About Open Voices

The Linux Foundation Podcast featuring conversations with Executive Director Jim Zemlin and Open Source Luminaries.

Entries RSS

Edward Screven Interview

Edward Screven Interview
March 20, 2008

Jim Zemlin: I’m here with Edward Screven from Oracle Corporation as part of our ongoing series called Open Voices, which brings leaders from the Linux and open source community together to talk about various issues—technical, social, and so forth—that affect the platform. Thank you, Edward, very much for joining us today.

Edward Screven: Thank you for the opportunity.

Jim Zemlin: I’d like to talk a little bit about your personal background because I polled a few folks in the Linux community, and told them that I was very excited about speaking with you today and many of them had never heard of you before. I don’t think in a bad way, but I think it would be good to let some folks hear about your specific role at Oracle and sort of where you have come from in the technology world.

Edward Screven: My title is Chief Corporate Architect. So I report directly to Larry Ellison, who’s Oracle’s CEO, and I help that product direction across the company and make sure our products are lined up with our company strategy. And I have a few more specific responsibilities and one of my favorites is I’m responsible for our Linux engineering efforts. So I’m responsible for our Unbreakable Linux program. So I’m sure many of the listeners probably have heard of Wim Coekaerts. So Wim reports to me and he has all the Linux engineering and support operations under him. So I’ve actually worked for Oracle for a very long time. I joined Oracle in 1986 and joined straight out of college, so Oracle is, you know, basically my only real job.

Jim Zemlin: That’s impressive. Have you reported to Larry almost that entire time?

Edward Screven: No. I started reporting to Larry probably mid 1990’s or so; maybe early 1990’s.

Jim Zemlin: Impressive indeed. That’s a rarefied few. And before that, you just joined Oracle out of college. Are you one of the few people at Oracle who’s been there almost the entire time for your entire career? That’s really impressive.

Edward Screven: Yes. There are not too many people who’ve been around Oracle as long as I have. I think, you know, there’s probably 20 or 30 people. When I joined Oracle of course, it was a much smaller company, but even when I joined in ’86 they had already gone public, like probably three or four months before, and they had existed as a company since, I think, ’79 or so. You know, Oracle had been well established and already well on their way to great success by the time I joined. I mean, I wish I had managed to join a few years earlier, but hey, you know, I was only like 17 back then.

Jim Zemlin: I suspect that a lot of our listeners can relate to that.

Edward Screven: Yes. {Laughs} Yes.

Jim Zemlin: I mean, if you’ve been around since 1986, you have seen a lot of ebbs and flows in technology, a lot of historical events; a lot of turning points in the history in terms of platform technology. Tell us a little bit about the history of Oracle’s use of Linux. When did Linux first hit Oracle’s radar?

Edward Screven: For any listeners who are familiar with Oracle database for a long time would know that we have a really strong focus on cluster database. And for Oracle database customers, they know that as RAC—which stands for Real Application Clusters. We started working on clustering technology in the database in the probably late ‘80’s and made great strides into the ‘90’s. And some time around, let’s say, mid to late 1990’s we realized that commodity computers—that is computers that are based on x86 architecture chips—were starting to get faster and more reliable and that we could use our database clustering technology to tie them together and solve significant database problems at a cost which was very appealing to customers.

But of course we still needed a kind of operating system. And, you know, if you looked around back then at what was available, I mean, basically it was Windows or not too much else commercially supported, and certainly nothing else was commercially successful. But there was Linux. And, you know, it was very early in Linux days, especially for trying to have a production deployment in a data center, but still we decided Linux was probably going to be very successful. And so we produced the first commercial database port to Linux in 1998, I think. After that we created the Linux engineering team and we began to hire engineers who knew Linux and also understood the server software space; in particular, you know, the kind of requirements that, say, the Oracle database might have. And so we worked to try to help make Linux a better operating system for that kind of server software and also helped make our software run better on Linux.

