Fujitsu Goes All Out on Linux, Investing 6.3 Billion Yuan and 10,000 Engineers (1)
Fujitsu recently made a major decision—a strategic shift in its enterprise server business: it aims to complete development of Linux backbone servers within 2 to 3 years, so that the open-source OS Linux can cover large-scale systems such as those in finance and public infrastructure that currently use large computers (mainframes). To this end, “Fujitsu will mobilize 1,000 developers and 10,000 SEs (system engineers), and invest 100 billion yen over three years (about 6.3 billion RMB),” said Fujitsu Vice President and head of the platform division, Tadayasu Sugita.
Strategic cooperation with U.S. Red Hat and Intel, with the Intel partnership to be announced in January 2003
As “the world’s first vendor to fully use Linux servers in backbone systems” (Tadayasu Sugita), Fujitsu first launched two highly strategic partnerships to achieve the above goal. One was forming the One Source Alliance with U.S. Red Hat. As a Linux distributor, Red Hat controlled 75% of the global market in 2001 (according to IDC in the U.S.). Fujitsu became the second company after Dell Computer in the U.S. to join hands with Red Hat. From Red Hat’s point of view, Dell mainly handles small and medium Linux servers, while Fujitsu is for large Linux servers.
Through this strategic cooperation, Red Hat will support the Linux backbone servers developed by Fujitsu. In this way, besides rapid installation of software around the Linux kernel, support for general software, simpler system management, and reduced costs, it is also expected to increase confidence among Fujitsu’s customers.
As for Linux scalability, the current Linux 2.4 version supports up to 8-way systems. However, Linux founder Linus Torvalds announced on October 24 that Linux 2.6, which supports large SMP (symmetric multiprocessing) systems of 16-way and above, “will be released by June 2003.” In addition, Dr. Irving Wladawsky-Berger, head of IBM U.S.’s “e-Business on Demand” business in Japan last week (formerly head of Linux), also expressed expectations for Linux: “By 2004, the reliability of commercial applications running on Linux servers with 16 processors and above will become higher and higher.”
The second strategic partner is Intel. As Vice President Sugita said, “The distinguishing feature of Linux servers lies in the system architecture, namely the chipset.” Fujitsu has embedded the RAS (reliability/availability/serviceability) functions of mainframes into the chipset of its self-developed Linux backbone servers. At the same time, it also plans to embed Fujitsu-developed RAS functions into IA processor chips and make them an Intel standard. Therefore Fujitsu decided to cooperate with Intel, and it is estimated that the two companies will formally announce the details of the strategic cooperation around January 20, 2003.
The difference is ‘developing pure Linux servers’
When describing Fujitsu’s enthusiasm for Linux, Vice President Sugita said: “IBM, HP, NEC and others are also developing servers—they install Windows and Linux on their servers, whereas Fujitsu is developing ‘pure Linux servers’ that may allow Linux to run with the best performance. That is where we differ from other companies.” This is exactly the purpose behind Fujitsu’s strategic cooperation with Red Hat and Intel. Of course, Fujitsu’s servers can also run Windows, but in terms of primary use, Linux will be given priority.
Fujitsu Goes All Out on Linux, Investing 6.3 Billion Yuan and 10,000 Engineers (2)
In addition, for Linux servers to run in backbone systems, not only must all Fujitsu-developed software such as middleware and applications be completely rewritten to support Linux, but active negotiations must also be carried out with ISVs (independent software vendors) such as Oracle in the U.S. and SAP in Germany. As for why Linux was chosen as the backbone foundation for Fujitsu’s next-generation servers, Vice President Sugita pointed out the main reasons as follows: (1) not dependent on specific IT vendors; (2) functions continue to improve through open-source communities; (3) the fastest technological innovation can be obtained; (4) customer needs can be responded to directly and functions can be strengthened.
As William Zeitler, IBM U.S. senior vice president and head of the server division, said at Linux World in January 2002: “Open-source technology has brought earth-shaking changes to the IT industry. The era in which the IT industry was dominated for 40 years by the proprietary technology power of various companies is coming to an end.” In order to break free from the constraints of proprietary technology, IBM has positioned Linux as the common OS for all servers.
