The Love of Linux Chapter 14: So What About the SCO Lawsuit?
I have been asked by many people what my opinion of what SCO did. I have not commented publicly about their actions, because it really requires more explanation than can be provided in a 30 second sound bite. First of all, I would like the record very clear that I do not stand to gain anything from what I am about to say as I sold all of my stock in SCO when they announced the suit with IBM. And the sale occurred long before the stock went to 8 dollars a share after the announcement that MicroSoft had made an investment or the subsequent spike to around $17 per share. I fundamentally did not believe in the suit, not because I do not think that elements of the suit against IBM are not justified (I think the suit against Linux is wrong), but because it had a strong chance of destroying SCO’s own business (which it has), hampering the success of Linux (very little) and seriously destroying UnitedLinux (which it did). It is not how I would have done it and that is obviously why I am not still there. Why are they doing this you might ask? To answer that will take some history.
One of the main reasons Caldera bought UNIX was to extend and enhance Linux. When we went to the main supporters of Linux, Intel and IBM, their responses were less than warm. Intel was almost rude. If anyone has dealt with Intel that is not that unusual J but their reaction was close to anger which did surprise us a little. We proposed that the UNIX kernel could be open sourced and used to supplement Linux and provide a very scale-up solution while the current Linux kernel was focused on providing a scale-out solution. We felt that we could create a common application and even installation environment that would be seamless. We were not only willing to Open Source much if not most of the technology but provide indemnification. They did not want to hear any of it. If I recall Will Swope’s comment was something to the effect, “We have tried UNIX on Intel and we are done with it.” I left the meeting, really questioning why Intel was even involved in Linux. Interestingly enough, Intel was a prominent sponsor of the Monterey project with SCO and IBM, whose stated purpose was to move UNIX on to the IA64 architecture. My questioning and suspicions of Intel’s motives continued when I encountered their resistance on almost every major initiative they sponsored publicly. They would organize industry wide imitative for Linux like OSDL, only to block any and all funding when it came to marketing and promoting Linux, they were supporting the technical advancement of Linux only. They insisted that they would not spend any of their money on marketing Linux. They clearly contributed technology and they were very vocal about wanting to use Linux to destroy any and all Intel UNIX. They seemed bent on using Linux against UNIX but not against MicroSoft. So why are they so much against UNIX when it actually could and should complement Linux? What we were proposing was an Operating System platform for Intel that could scale and possibly displace the majority of proprietary, high-end RISC hardware systems. Our strategy was borrowing from MicroSoft’s own. They had Windows running on top of a DOS operating system and NT, MicroSoft’s own multiuser operating system. Both MicroSoft kernels shared the same application environment so that ISV’s could develop a windows application and have it run seamlessly in both environments. Caldera would open source the UNIX kernel, utilize the Lunix application environment and allow our Replicated Site organizations and VARs develop applications that could either scale using a cluster of smaller servers or for some applications, the power of symmetrical multiprocessing on one server. The difference between what we would offer and what Microsoft could offer at the time, is our platform actually worked and scaled as promised. It could have taken Intel to a very different place much faster in the Enterprise. So why were they so against the concept?
The only obvious conclusion I could come to was they wanted to use Linux to hurt SUN Micro Systems. So why wouldn’t they embrace Intel UNIX to help Linux scale and truly compete with Sun on Intel hardware? The UnixWare kernel of UNIX could scale to 32 way clusters on Unisys Intel machines. All we were lacking was a stable, credible hardware platform that would scale. Obviously, if our UNIX kernel could be made to run on Intel, so could Sun’s, so Intel would be concerned. Intel felt that the commodity platform would eventually push the proprietary RISC processors out of the Enterprise because the processing power was doubling every 18 months. They knew they would be pushing Sun and its competitors out, so they just did not want to leave them any turf on their own platform to prosper.
Ironically, Sun actually did what Caldera proposed to Intel. They published the entire Solaris Operating System under an open source license, but it was a little too late for the company. Had Intel simply supported Caldera’s efforts, we would have been more of a deterrent against Sun as there would have been almost no interest in the open source community to support another Unix kernel on Intel and ISVs would not have tolerated yet another platform to certify and test. However, I think there was more to Intel’s reaction then just an attack on Sun.
It is interesting to note that Intel’s support of Linux began right about the time that Microsoft was in the government antitrust suit claiming that Linux was a major threat to them and their business. We had the biggest laugh when Microsoft actually used the Caldera desktop product in court showing how it was a competitive threat to Windows. This was the time when the infamous “Halloween” letters were leaked to the press. The Halloween documents were supposed internal memos that got published by accident. Many in the Linux community really believed the marketing hype that Microsoft was somehow threatened by them and the documents ran like wild fire through the open source community. I have always been skeptical. Microsoft never makes mistakes like that and Intel and Microsoft have an incredible working relationship. Some say it is a love, hate relationship, but I seriously wonder especially in something as critical as an investment in alternative, competing operating system platform.