Jim Zemlin: Yes, I mean, most people in the Linux world see Oracle’s decision to port Linux and improve it up there if not equal to IBM’s large investment at a similar time frame. And certainly it’s something that really enabled the ecosystem from a business perspective around the Linux platform. And the way you describe it, it’s sort of intuitively obvious that in retrospect it was with a great deal of foresight, though, back then. In terms of when that decision was being made, was there resistance or was there, in terms of intellectual property, open source issues, a lack of understanding or a risk aversion when looking at, you know, what does Linux really mean to the organization?

Edward Screven: I don’t think there was any risk aversion then. At Oracle we have the benefit of a CEO who has a very strong technical sense and a really excellent business intuition. And, you know, he was able to grasp the potential of Linux and he was able to understand the implications of the licensing model on business. And he realized easily that, in fact, there was no downside for us to support Linux. Now, of course, we were careful to obey the GPL and, you know, not entangle any of our other intellectual property in such a way that would be improper. But we didn’t view GPL as something that was going to get in the way of business in the least.

Jim Zemlin: It’s interesting because you hear a lot about companies where they describe entering into open source as, “Well, at first it was an experiment and we weren’t sure, and there were all these reservations and concerns,” and every time I talk to someone from Oracle they’re, “Yeah, no it was the obvious choice and we made sure we lived up to the letter of the licensing, but we just went ahead and did it.” In terms of Oracle’s use of Linux internally, I often hear that Oracle runs its entire business on Linux. Is that the case today?

Edward Screven: Oh, absolutely. Oracle definitely runs on Linux. We have very few servers in our infrastructure that are not Linux; that support, you know, internal IT systems, very few. And even the ones that continue to exist are on a plan to be phased out. So we definitely run our business on Linux. In fact, I mean, our entire IT infrastructure is Linux, our entire development infrastructure as well. So, you know, our development platform is Oracle Enterprise Linux. Our test platform is Oracle Enterprise Linux. You know, our On Demand business where we have a few hundred customers that we host running mostly business applications, the entire thing runs on Oracle Enterprise Linux.

Jim Zemlin: That’s interesting. In fact, I have a question a little later on about your application – the Oracle On Demand business and the use of Linux in virtualization in that business. But I want to save that for the end because I think that, while a lot of folks know that you guys have been long supporters in terms of supporting Oracle and running on Linux and that you use Linux in your business, a lot of people would like to hear about why Linux decided to get into the Linux distribution business in terms of supporting Linux in that way through Unbreakable Linux. What was the decision making process that you went through to decide to make that announcement almost a year and a half ago, I believe.

Edward Screven: Yes, we don’t really view ourselves as being in the distribution business.

Jim Zemlin: Right.

Edward Screven: We see ourselves as being in the Linux support business. I think there’s an important difference there. I mean, we don’t try to compete by creating a differentiated distribution. We don’t try to compel customers to subscribe by withholding binaries. You know, anyone on the planet can download and use Oracle Enterprise Linux binaries for free. You know, if you want support from us, you pay us. But we’re not trying to compete in the distribution business.

Now, you might ask then, “Well why did you get into the Linux support business?” Well, our goal is for Linux to be a practical and cost effective choice for our customers, and especially enterprise customers. The existing Linux vendors I think have a little bit different point of view and I don’t think that they were doing a very good job. You know, they were charging a lot of money for support levels that, in our minds, were insufficient for many enterprise customers. And the implication is that a lot of those customers were discouraged from using Linux for mission critical systems in their data centers. Now, we really want Linux to be the default choice for Oracle customers in their data centers. So we got into the business to fix it.

Jim Zemlin: How hard was that to do? And I think you’re right, it is very important to understand you’re in the Linux support business and that you enable supporting what you’re calling Oracle Unbreakable Linux, which is using the Red Hat code base, which I’ll come to in a second. But in terms of building up that support infrastructure, how difficult was it? Could any company do this? Is there unique attributes at Oracle that enabled this to happen for you?