To achieve this goal, IBM announced in the fall of 2000 that, while integrating with eServer, it would invest 1.3 billion dollars over the next several years starting in 2001 in the development, marketing, and strengthening of service structures for Linux hardware and software. The result won by this positive attitude toward Linux is that IBM’s major European and American customers have successively begun adopting Linux. In 2002, Linux-related business even exceeded 2 billion dollars. IBM’s policy is that both Grid Computing and the standard OS of the future information application era will center on Linux.
It can be said that Fujitsu is another company transforming toward Linux after IBM. During this period, IBM, Fujitsu, Hitachi, and NEC reached an agreement to jointly develop technology that can ensure Linux runs stably in mission-critical fields, and they plan to release research results starting in 2003. Current Linux servers are mainly used for Internet-related purposes, while the purpose of the cooperation among these four companies is to make Linux servers usable in backbone systems.
Providing a path of transition from proprietary systems to Linux
Judging from Linux’s own performance, U.S. Gartner predicted that by 2007, Linux server shipments will exceed 10 billion dollars, accounting for about one quarter of all servers. Linux server sales will respectively surpass servers running Sun Microsystems’ UNIX OS Solaris, HP’s HP-UX, and IBM’s AIX. In addition, according to IDC forecasts, whether in the Japanese market or the global market, by 2006 SIAS (Intel standard architecture servers), including both 32-bit and 64-bit, will account for more than half of shipment value.
Facing Linux’s excellent prospects, Fujitsu Vice President Sugita added: “Right now, probably no one can imagine Linux delivering mainframe-class performance, functionality, and reliability. But the probability of Linux being accepted by backbone systems is far higher than at a Las Vegas casino. Since Fujitsu has chosen Linux, it has the responsibility to provide all customers with solutions (for transitioning from proprietary technology to Linux).”
The vice president also stated that Fujitsu will continue to sell Windows and UNIX servers over the next 5 to 10 years. Because this major shift in server strategy may affect sales of existing Solaris servers, Windows servers, and mainframes, some within Fujitsu also believe: “It is still not appropriate to publicly say that Fujitsu’s server strategy is centered on Linux.” Vice President Sugita believes: “However, if we don’t start now, it will be too late by then. I hope this strategic shift can have a positive effect on current business.” In fact, these words were directed even more toward people inside the company.
Fujitsu Goes All Out on Linux, Investing 6.3 Billion Yuan and 10,000 Engineers (3)
Although IBM is actively promoting Linux, it has not had any particular impact on sales of IBM servers running other OSes. In server shipment value, IBM tied HP in the second quarter of 2002, and in the third quarter even surpassed HP by 2.6 percentage points (according to IDC). Perhaps Vice President Sugita deeply hopes that Fujitsu can also learn from IBM’s successful transition experience. Unfortunately, as this vice president worries, an unusually cautious atmosphere is now spreading within Fujitsu.
“Linux is the common property of mankind”
Vice President Sugita also once said, “We strive toward the lofty goal of making Linux the common property of mankind.” In fact, the term “common property of mankind” was something the writer first heard in his early days from Toshio Ikeda, now a legendary figure at Fujitsu and then managing director. When he resolved to develop IBM-compatible processors, Toshio Ikeda once said this: “IBM’s OS is the public property of mankind (Public Domain), and Fujitsu will offer the world processors that let mankind’s common property—the OS—run with the highest efficiency and greatest speed.”
Contrary to Ikeda’s original intentions, seven years after his death (1981), U.S. copyright law was revised, and OSes also became protected intellectual property, resulting in disputes between Fujitsu and IBM. However, in a 1978 survey commissioned by Fujitsu to five famous American research companies led by IDC—“Will the compatibility business equipped with public software—IBM’s OS—still be guaranteed in the future?”—all five companies answered: “It should continue to be guaranteed.” But the subsequent changes were less a trend of the times than an American strategy.
By changing its target OS to Linux, Fujitsu has once again returned to the old path of the “Ikeda strategy.” The writer believes this is precisely the true meaning of this major shift in server strategy. Because Fujitsu believes that perhaps on open-source Linux, the unfortunate disputes of the past will not happen again. In this way, in the arena of “developing servers that make Linux run at the greatest speed,” it can fully bring its experience and wisdom into play and make a major effort.
Can it once again ignite the spark of a corporate culture that “rises up because of an imagined rival”?