As mentioned, it is interesting that Intel was a major sponsor of the Monterey project. To refresh the readers mind, the project was between SCO, prior to their purchase by Caldera, IBM, Intel and Compaq, prior to their purchase by Hewlet Packard. The intent was to move Intel Unix on to the IA64 processor. So why did their good intentions turn so dramatically against Intel UNIX even when it was offered as a way to enhance and move Linux forward?
Regardless of the motives of both IBM and Intel in regards to Intel UNIX, the Monterey Project, later called AIX 5L, was one of the major reasons Caldera purchased SCO. SCO had a very good agreement with IBM that would enable them to move their large customer’s like Nasdac to IA64 as they needed the capacity. IBM had spent considerable effort to ensure Caldera of their intent to continue to support Monterey when we announced the purchase agreement at the SCO Open World event in Santa Cruz. A team from IBM came to the event with a specific agenda to assure both the SCO and Caldera management of their ongoing commitment to Monterey. At the time, the move to IA64 was imminent. Intel was using its marketing muscle on its OEMS and the industry to get them all to embrace Itanium which was going to be their very next generation chip set. All of SCO’s large UNIX customers were concerned about an upgrade path. Most of SCO’s install base did not need IA64. The mainstream OpenServer customers did not need or want the large foot print or enterprise functionality of UnixWare. IA64 was additional cost and overhead the VAR channel and its customers were not interested in. However, what the OpenServer customers and the VARs that serviced them needed was hardware driver and application support. OpenServer was not staying up with the new hardware platforms and the application developers were more interested in the growing market share of Linux.
Caldera felt it could marry the two issues by making the UNIX kernel another kernel option for Linux. Placing the Linux application environment on UNIX would allow software developers and hardware OEMs the opportunity write one application and have it scale across multiple hardware platforms. Linux had won the battle for the developers but the UNIX kernel had a good chance to be a deployment option. Linux was ideally suited for the SCO OpenServer and VAR market if Caldera could get its version of Linux certified by all the Intel OEMs. When those VAR’s and customers needed more scalability, they would not have to change their application. This solution was also appealing to the distributed enterprise or replicated site customer. Marrying UNIX and Linux would enable applications developed for a small retail outlet of McDonalds or Kentucky Fried Chicken which were currently using and outdated UNIX to deploy very low cost software and hardware solutions at the retail outlet but interface with large back end UNIX needs at the corporate offices with a single application. The marriage of UNIX and Linux would enable software developer’s access to the most flexible, scalable, fastest growing Intel hardware platform in the Industry and fully utilize all of Intel’s capabilities and growing maturity. At the time, no one really took Intel serious in the enterprise. IA64 was going to change that. The marriage of UNIX and Linux on the Intel platform would have been unstoppable. The market projections and popularity of Linux on a mature, scalable kernel or on a distributed, low cost Linux kernel was ideal for Caldera’s mixed customer base and for the world at large. The key of course would be to have the UNIX kernel on Itanium, the latest Intel platform in order to have a complete story. Having UNIX on Itanium was more critical for Caldera’s Linux business than even its UNIX business.
Caldera needed the OEMS, ISVs and IHVs support. The popularity and growth of Linux, the UNIX technology Caldera could have provided to ensure compatibility of these applications on Linux along with the power of UNIX, placed Caldera at the center of an incredible industry opportunity. Apparently, the opportunity for Caldera really scared some of the major players like IBM and Intel, whose investment in Linux was not to create another Microsoft, but drive their own agendas. For Intel, while they would finally have a platform that would enable its hardware to be seriously considered in the enterprise, it would allow Sun a foothold on Intel as they could also create or license the Linux personality for Solaris and the married UNIX/Linux platform could really pose a threat to Microsoft. It would appear that all some in Intel wanted was to help produce enough of threat of competition to Microsoft to ease the government’s investigation of antitrust but not enable a true competitor to Microsoft that could erode current business. And as I mentioned, Intel also wanted to launch a Linux/Intel attack on Sun.
IBM’s concerns seemed much less founded. Caldera should have been the ideal Linux partner. IBM did not want to sell the Linux OS. They had stated that they wanted to enable Linux for the enterprise. How better than to utilize UNIX technology like the kernel and compatible libraries and utilities? They needed a Linux provider that could support the OS globally. Caldera had offices all over the world and a global reseller channel allowing global roll outs of products and services with a single Linux vendor. IBM was also very strong in the distributed enterprise or replicated site business market. They sold cash registers and software to the same VARs and retailers as Caldera did. We could not get IBM software or the retail group of IBM engaged for a long time. Steve Mills, who headed IBM software, refused to support Caldera until we had successful negotiated the UnitedLinux relationship. This really hurt Caldera. Without the support of the IBM and other major ISVs and IHVs (IBM was both in our major market) we were losing business to SuSe and RedHat even though we had long established relationships with the distributed enterprise customers. We lost the Target account to Red Hat primarily because Red Hat was supported by IBM.
So when IBM just chose not to continue the Monterey project on Intel, they really hurt our credibility. IBM continued what became to be known as AIX 5L on their RS6000 platforms but not on Intel. When we went back to talk with IBM’s legal department, they simply said we are not gong to deliver. When we asked for some compensation, the offers were absolutely ludicrous considering the economic cost to Caldera for not offering an IA64 UNIX offering to their existing customers. I believe they offered something like $50,000. We considered the matter unresolved, but we focused our attention on trying to collaborate with IBM on UnitedLinux and our joint, replicated site retail business. The lion share of Caldera’s business was in the distributed enterprise, replicated site or retail. SCO and IBM had a substantial amount of common retail business.