Edward Screven: It was easy for us, but only easy for us because we had made such a substantial investment in Linux along the way. When we first put together the Linux engineering team, it wasn’t our intent then to be in the Linux support business. But it was our intent, you know, to help make Linux better, help make our products better in Linux, help our customers use Linux. And so over time, you know, we hired more and more Linux folks who knew how to create builds and knew how to create patches because, you know, in 2003 or so, we started directly giving customers patches to Red Hat and SuSE Linux when they ran into P1 problems. So we had that infrastructure all built up. So by the time we actually launched Unbreakable Linux as a full commercial support offering in, I guess, October 2006, we had already had hired the people in place and we already had the support infrastructure; we already had the build systems, we already had the test systems. You know, we had everything ready to go. All that infrastructure existed because Linux is so strategic to Oracle and because it was so important to our customers. You know, if we had to start from scratch, it would be expensive, but we didn’t start from scratch. We started from, you know, an existing strategic commitment to Linux.

Jim Zemlin: What was the decision making process behind using the Red Hat code base?

Edward Screven: It’s really our desire to encourage the market to move to a single distribution. Red Hat has by far the largest market share in the data center, and especially for Oracle customers. So it made sense to pick Red Hat as our base. Now if the Red Hat and Novell numbers were reversed we would have picked SuSE.

Jim Zemlin: And could you describe for people who aren’t familiar with exactly what that means, what exactly does it mean for Oracle to be supporting this code base?

Edward Screven: What we do is we decided not to create a distinct distribution. We decided to instead support the Red Hat distribution. And the way we do that is that we continually pull from Red Hat source code for Enterprise Linux and we compare it to our own source code repository where we have incorporated, you know, some bug fixes. You know, most of the bug fixes either come from customers or they’re generated as a result of our extensive server testing. And we fold our bug fixes in, we compile them and we distribute them. And we distribute them as freely downloadable binaries from oracle.com. We also, of course, have an update network where our subscribed customers, you know, get incremental fixes over time.

In all cases, we’re very, very focused on making sure that what the binaries that we distribute either as, you know, individual package updates or as complete installs is completely compatible with Red Hat Linux. And, you know, our goal is to make sure we do not cause any fragmentation in the Linux market space. And also our goal is to make it very easy for customers to switch from Red Hat Linux to Oracle Enterprise Linux. So, in fact, a customer who decides to subscribe to Oracle for support and has Red Hat Linux installed doesn’t have to actually change anything about their installation. The only thing they need to do is point to our update network instead of the Red Hat network and, you know, they’re supported by us. So any bug fixes they need will get propagated down; any incremental, you know, updates will get propagated down. They never have to do a machine reinstall.

Jim Zemlin: How do you respond to some of the criticisms from Red Hat that there’s no way to achieve that 100% binary compatibility promise?

Edward Screven: We’ve been in the business now for, I think, 17 months and during that time we’ve gone through a few updates and a major release. And so I would ask Red Hat or any other potential critic to find a binary incompatibility; find a place where there is a functional difference between Red Hat Enterprise Linux binaries and Oracle Enterprise Linux binaries. The only practical difference that I know of is the difference in label string. So let me give you an example of the extent to which we at Oracle believe that we are consistent and compatible. Obviously, we produce a lot of conventional software running on Linux, including the database, that we develop and we test on Oracle Enterprise Linux. We do not test on Red Hat Linux, yet we release our products to our customers certified and supported on Red Hat Linux. And we can do that because we know with certainty that they are the same.

Jim Zemlin: So as an ISV, you’re willing to support the Oracle database on Red Hat because you’ve done that testing on Unbreakable Linux and that’s something that as a company you’ll stand behind.

Edward Screven: Absolutely. That’s right. If a difference were to emerge, so someone found we made a mistake when we fixed a bug or some similar thing, we would view that as a bug in Oracle Enterprise Linux and we would fix it.

Jim Zemlin: Do you think that other organizations could follow suit in terms of what Oracle has done in supporting Linux?

Edward Screven: It’s possible. I don’t think there are that many organizations that have the same, you know, strategic commitment to Linux as Oracle. I think there will be new distributions arising in the future, but I think those distributions would be focused on some more niche functionality or niche application. So perhaps some focus on embedded, or focus on real time, you know, something more specialized than the general purpose server market. But it is very possible for another company to enter into the same market.