Another point is Fujitsu’s corporate culture. After the introduction of the individual evaluation system, there really have been fewer brave people inside Fujitsu who dare to act as before. But later generations have inherited the tradition of “rising up for the sake of an imagined rival.” Perhaps there is some personal fondness for Fujitsu in the writer’s view here, but there is still hope that the current generation of Fujitsu people can accomplish something. In the 1990s, IBM’s power weakened, and Fujitsu lost its target to “catch up with.” But IBM revived through its service business, and after entering the 21st century, it regained a leading position in all IT fields including servers.
Through this strategic server shift, Fujitsu and IBM are now in direct competition over the development of Linux backbone servers, and what remains is how to concentrate resources and determination. As for this talk of returning to the Ikeda strategy and the theory of creating an imagined rival, Vice President Sugita merely smiled and gave no clear response. But recently Vice President Sugita upgraded the Linux server development organization called BCC (Business Critical Computing), established in July, into a 300-person body and began formal development. At the same time, he clearly stated: “This time we’re serious. We’re going to bring Fujitsu’s strength together.”
Core technical personnel were transferred from the mainframe development division to the UNIX server division, and this time even more effort is being made to put core technical personnel into Linux. With Fujitsu’s axis having shifted several times, what will happen this time? That said, back in 1993 when Fujitsu changed the OS of its personal computers to the industry-standard Windows, it was Sugita who was in charge of the PC division. Whenever Fujitsu makes a major decision, Sugita is always involved—perhaps that too is fate.
In this major shift in server strategy, the most important thing is customer response. Vice President Sugita said, “Since we’re going to do it, we will strive to make Fujitsu number one in the world through backbone systems based on Linux,” and it is said that this has already won approval from several major customers in Japan. Fujitsu Managing Executive Director and head of the solutions business headquarters, Hiroya Madarame, clearly stated, “We will soon begin consultations with major customers and decide on trying to recommend Linux to them.” In this way, Fujitsu’s service business side is also giving full support to Linux. (Reporter: Kenichi Kitagawa)
Fujitsu recently made a major decision—a strategic shift in its enterprise server business: it aims to complete development of Linux backbone servers within 2 to 3 years, so that the open-source OS Linux can cover large-scale systems such as those in finance and public infrastructure that currently use large computers (mainframes). To this end, “Fujitsu will mobilize 1,000 developers and 10,000 SEs (system engineers), and invest 100 billion yen over three years (about 6.3 billion RMB),” said Fujitsu Vice President and head of the platform division, Tadayasu Sugita.
Strategic cooperation with U.S. Red Hat and Intel, with the Intel partnership to be announced in January 2003
As “the world’s first vendor to fully use Linux servers in backbone systems” (Tadayasu Sugita), Fujitsu first launched two highly strategic partnerships to achieve the above goal. One was forming the One Source Alliance with U.S. Red Hat. As a Linux distributor, Red Hat controlled 75% of the global market in 2001 (according to IDC in the U.S.). Fujitsu became the second company after Dell Computer in the U.S. to join hands with Red Hat. From Red Hat’s point of view, Dell mainly handles small and medium Linux servers, while Fujitsu is for large Linux servers.
Through this strategic cooperation, Red Hat will support the Linux backbone servers developed by Fujitsu. In this way, besides rapid installation of software around the Linux kernel, support for general software, simpler system management, and reduced costs, it is also expected to increase confidence among Fujitsu’s customers.
As for Linux scalability, the current Linux 2.4 version supports up to 8-way systems. However, Linux founder Linus Torvalds announced on October 24 that Linux 2.6, which supports large SMP (symmetric multiprocessing) systems of 16-way and above, “will be released by June 2003.” In addition, Dr. Irving Wladawsky-Berger, head of IBM U.S.’s “e-Business on Demand” business in Japan last week (formerly head of Linux), also expressed expectations for Linux: “By 2004, the reliability of commercial applications running on Linux servers with 16 processors and above will become higher and higher.”
The second strategic partner is Intel. As Vice President Sugita said, “The distinguishing feature of Linux servers lies in the system architecture, namely the chipset.” Fujitsu has embedded the RAS (reliability/availability/serviceability) functions of mainframes into the chipset of its self-developed Linux backbone servers. At the same time, it also plans to embed Fujitsu-developed RAS functions into IA processor chips and make them an Intel standard. Therefore Fujitsu decided to cooperate with Intel, and it is estimated that the two companies will formally announce the details of the strategic cooperation around January 20, 2003.
