dcsimg

The Love of Linux

Ransom Love's Caldera Linux Systems became a powerhouse overnight when it acquired traditional Unix vendor SCO. Step two is leveraging SCO's technology to help Linux scale to the enterprise.








Love opener
PHOTOS © GARY WAGNER


When Caldera Systems, Inc. announced that it was purchasing the better part of the Santa Cruz Operation, Inc. (SCO), it seemed like either the greatest move a Linux distribution vendor could make — or the worst. In buying SCO, Caldera picked up a 21 year-old Unix company with developers, a sales force, and a worldwide channel — all crucial elements in its fight to contend in a market increasingly dominated by rival Red Hat software. But, Caldera also picked up some less desirable baggage — namely a reputation for being anti-Linux and an overwhelming portfolio of technologies that may or may not be rendered obsolete by the very technology that helped Caldera raise the capital to buy SCO: Linux. Last August, Linux Magazine editors Robert McMillan and Jason Perlow, along with publisher Adam Goodman, sat down with Caldera CEO Ransom Love to discuss his plans for SCO and Caldera.

Linux Magazine: What was the most valuable thing you got from acquiring SCO?

Ransom Love: Well, number one was global, knowledgeable infrastructure. We have regional Linux companies and Linux resources that are so scarce they’re like gold. You have to steal from your competitor in each of those regions to even come close to having global infrastructure. We get it instantaneously.

Number two, we get some incredibly powerful technologies. One of the keys to the open source model is constant innovation. You can constantly be providing technology, constantly showing yourself as a leader of innovation, and that will drive people to your platform — your “brand” as people like to say. But, it’s more important than brand. It’s actually delivering on the fact that you know where you’re going. There’s a direction, a road map. There’s an element of confidence. The clustering technologies we have on Linux today are not even close to what we acquire with the SCO relationship. And, when you think about it, who else can do that in the industry? Well, there’s only one other software company that can; and that’s Microsoft. Being the alternative to Microsoft is a good thing.

The third thing is the channel. When you talk server and Internet infrastructure, you need someone who understands the Internet, its complexities, and how to provide those services; and so we have the broadest channel in the industry ready-made for Linux.

Then, there’s the application base. One of the things we’ve always hit upon is that we don’t have enough applications online. So, now we have business applications. So, as we move out the OpenServer application base, we’re talking hundreds of thousands of dentist’s offices and medical billing packages. We move those onto Linux seamlessly. And guess what — in a couple of years they will be Linux applications. And, you won’t even realize it, because it will all seamlessly migrate.

LM: But, it’s interesting that while SCO is in all those places, there’s also the general perception from the dot-coms and the Internet infrastructure companies that SCO is this old, stodgy company. How are you going to get past the whole stigma that SCO has with that market?

Love: Well, the beauty of that is Linux. Look at the ISPs; they love Linux. Now we can give them the technology that will scale, and it’s Linux. It will look, feel, act, be Linux, and yet all of a sudden, magically, it scales to the high-end data center needs.

LM: So how do you get Linux to magically scale? My understanding is that you are planning to add Linux services to the UnixWare and OpenServer kernels.

Love: Let’s just look at both kernels. You have OpenServer kernel and you have UnixWare. The transition from OpenServer to UnixWare is a phenomenal leap of technology. The transition from OpenServer to Linux is much more seamless. It fits from an economic perspective. It fits from a technology perspective. If you look at those vertical applications, they need driver support. Linux provides that. You don’t get that from UnixWare. So, you see, the glue that Linux provides in this story — where you can take the OpenServer applications and move them onto Linux but where you have Linux scale with the UnixWare kernel on the high end — now you’ve got a platform that’s literally seamless for the customer.

LM: But, you have a vested interest in the UnixWare kernel over the Linux kernel, don’t you?

Love: No. Unix doesn’t scale down. If we are successful to where we need to be, the actual base distribution will have much more in common than it will be different. The differentiation will take place at the layers up above that. And so if that’s the case — and we want it to be the case — we really want to see the Linux Standard Base [LSB] succeed. Otherwise, we’ll never get the penetration into the commercial market that we need and the support from the ISVs, never.

Let’s hit that issue of LSB. The LSB needs to go far beyond what it is today. The goal of LSB should be a develop-once deploy-all-around Linux. So, you have to begin to address the standards across the board.

love2
love1
love3

LM: So why do you think the LSB has not been more productive?

