinventing virtual machines",[5] and had previously received the ACM SIGOPS Mark Weiser Award (2002).[6]


  1. ^ "Stanford School of Engineering - Personnel Profile". 1969-12-31. Archived from the original on 2005-03-17. Retrieved 2011-12-16.
  2. ^ "VMware Leadership". Retrieved 2011-12-16.
  3. ^ a b "VMware loses Mendel Rosenblum, co-founder and husband of fired CEO Diane Greene". 2008-09-10. Archived from the original on 2011-06-13. Retrieved 2011-12-16.
  4. ^ "ACM Fellows". Archived from the original on 2012-01-21. Retrieved 2011-12-16.
  5. ^ "ACM: Fellows Award / Mendel Rosenblum". Retrieved 2011-12-16.
  6. ^ "Mark Weiser Award". SIGOPS. Retrieved 2011-12-16.
This page was last edited on 26 September 2019, at 05:38
Basis of this page is in Wikipedia. Text is available under the CC BY-SA 3.0 Unported License. Non-text media are available under their specified licenses. Wikipedia® is a registered trademark of the Wikimedia Foundation, Inc. WIKI 2 is an independent company and has no affiliation with Wikimedia Foundation.