Jim Zemlin: One of the things that you mentioned was that this is not fragmenting the Linux marketplace; that you really believe in having a cohesive Linux platform. What is Oracle’s opinion on the need for standards? Things like the Linux Standard Base which helps prevent “forking” of Linux. As an ISV yourself, you just gave an example of where Oracle supports your database on Red Hat, what about on other Linux distributions? What as an ISV are some of the things that you see are either difficulties or challenges that you have in supporting incremental Linux distributions; Novell, Asianux, and so forth?

Edward Screven: Well first I think Linux Standard Base is a very important baseline. It does provide a target for ISV’s and it helps, I think, to keep, you know, the various distributions in line. From our perspective though, we really believe that for general purpose server deployment the kind of distribution that a customer is using should really be something that’s very close to mainline. We think the right model for Linux vendors is not to try to compete through differentiated distributions, but rather to compete on support, right. And that the vendors in general purpose Linux distribution support market would be better off cooperating on creating a single distribution and competing on support. Customers would be better served, the Linux market would grow faster, and there would be, I think, more commercial opportunity in general than there is right now.

Jim Zemlin: Just to follow up on that thought, you know, I travel all over the world and I meet with government organizations in China who are sponsoring efforts to create a Chinese Linux distribution or in Brazil or Korea or all over the world. And there seems like there’s a lot of efforts to have a national Linux distribution. Any advice you’d give these efforts in terms of the experience that you’ve had at Oracle?

Edward Screven: What I would tell them is work with mainline. Work to make mainline closer to what you want. There is of course an aspect of packaging that makes sense; that helps make Linux more consumable by general customers. But creating forked, different versions of Linux is not healthy for Linux. It’s not healthy for the entity that’s creating that forked, different version of Linux. You know, ultimately everyone would be better served if we work together to make Linux—all Linux—better.

Jim Zemlin: Who do you see as your primary competitor for Unbreakable Linux?

Edward Screven: Microsoft and Microsoft Windows. You know, we, as a company, we have a few large primary competitors. One of them is Microsoft. So Microsoft wants customers to deploy Windows as a vehicle for locking them into Microsoft’s monopoly. Linux is an important counterbalance to Microsoft’s strategy. We view Linux as a much better open choice for enterprise customers running x86 computers. So we really want Linux to succeed in that space. You know, of course we fight with Red Hat over customers, we fight with SuSE over customers. We’re competing with each other in this Linux space. But the real opposition here is Microsoft.

Jim Zemlin: And certainly at its core, due to the GPL license, you do share in that code base as a requirement.

Edward Screven: Oh, absolutely. Absolutely.

Jim Zemlin: How is the business growing? How is Oracle Unbreakable Linux doing?

Edward Screven: Good. I mean, we’ve had the doors open for 17 months. We’ve got, I think, more than 2,000 customers. So we’re very pleased with our progress.

Jim Zemlin: Can you give us a couple of anecdotes?

Edward Screven: You know, right now we’re actually in our quiet period, which means we’re going to announce our results in a few days, and I’m forbidden from talking about specific results. But we have many, very large, strategic customers who are adopting Oracle Enterprise Linux, and including some large customers who are standardizing on Oracle Enterprise Linux.

Jim Zemlin: It’s interesting. I think I saw an article about a customer who had been on Windows and migrated over to Oracle on both the database and on the Unbreakable Linux side.

Edward Screven: Like I said, Linux is a strategic element of Oracle’s full offering to the customers. And you think about, “What is it that we’re really trying to provide customers?” What we’re trying to provide customers is a complete and integrated solution that’s based on open standards. And, you know, within that large suite, Linux is the operating system. And Oracle VM is the virtualization layer. And then of course we have the database and we have middleware and we have applications and lots of other components all fit within that. You know, so we have a different approach than Microsoft. Microsoft has this really proprietary, locked in strategy, right? We’re different. You know, we want everything we build to be based on open standards. Some of the pieces are also based on open source, right? But the key element of our strategy is that we want the components of our solution, while they’re tested and integrated together, to be replaceable by the end customer. It maybe sounds a little strange to have a vendor say we want our components to be replaceable, but the truth is most customers are risk averse. And by having our components be replaceable, they’re more likely to choose them to begin with, because if they don’t like them, they can get rid of them.