Love: Well, there are a number of factors, I think. First of all, it’s a cooperative effort, which takes time. It’s not nearly as simple as, “I have my source code; I’m going to go off and do what I think is exciting.” So, there’s an element of collaboration that’s needed.

And, for some, there’s less motivation. I think those attitudes are a little short-sighted. I think we’re moving to the stage of productization of Linux where there’s plenty of differentiation to be made on the products that you’ve delivered.

LM: So, is the LSB’s lack of progress a real problem?

Love: When you talk to [LSB Chairman] Dan Quinlan, he feels very comfortable with the contributions that are being made. I think that some of it is because the industry is growing up rapidly. The need for LSB is expanding rapidly, and I think it’s going to take a little bit more of a proactive effort on all parts. And, I think we all need to begin to implement. In other words, there are aspects of the specification that we can implement; so let’s implement them now. Let’s have the distributions show by our actions that we’re willing to perform to it. Then, you’ll start seeing who is and who isn’t willing to support the LSB.

LM: What kernel will your customers be using on the new Intel IA64 processors?

Love: If you take what we’re doing with Monterey [IBM and SCO's joint venture to create a next-generation Unix for IA64; it was recently renamed AIX 5L] and IBM, we can, again, move that layer across that platform just as easily as we have done with UnixWare. Obviously, there’s a whole relationship as we come in and begin to negotiate with IBM what Monterey is moving forward. We have the contract that says what we have to deliver today, and these are all areas of discussion that we can have with IBM.

LM: What parts of UnixWare do you plan to open source?

Love: Let me back up first. We will provide open access to the UnixWare source under an open source license. Our commitment was, when we made the announcement — for every product that we own and every component that we own — that we will provide open access to the source.

Now the key is, in some cases, it may make sense to use the GPL –depending on what the technology is. In other cases, the BSD license may make a lot more sense. It just depends on what we want to accomplish with the technology. I don’t believe that all technology has to be open sourced under a GPL license. I think that where the commodity platforms are going is, if you don’t provide access to the source, you’re not going to get the breadth, the market share, the scale of economy that you could get if you did.

LM: Is it going to be a major job to untangle all of the third-party proprietary code in UnixWare?

Love: The process of de-entanglement goes clear back to when Novell owned UnixWare. And, Novell made huge progress in separating out all of that. SCO has continued that effort. So, it’s a lot cleaner than people would expect. We haven’t gone down to the depths, but, as we talk to those who know the technologies, they do not foresee any major obstacles with our initiatives at all. This is something that I think has been contemplated for some time.

LM: Do you believe that you could GPL all of your software and still derive revenue from it?

Love: Sure, no question. But, I think the real value that we provide is this constant innovation, constant enhancement, constant development.

What we’re saying is as infrastructure becomes the standard, it has to be in a GPL or an open source license for-mat. There are layers above that don’t necessarily need that license. But, before you’re forced to, you need to give it back. And then you move on.

LM: Before you’re forced to?

Love: Well, no. This is what happens in a traditional IP model, a proprietary model: We have our IP. We’re going to protect it. We’re going to nurse this baby. And, perhaps it becomes a cash cow, and we can never make the choice to move on to something else because we’re trying to protect this cash, this revenue. And the problem is, in a publicly traded environment, that becomes an obsession, because you have to meet your quarterly numbers. So, that’s one of the problems with our industry today. What we’re saying is, you’re voluntarily giving it back long before it ever becomes a cash cow. That forces you to constantly innovate, to move forward.

LM: So, you subscribe to the Aladdin Ghostscript type of model where part of the software is open source, but the most cutting edge part is not?

Love: I think there’s room for a lot of different models. But, I don’t criticize any of them as long as there’s a social contract that says that they are going to give it back. It may depend on the technology. If I wanted to create a standard GPL or BSD, one of the standard licenses is clearly the license that I would use.

LM: So, here’s the potential problem. Let’s say you guys create this UnixWare kernel that has Linux APIs, and everybody is writing applications to Linux. What’s to stop your customers from deciding they don’t want to buy the next Caldera release and instead migrate to Red Hat? How do you keep your market share? What is the incentive?