The difference is ‘developing pure Linux servers’
When describing Fujitsu’s enthusiasm for Linux, Vice President Sugita said: “IBM, HP, NEC and others are also developing servers—they install Windows and Linux on their servers, whereas Fujitsu is developing ‘pure Linux servers’ that may allow Linux to run with the best performance. That is where we differ from other companies.” This is exactly the purpose behind Fujitsu’s strategic cooperation with Red Hat and Intel. Of course, Fujitsu’s servers can also run Windows, but in terms of primary use, Linux will be given priority.
Fujitsu Goes All Out on Linux, Investing 6.3 Billion Yuan and 10,000 Engineers (2)
In addition, for Linux servers to run in backbone systems, not only must all Fujitsu-developed software such as middleware and applications be completely rewritten to support Linux, but active negotiations must also be carried out with ISVs (independent software vendors) such as Oracle in the U.S. and SAP in Germany. As for why Linux was chosen as the backbone foundation for Fujitsu’s next-generation servers, Vice President Sugita pointed out the main reasons as follows: (1) not dependent on specific IT vendors; (2) functions continue to improve through open-source communities; (3) the fastest technological innovation can be obtained; (4) customer needs can be responded to directly and functions can be strengthened.
As William Zeitler, IBM U.S. senior vice president and head of the server division, said at Linux World in January 2002: “Open-source technology has brought earth-shaking changes to the IT industry. The era in which the IT industry was dominated for 40 years by the proprietary technology power of various companies is coming to an end.” In order to break free from the constraints of proprietary technology, IBM has positioned Linux as the common OS for all servers.
To achieve this goal, IBM announced in the fall of 2000 that, while integrating with eServer, it would invest 1.3 billion dollars over the next several years starting in 2001 in the development, marketing, and strengthening of service structures for Linux hardware and software. The result won by this positive attitude toward Linux is that IBM’s major European and American customers have successively begun adopting Linux. In 2002, Linux-related business even exceeded 2 billion dollars. IBM’s policy is that both Grid Computing and the standard OS of the future information application era will center on Linux.
It can be said that Fujitsu is another company transforming toward Linux after IBM. During this period, IBM, Fujitsu, Hitachi, and NEC reached an agreement to jointly develop technology that can ensure Linux runs stably in mission-critical fields, and they plan to release research results starting in 2003. Current Linux servers are mainly used for Internet-related purposes, while the purpose of the cooperation among these four companies is to make Linux servers usable in backbone systems.
Providing a path of transition from proprietary systems to Linux
Judging from Linux’s own performance, U.S. Gartner predicted that by 2007, Linux server shipments will exceed 10 billion dollars, accounting for about one quarter of all servers. Linux server sales will respectively surpass servers running Sun Microsystems’ UNIX OS Solaris, HP’s HP-UX, and IBM’s AIX. In addition, according to IDC forecasts, whether in the Japanese market or the global market, by 2006 SIAS (Intel standard architecture servers), including both 32-bit and 64-bit, will account for more than half of shipment value.
Facing Linux’s excellent prospects, Fujitsu Vice President Sugita added: “Right now, probably no one can imagine Linux delivering mainframe-class performance, functionality, and reliability. But the probability of Linux being accepted by backbone systems is far higher than at a Las Vegas casino. Since Fujitsu has chosen Linux, it has the responsibility to provide all customers with solutions (for transitioning from proprietary technology to Linux).”
The vice president also stated that Fujitsu will continue to sell Windows and UNIX servers over the next 5 to 10 years. Because this major shift in server strategy may affect sales of existing Solaris servers, Windows servers, and mainframes, some within Fujitsu also believe: “It is still not appropriate to publicly say that Fujitsu’s server strategy is centered on Linux.” Vice President Sugita believes: “However, if we don’t start now, it will be too late by then. I hope this strategic shift can have a positive effect on current business.” In fact, these words were directed even more toward people inside the company.