Jim Zemlin: Yes. I think one of the things that a lot of folks that listen, perhaps, to our broadcast are maybe not as used to hearing is whenever you speak with somebody at Oracle about open source, you get a general matter of fact answer of, “Yes, we use it when it’s helpful and when it’s not we don’t. And it’s just part of our day to day business and that’s how we’re being competitive.” {Laughs} There’s no huge hour long strategy discussion about it or the implications from licensing perspectives or whatnot. It’s just like, “Yes. No, it works for us.” And that’s fascinating to continue to hear it from you as well.

I want to back up and talk a little bit about broader market issues around the Linux platform. You already mentioned Microsoft as a competitor to Linux, certainly, and Unbreakable Linux and, certainly from a strategic point of view they’re a huge competitor of Oracle. How about Unix? How do you see the future of Unix? Who are your competitors in that market? And do you really see the world increasingly consolidating on a Linux/Windows type of framework?

Edward Screven: We have a lot of customers that use Unix. And for a lot of those customers, you know, a big iron solution based on Unix is a good choice for their application. Now, I would say I do believe from talking to customers that, you know, they are migrating over time to data center architectures that are grids, right? So there are lots of racks of identically configured x86 architecture machines running, hopefully, Linux. And at Oracle, we’re really trying to help customers as they move into that kind of infrastructure by supporting Linux and by creating the Oracle VM product, by creating the Grid Control product, and making sure that our software works together well in clusters. So both middleware clusters and database clusters. I really see many more conservative customers beginning to migrate in that direction. Now, a lot of customers of course have already moved there. Oracle moved there years ago, right. If you walk around in our Austin data centers, it’s huge, right, and what you’ll see is racks and racks and racks of identical computers all connected and wired in identical ways, all running Oracle Enterprise Linux and now beginning to run Oracle VM.

Jim Zemlin: In some ways, is that going to benefit the mainframe business in terms from a hardware perspective – that cluster environment that you’re describing sort of does hearken back to those shared workloads that you remember from the mainframe computing days?

Edward Screven: I think in a way you could look at grids as the new mainframe. Yes, we called our virtualization product Oracle VM for a reason. IBM invented that notion, you know, I guess, or maybe perhaps it was the Cambridge Lab who invented that notion many years ago. And it was a good one. I mean, it’s a good way to make hardware fungible, right. And so if you combine together virtualization and clustering, you really can turn the problem on its ear. So instead of saying, “Well, I’ve got to figure out for this application what hardware to buy to run the application?” instead you say, “Well, for this application today, what level of resource do I need?”, right. So then I can go into my pool of hardware resources and consume the level of resource that I need to meet a given quality of service for that application today. And when the resource level changes tomorrow, well, I can use my virtualization and clustering technology to adjust the level of resource that’s dedicated to it. So overall, the amount of hardware resources you need would be less, the amount of power you consume will be less, the amount of management—labor—that you need, will be less.

Jim Zemlin: And you can certainly charge back on a business unit level based on application utilization in a much more transparent and clear way.

Edward Screven: That’s right. A lot of our customers, you know, that’s one of the first things they’ve asked for—Oracle VM—“Is [there] some way to help me account?” And it’s just like it was in the mainframe base, right; charging back.

Jim Zemlin: Tell me a little bit about some of those management challenges and how Oracle is addressing them in that world of sort of these Linux grids with virtualized environments and being able to track those; charge back. I know, I was actually one of the founders of a company called Corio, which was very early on hosting Oracle applications, and we had a heck of a time tracking back single customer economics, you know, like per customer sort of charge back mechanism. And what you’ve described is a way to do that through this virtualization environment. What are some of the management challenges from the software perspective in terms of the tools that are required to that, and then from a business perspective?

Edward Screven: You know, the first challenge is make sure that you’ve got a management level that understands both the virtualization layer, you know, and the application layer, and of course let’s not forget the OS in between. We’ve done that with Oracle Grid Control. So this year we’ll release a version of Oracle Grid Control that integrates virtualization management. So through Oracle Grid Control, we’ll enable customers to deploy and redeploy and undeploy software in a way which is convenient and simple and Grid Control can help take care of the details. Now, Oracle VM has a separate management environment for customers who don’t want to use Grid Control, it can handle all of the virtualization aspects, but it doesn’t handle the software aspect on top. So it doesn’t help you manage the database. With Grid Control, because it understands RAC and understands virtualization and understands, you know, all the bits in between, you know, it’s very simple for customers to manage this kind of much more fluid grid environment.