IBM was also not very receptive to our desires to merge UNIX with Linux. Early on, we approached IBM about an investment and they flatly turned us down. We had a terrible time getting the software division to continue to support even our Linux products until we were successful in creating UnitedLinux but by that time we were loosing very significant customers. All other Linux companies received major investments and assistance from IBM except Caldera even though it was primarily Caldera that was driving UnitedLinux. IBM needed UnitedLinux because they did not want a dominate Linux provider but they could not afford to test and support their software on too many independent Linux versions. IBM, because of their interest in UnitedLinux, was a major influence in helping to shape the UnitedLinux relationship. They were not helpful in assisting Caldera in being successful and in fact were a major impediment. There was substantial justification for a suit relating to the Monterey project, but we choose to focus on building a business.
I am not a conspiracy promoter, but we kept hearing rumors that Intel’s and IBM’s motives were not pure in relationship with Monterey or ultimately AIX 5L. They did not want UNIX to succeed on the Intel platform. I do believe IBM’s motives with Linux are business pure. In other words, they are making money on Linux so they genuinely want Linux to succeed. I believe that while they have a good relationship with Intel and Microsoft, they also would like nothing more than to see something at least threaten the dominance of Windows. What IBM did not want was another MicroSoft to deal with. The felt they could manage Linux if it remained even one or two small players. I honestly do not see that same motivation for Intel. What do they really stand to gain with Linux except as a weapon against Sun who remains their only real chip competitor? However, AMD has made some inroads. They had absolutely no desire to really help Linux except as a weapon against UNIX. Consequently, their motives with Linux did not appear altruistic or even pure business related.
So to summarize, while we had gotten Caldera to break even, the investment to get Tarantella off the board had completely drained the cash reserves. Darl needed an investment to provide a cushion and fund new business growth opportunities. He did not come through on his promise of his own investors. Bob Bench and I had been working to attract other investors and had developed a package that both Bob and I had begun to present to the investment community highlighting our strengths. Some of our efforts were beginning to bear fruit but we did not have any major success.
The board had seen the major set backs Caldera had suffered both at the hands of large corporate entities like Novell, Netscape, Corel, Star Division, IBM and Intel as well as the abuse of a well meaning but misinformed Linux community itself. Many of these companies had literally stolen millions of dollars from Caldera only to be held up as icons by the Linux and Open Source communities. Other companies, either intentionally or unintentionally really hurt Caldera by withholding their support both financially and or through supporting Caldera’s Linux version with their hardware and software products, making it extremely hard to meet customer demands. This lack of support was particularly puzzling given their public statements and Caldera’s excellent positioning to help and move Linux forward on a global basis in comparison to the other extremely regional players. The company had spent millions of dollars in supporting Linux only to be ridiculed, brow beaten and falsely accused by the community of trying to take advantage of the Linux developers when we were only trying to make a legitimate business while these other companies, who behind the scene, were stealing money by breaking contracts and at the very least, had very mixed motives for Linux, were heralded as almost heroes by the open source community.
The Taratella influence on the board in support of UNIX came at a critical time. With the economic down turn, it appeared that the Linux bubble could have burst along with the dot coms. At least, I think Alok and Doug Michaels must have felt that way. They had completely missed both the Internet and Linux opportunity and I think they were almost hoping it would fail just to vindicate their poor choices. Our board was somewhat receptive because of all that had transpired. I do not critize the board, but this was a critical time when real leadership was needed. Unfortunately, since I was not the chairman of the board, Canopy was running scared (concerned about minority share holder suits at a time when they were particular vulnerable), and the bottom was falling out of the market, there was no one to stay the course.
When Darl began to get enamored with the heritage of UNIX, he felt he had found something that glittered. He felt he might have something that could either bring IBM or Mircrosoft to the table for very different reasons. Apparently, the tactic worked short term as a major investor came to the table very shortly after the announcement of the suit. There were a number of very legitimate suits that could have been made because of the many injustices that were done, but the path they took was apparently determined by the investor that came to the table. Unfortunately, as is the case with these suits, those closely involved began to get caught up in the frenzy and publicity that occurs as a result of the public attention. They start to believe themselves.
When I heard about the suit against IBM and Linux I immediately sold my shares. When the investment came into Caldera from one of Microsoft’s investment companies and the stock price rose to $17 per share, I had long since cashed out without benefiting from the hype. I absolutely do not regret my decision because I had lived through a suit and knew what it could do to a company and had lost faith in the management’s ability to carry out our vision. Their decision had put a knife through our Linux vision and over 9 years of efforts when they were finally in a position to capitalize on the investment. The management team and board in my mind had gone for the quick pay off rather than opting to build the business. The board and management team literally gave all our Linux work to SuSE and then to Novell. The incredible ride was over but the vision of Caldera lives on in Novell’s core engineering team that was acquired by SuSE to complete United Linux.