Fujitsu Goes All Out on Linux, Investing 6.3 Billion Yuan and 10,000 Engineers (3)
Although IBM is actively promoting Linux, it has not had any particular impact on sales of IBM servers running other OSes. In server shipment value, IBM tied HP in the second quarter of 2002, and in the third quarter even surpassed HP by 2.6 percentage points (according to IDC). Perhaps Vice President Sugita deeply hopes that Fujitsu can also learn from IBM’s successful transition experience. Unfortunately, as this vice president worries, an unusually cautious atmosphere is now spreading within Fujitsu.
“Linux is the common property of mankind”
Vice President Sugita also once said, “We strive toward the lofty goal of making Linux the common property of mankind.” In fact, the term “common property of mankind” was something the writer first heard in his early days from Toshio Ikeda, now a legendary figure at Fujitsu and then managing director. When he resolved to develop IBM-compatible processors, Toshio Ikeda once said this: “IBM’s OS is the public property of mankind (Public Domain), and Fujitsu will offer the world processors that let mankind’s common property—the OS—run with the highest efficiency and greatest speed.”
Contrary to Ikeda’s original intentions, seven years after his death (1981), U.S. copyright law was revised, and OSes also became protected intellectual property, resulting in disputes between Fujitsu and IBM. However, in a 1978 survey commissioned by Fujitsu to five famous American research companies led by IDC—“Will the compatibility business equipped with public software—IBM’s OS—still be guaranteed in the future?”—all five companies answered: “It should continue to be guaranteed.” But the subsequent changes were less a trend of the times than an American strategy.
By changing its target OS to Linux, Fujitsu has once again returned to the old path of the “Ikeda strategy.” The writer believes this is precisely the true meaning of this major shift in server strategy. Because Fujitsu believes that perhaps on open-source Linux, the unfortunate disputes of the past will not happen again. In this way, in the arena of “developing servers that make Linux run at the greatest speed,” it can fully bring its experience and wisdom into play and make a major effort.
Can it once again ignite the spark of a corporate culture that “rises up because of an imagined rival”?
Another point is Fujitsu’s corporate culture. After the introduction of the individual evaluation system, there really have been fewer brave people inside Fujitsu who dare to act as before. But later generations have inherited the tradition of “rising up for the sake of an imagined rival.” Perhaps there is some personal fondness for Fujitsu in the writer’s view here, but there is still hope that the current generation of Fujitsu people can accomplish something. In the 1990s, IBM’s power weakened, and Fujitsu lost its target to “catch up with.” But IBM revived through its service business, and after entering the 21st century, it regained a leading position in all IT fields including servers.
Through this strategic server shift, Fujitsu and IBM are now in direct competition over the development of Linux backbone servers, and what remains is how to concentrate resources and determination. As for this talk of returning to the Ikeda strategy and the theory of creating an imagined rival, Vice President Sugita merely smiled and gave no clear response. But recently Vice President Sugita upgraded the Linux server development organization called BCC (Business Critical Computing), established in July, into a 300-person body and began formal development. At the same time, he clearly stated: “This time we’re serious. We’re going to bring Fujitsu’s strength together.”
Core technical personnel were transferred from the mainframe development division to the UNIX server division, and this time even more effort is being made to put core technical personnel into Linux. With Fujitsu’s axis having shifted several times, what will happen this time? That said, back in 1993 when Fujitsu changed the OS of its personal computers to the industry-standard Windows, it was Sugita who was in charge of the PC division. Whenever Fujitsu makes a major decision, Sugita is always involved—perhaps that too is fate.
In this major shift in server strategy, the most important thing is customer response. Vice President Sugita said, “Since we’re going to do it, we will strive to make Fujitsu number one in the world through backbone systems based on Linux,” and it is said that this has already won approval from several major customers in Japan. Fujitsu Managing Executive Director and head of the solutions business headquarters, Hiroya Madarame, clearly stated, “We will soon begin consultations with major customers and decide on trying to recommend Linux to them.” In this way, Fujitsu’s service business side is also giving full support to Linux. (Reporter: Kenichi Kitagawa)
我完全同意设想建立DOS组织“DOS联盟” ,也就是说和Wengier、以及“起步”站长莫老师等DOS战友一起来建立这个“DOS联盟”,以发展我国自主OS(操作系统)的高度去完成我们共同的愿望。
------党委书记
------党委书记