Now, I have to say—admit—we didn’t actually predict the chargeback feature request from customers, and so that’s something we’ll be adding over time.

Jim Zemlin: Got it. But it’s certainly an area that’s growing in technology and there are lots and lots of folks out there that are getting into that space - virtualized environments. There’s also a new category being created that folks are calling ‘cloud computing’ where, you know, you’ve got the – IBM’s got their Blue Cloud, Amazon’s got EC2. There are people predicting a world hub hearkening back to – I can’t remember who it was who predicted in the ‘70’s that there would only be five computers in the entire world and people would tap into those resources. Do you at Oracle see a world emerging where you have these huge compute farms where people tap into them to be able to build these virtual environments from companies like Amazon and IBM and Google and perhaps Oracle?

Edward Screven: I think what you’ll see is a wide spectrum of choices. The customers make {decisions} because of differences in their business, differences in their requirements, differences in some aspect of their business, so for example one kind of point in that space is cloud computing, UC2 style; great for stuff which is CPU intensive and which rises and falls over time. You know, another kind of example is software as a service, so at Oracle we have CRM On Demand offering where customers, you know, subscribe and sign up and do saleswares automation, right. Another data point might be something like our E-Business Suite On Demand offering where there’s computers in our data center that host our business software for one customer, right. And then of course lots and lots of customers run machines in their own data centers. And I think all of those kinds of deployment scenarios will continue to exist. You know, not every customer is going to want to have every application administered by someone else, and certainly they’re not going to want to have it deployed in a cloud. They’re going to demand a certain level of reliability and stability that may not be achievable in every kind of deployment choice.

So I think the more interesting question is, well, what’s common about all of those things, right? So, and I think that what’s common about it is a grid architecture actually works in all of those scenarios. You know, when I see customers coming to Oracle, and especially customers who go down and look at Oracle’s Austin data center, what I see them asking about and talking about and thinking about is that, even in the data centers that they are going to continue to own and operate, what they want is they want a grid environment which looks essentially like a cloud computing. The only difference is it’s owned and maintained in a private way as opposed to owned and maintained in a public way.

Jim Zemlin: And Oracle would obviously be enabling technology in both respects.

Edward Screven: That’s right. And also a vendor of those kind of environments in that we have an on demand business. The way our On Demand business is built is using, you know, grid architectures and using the technologies of Oracle VM and Linux and our cluster technology to provide that computing environment to hundreds of customers we currently have.

Jim Zemlin: In many ways you can look at what infrastructure the hosting providers and Oracle On Demand is using as a leading indicator because they’re so driven to have better margins by reducing the overall support footprint and cost of delivering that software as a service that they, you know, there’s a high incentive to utilize new and innovative technology to get there.

I wanted to move to a couple questions about open source in general. You’ve talked about how Oracle uses open source software, where it fits the business and where it does not. What do you think of open source applications beyond infrastructure software like Linux and that core of technology layer? In terms of the application arena, you see things like SugarCRM, Content Manager, there’s a business intelligence company called Caspersoft. Are any of these projects companies on Oracle’s radar screen?

Edward Screven: Well, we’re aware of them. We know about them. I think it’s very possible to have an applications business which is built on an open source model. I don’t think open source only works for system software. I think it can work for application software too. It’s not the way our application software is built or the way our business model runs, but, you know, it’s a perfectly fine model for SugarCRM, I’m sure.

Jim Zemlin: You’re not seeing them as much in direct competitive situations, but you believe that there is a business to be had around delivering open source applications, even at that higher level of enterprise software?

Edward Screven: Sure. Yes.

Jim Zemlin: I’d like to ask a couple questions about some recent industry events that have come up. The first one I’d like to ask about is Sun’s acquisition of MySQL. How does that affect Linux? Obviously you’re supporting Linux with Unbreakable Linux - part of your Enterprise. How do you think that acquisition affects it? Does it affect it any way? Do you even care?