Love: The incentive is, again, the social contract. What has Caldera done? Have we innovated? Have we given back? When you talk about the layers, as you start moving up above the kernels, what have we built? Products like Cosmos [Caldera's Web-based management software]. You can use it to administer Red Hat Linux. So, on our services side, if customers want to deploy Red Hat as the kernel, hallelujah, we don’t care. Because they’re going to use Cosmos to do it. In other words, we will make it just like bottled water; switching between brands is not that big a deal. It’s the value that we can provide above that.

LM: You mentioned before that you were interested in becoming the alternative to Microsoft. When you say that, do you mean that you see Caldera being an alternative to Microsoft, or do you see Linux being an alternative to Microsoft? How do you see your relationships to the rest of the Linux community in that regard?

Love: Clearly, Linux is a community, collaborative effort. We’ve never tried to superimpose ourselves and say we represent Linux. And yet, being a vendor of Linux, we do have to stand up and say this is who we are and we’re willing to take responsibility for the products that we ship. We’re willing to take all of this technology that’s clearly produced in a collaborative effort, and which we clearly contribute back to as part of that social contract, and we are going to put it together and productize it and make it viable for businesses. When we stand up and provide businesses with some sort of road map of where it’s going to take them, and how they’re going to migrate, that’s being responsible to the customers.

We will have the breadth of technology, the breadth of services, the global infrastructure, the breadth of partnership relationships that will allow that customer to deploy our products worldwide and thus have a viable alternative to Microsoft. That’s where I think we’re unique as a company with this SCO acquisition. Does that mean that we want to take over and dictate to Linux where it’s going to go? No.

LM: I think that there are a lot of questions raised by the Novell and SCO history and its antipathy toward Linux. Caldera’s mission seems somewhat less clear than, say, Red Hat’s mission…

Love: I disagree with that. I actually think that it’s very clear what Caldera is doing. I think it’s clear that we’ve been going after the reseller channel, the vertical solution provider, and the ISP market space. But, I see Red Hat as more of a retail product.

Let me give you some history. When you talk about our Novell roots, Novell gave us a very good broad industry exposure to all the major partners, to all of the networking technologies, lessons on how to partner, and experience with what’s necessary for an OS infrastructure to be successful. That’s what Novell gave us. They gave us a lot of negative experiences too — of things not to do in a company.

LM: Like what?

Love: Well, don’t be arrogant; they’re very difficult to partner with.

LM: How did Novell view Linux?

Love: Even inside of Novell we saw Linux; and we said, “the industry needs an alternative and you cannot compete against Microsoft on their own terms.” The OS/2s of the world have proven that where you go out and try to compete and say, “We’re going to hire 600 engineers, and we’re going to try and create an operating system that’s going to compete on commodity prices against an entrenched leader,” the economies of scale simply are not there. It’s a huge investment to get anywhere close to the availability. So, we knew that we had to change the paradigm. This was eight years ago.

I hold up [Novell founder] Ray Noorda as a mentor. I don’t agree with a lot of things that Ray did. But, he has an incredible knack for understanding technology from almost a gut, raw level. We proposed Linux to him eight years ago. There were, maybe, only 20 or 30 people on the kernel list when we proposed to him that Linux was actually a viable alternative. And, he stood up to all the opposition inside that company to endorse this skunkworks project and let us start to create this stuff. And then, when we spun out, he funded it to the tune of millions of dollars. And, kept with it long before anybody ever validated it.

We knew, and Ray knew, that something had to be done with Microsoft. And, you couldn’t beat them with just great technology. We knew we could change the paradigm using open source. But, we knew that legally you had to address them, because they had a lock-hold on the distribution network, and you could never get your great technology out because of Microsoft’s pricing practices and the things that were going on that were totally illegal.

So, we knew we had to fight them in court as well as fight them with better technology. So, again, Ray took that on. It wasn’t a vendetta against Bill Gates. It was, “If this industry wants to move forward, you have to break that environment.” Ray knew eight years ago it had to be done. And, in both cases, he’s been vindicated.



Robert McMillan is executive editor for Linux Magazine. He can be reached at bob@linux-mag.com.

Fatal error: Call to undefined function aa_author_bios() in /opt/apache/dms/b2b/linux-mag.com/site/www/htdocs/wp-content/themes/linuxmag/single.php on line 62