Edward Screven: Well, the company I’d say, overall, we just don’t care. I mean, we don’t see MySQL very often, again, in competitive deals. It’s out there, but it’s not very often that a database sales rep comes back and says, “I had to compete for the business against MySQL.” I don’t really think Sun’s acquisition of MySQL has a lot of implications for Linux. I mean, certainly there are Linux deployments that exist because of MySQL and if Sun screws up MySQL, then of course that may mean fewer Linux deployments. But, you know, there are plenty of database choices on Linux, commercial and otherwise. I think Linux is much stronger than anything that Sun could possibly do to MySQL.

Jim Zemlin: One other competitor you mentioned earlier, Microsoft, recently opened up its protocols; put 30,000 pages worth of protocols on the internet and stated that it wants to reach out to other industry players and to the open source community. What is your take on that announcement; why they did it and how effective you think it will be for Microsoft and for other software organizations and companies to interoperate with their platform?

Edward Screven: First of all, I don’t know if you’ve looked at those protocol descriptions…

Jim Zemlin: {Laughs}

Edward Screven: …but a lot of them are essentially unreadable; unusable.

Jim Zemlin: I’m on page 32 of 30,000.

Edward Screven: Yes, okay. So second, what some lawyers have told me is that it’s a license term for those protocols that are incompatible with GPL.

Jim Zemlin: That’s the Rand licensing terms for the patents, I believe.

Edward Screven: Yes. So that’s kind of a nonstarter. And finally, I think what customers need— what the world needs—is for Microsoft to decide that open standards matter. That publishing your own proprietary standards that you’re free to change at will, right, and free to make undocumented incompatible implementations of are not helpful to customers. Open standards mean that everyone who has an interest has a chance to participate. And one vendor doesn’t totally dominate. So publishing protocols is not the same thing as using open standards.

Jim Zemlin: Do you think this is an internal dilemma at Microsoft; that they’re sort of going through fits and starts seeing the world moving towards using open standards, creating technology and innovation both in and outside of the enterprise, that they’re sort of conflicted and want to hang on to their practices of the past? Or is this just another way for them to compete and continue to dominate?

Edward Screven: I doubt that they are conflicted in the least. You know, I think they are just tired of paying fines to the European Union.

Jim Zemlin: {Laughs} What do you think are some of the most important issues that are facing technology today in terms of open source? You mentioned open standards and I think that there’s a lot of people in the industry who have a difficult time understanding the difference between the two. Or they often get confused. Why are open standards so important and is open source complementary to developing those type of standards?

Edward Screven: Well, open standards are important because they enable multiple competing or alternate implementations. That’s important if you’re using open source based implementations or not. Now I think there is a natural connection between open source and open standards because open source implementations often are the first implementations in open standards or are often a widely available implementation of open standards. So they kind of create the market for products based on that open standard. You know, it’s very important for the database world that there are multiple implementations of SQL. SQL’s a very important standard.

Jim Zemlin: Right.

Edward Screven: Now, there didn’t happen to be any open source implementations of SQL for a very long time, right? But if we look at more and more recent times, say in the middlewares space, you know, there are lots of implementations of open standards which are based on open source. And that has helped create this dynamic market for open standards based middleware. And that’s been very beneficial to customers. And for companies that are cognizant of this - that means they’ve been able to take advantage of it and produce commercially successful products.

Jim Zemlin: One of the things I hear you saying is that open standards enable multiple competing implementations of that standard, whether it’s closed or proprietary. And getting the two confused, let’s say we use open source software but it doesn’t comply with any standard; it’s not within the mainline of a particular large project that people are utilizing. Could you get just as well locked in to an open source implementation that doesn’t use an industry standard as you could a proprietary one?

Edward Screven: Absolutely. Let’s say that Tomcat were the only servlet engine that existed in the world; there was no standard for it. And you decided to use Tomcat and then Tomcat went down a road that just didn’t work for you anymore. What would you do? I mean, you’d have to hire programmers to fix Tomcat yourself. Whereas because it’s based on open standards, you know, if Tomcat doesn’t cut it for you anymore, then you can choose to use something else. You can choose to use a commercial offering. So the fact that it’s an open source offering which is based on open standards gives protection to consumers of Tomcat, you know, even though Tomcat is open source.

Jim Zemlin: Yes. This is one I think {of those} very little known things, because you hear so much about open source being equivalent of guaranteeing choice. But what you’re describing is that access to the source code alone does not guarantee the choice. The key is to reduce the cost of change by enabling your applications to work with an open standard and therefore, whether it’s open or closed, you can move at a very low cost to make that move.

Edward Screven: Yes, that’s right. Exactly.

Jim Zemlin: Let me talk about a couple of other fun things that all the Linux folks love to speculate about and talk about because it’s something that almost everyone who uses a computer can relate to, and that’s the Linux desktop. And I realize that Oracle is largely on the server side of the technology business, but what are some of your thoughts on the Linux desktop; the growth of Linux desktop and the future of that as a meaningful computing platform?

Edward Screven: Well, I hope it succeeds. I hope it grows and I hope it manages to begin to displace, you know, Windows in certain markets and certain applications. But you’re absolutely right. It’s also not Oracle’s business. We’re not in the desktop business. If a customer came to us and wanted us to support their desktop application on Oracle Enterprise Linux, of course we would, right. But we’re not working towards trying to get customers to adopt Oracle Enterprise Linux on desktops. It’s a hard road though. I mean, Microsoft’s desktop monopoly is really driven from their Office products. Most business users, you know, need to use some combination of Microsoft Word, PowerPoint, and Excel. I think the OpenOffice people have made a lot of progress, but it’s really hard to get the typical business person, you know, to let go of those Microsoft Office products.

Jim Zemlin: I’d like to wrap up by asking a few questions about a sticky issue that the software industry faces and that is around patents and intellectual property. I believe that Oracle is a member of the Open Invention Network and has opinions on patents and, you know, has seen saber rattling from Microsoft about the patents as they relate to the Linux platform. Could you give me a few thoughts from Oracle’s perspective on the role in patents in the software industry; whether or not you believe there is need for reform along those lines, and then specifically about the issue of patents related to the Linux platform?

Edward Screven: We have patents. We have a program for acquiring patents. It’s a necessary part of being in the enterprise software business. Because there are lots of players out there who view patents as a means to get rich. So we need to have patents in order to help defend ourselves against those folks. Now I think there is very clear need for patent reform. I mean, there are predators out there who essentially buy up patents from failed companies with no intent other than to make money by suing people. Now I think that is probably not the most constructive use of our economic potential. I mean, it’s one thing—and folks can argue about this, right, some people may think there should be no software patents at all—but I think even people who think that patents are a bad thing in general for software would agree that there is a very strong difference between a company that gets a patent and builds a product and delivers the products to customers, right, and a law firm that buys a patent in order to sue someone to make some money. So the way they make money is by suing people as opposed to building products, right. And so I think that is where we really need patent reform. Buying patents in order to make money by suing people seems to me to not be the kind of behavior that we want to encourage.

Jim Zemlin: If you think of patents in cold world terms where, you know, you’ve got your defensive pool and everybody else has got their defensive pool, these legal engineering efforts are the terrorists…

Edward Screven: That’s right. You know, they have nothing to lose, right.

Jim Zemlin: What about Microsoft?

Edward Screven: I think in fact Microsoft’s saber rattling about Linux patents {audio dropout} you know, is good news. Because what it means is they are struggling to find ways to demonstrate the value of Windows with respect to Linux. So they’ve resorted to fear mongering. And that’s exactly how we view it. And that’s exactly why we fully indemnify Oracle Unbreakable Linux customers. You know, we’re a big company and we’re going to stand up for our Linux supported customers.

Jim Zemlin: Interesting, and obviously people in the Linux community agree wholesale; we don’t believe that there is a lot of merit to their patent saber rattling. I will ask one final open ended question. If there is anything else that you would like to share in terms of Oracle, its role in Linux open source or future technology trends, and I will let you go.

Edward Screven: Okay, well thanks a lot Jim

THE END

Popularity: 1% [?]

These icons link to social bookmarking sites where readers can share and discover new web pages.