Learn about Linux
Download Linux
Get Linux helpSpecial Offers
Feature: Linux
Linux manufacturing a lead in Asia
By on
July 05, 2002 (8:00:00 AM)
Share
Print
Comments
- By Jack Bryar -
What better application for an industrial-strength operating system than heavy industry? It isn't making much news in the West, but Linux is on the verge of becoming the dominant platform at many integrated manufacturing companies in East Asia.
What better application for an industrial-strength operating system than heavy industry? It isn't making much news in the West, but Linux is on the verge of becoming the dominant platform at many integrated manufacturing companies in East Asia.
Most people familiar with the history of computing can cite a variety
of government and industrial applications that drove the early market for
computers. The very earliest applications of all involved the automation of
industrial processes. In fact, the granddaddy of all "data processing"
applications allowed French garment manufacturers to automate the process of
creating patterned sweaters -- back in 1801.
In the 1970s and '80s IT pioneers such as Digital Equipment Corp. spawned an entirely new business focused on industrial automation, and heavy industry grew to become an important market for software developers. Over the last 20 years, American enthusiasm for computer assisted industrial control systems has faded, along with big chunks of the country's traditional manufacturing base.
In the 1970s and '80s IT pioneers such as Digital Equipment Corp. spawned an entirely new business focused on industrial automation, and heavy industry grew to become an important market for software developers. Over the last 20 years, American enthusiasm for computer assisted industrial control systems has faded, along with big chunks of the country's traditional manufacturing base.
However, in the rapidly expanding industrial centers of East Asia,
computer integrated manufacturing is alive and well, and increasingly a
Linux-only environment. As vendors are quickly discovering, if you want to
market computing systems to heavy industry, they'd better be Open Source.
The most recent evidence for this Linux breakthrough is in the Japanese
steel industry. From a global perspective, there is far more steel-making
capacity in the world than the market would justify. The result has been ruinous
hyper-competition. Many North American and European steel makers have been run out of business. Japanese steel makers have survived by a mix of ruthless cost cutting
and by heavy investment in automated technologies.
Steel making is like most other highly integrated manufacturing
processes. Steel plants almost always require the development of highly customized
IT equipment and software. Most have relied on a limited number of
specialty integrators to develop and operate these closed systems. Toshiba and
Hitachi have traditionally won the bulk of these deals. While these business
relationships have produced reliable systems, they have also proved to be extremely
expensive. In the current competitive environment, it is not sustainable.
This past week, Nippon Steel, NKK Corp., and a number of other Japanese
steel firms let it be known that they are requiring vendors to
convert their systems to Linux platforms and to open up their source code. These steel companies are also expanding their vendor lists to include in-house Open Source
developers and IBM.
For example, starting in a few months, NKK will introduce a new
general-purpose computing system that will handle all 118 production lines at its huge
plant in the city of Kawasaki. Based on its internal cost of ownership
study, the company expects to save over $8 million a year in licenses,
support and administration costs. The company's Fukuyama operations are
undergoing a similar changeover.
According to the company, this conversion program began with the
conversion of its galvanized strip steel line back in early 2001. As part of that
conversion, the company dumped its centralized, special-purpose systems and built a
decentralized system of process control computers. NKK used modified white box PCs
running Linux and open software. Its initial studies suggested the firm would
save as much as 40% a year in IT costs. In addition, the company
hoped that it would be able to reuse many system elements as it upgraded. In
the past, the company had to strip out entire systems, forcing periodic
shutdowns and other costly disruptions to manufacturing.
Much of the code has been developed by an in-house team at the
company's KDK subsidiary, and the company hopes to market its expertise to other
steel makers.
Nippon Steel is undergoing a similar conversion. It recently introduced
a Linux-based process control system at its steel bar plant in Hokkaido
and at its blast furnaces in Kimitsu. The company estimates it will save
better than 50% over the proprietary systems it had used in the past.
These deployments presage a major shift in the Asian IT market.
According to the newspaper Nihon Keizai
Shimbum, Japanese manufacturers are rapidly converting to Linux systems and
leading much of the Japanese market in the process. The Linux server market in
Japan has exploded. Compound growth rates have approached 40% a
year since 2000. By 2004, computer makers are projecting Linux will be
on nearly a quarter of all servers in use across the country, and that
figure could prove to be conservative.
The vendor community is getting on board. Recently IBM, Fujitsu, NEC, and Hitachi announced they would be combining resources to develop a Linux distribution optimized for large Japanese corporate systems.
Other Asian manufacturers are jumping on the bandwagon. For
example, Linux is poised to become an important element in the IT strategies
of India's pharmaceuticals business. In China, government -linked
manufacturers are leading a phalanx of other companies and institutions by adopting
Linux or using the threat of Linux to force price wars among closed systems
developers.
Chinese and Hong Kong news outlets have been watching with some
amusement as Microsoft has become embroiled in a ruinous race with Sun
Microsystems to see which company can give away the most software to Chinese
businesses and institutions. Last spring, Sun donated millions of StarSuite
licenses to Chinese schools and business-affiliated "research
institutions." In response, Microsoft announced it would donate what it claimed would
be nearly $750 million worth of product and training services to China's
State Development Planning Commission for distribution to these same
institutes, universities and government-affiliated businesses. Many of these
institutes are little more than dressed-up manufacturing centers.
It all represents a staggering turnaround from the bullying stance
taken by closed system developers toward businesses and governments in the
area, and suggests how quickly Linux has begun to emerge as a viable and
potentially dominant player in the manufacturing sector.
Comments
on Linux manufacturing a lead in AsiaNote: Comments are owned by the poster. We are not responsible for their content.
It's like ecommerce - many of the big, noisy players are having problems, but there are plenty of small, quiet successes. Keep reading NewsForge. We write about successful Open Source companies regularly.
- Robin
- Robin
do you have any list of such small, quiet successes?
I suggest to keep such list on a sidebar, like "most active companies" in financial sites.
If you have a problem with such list content - your readers would be glad to help using some sort of poll application :)
Its a good idea. It would be nice to have that kind of a reference. Most open source companies are not publicly traded, so it might be difficult to include standard financial metrics. But a reference source for successes and failures would be very helpful.
It should be more in-depth that the "most active companies sidebar", though. For example, this was Yahoo Finance's most active sidebar for 2002-07-05:
Nasdaq
Price Change 52-wk Range
SUNW 5.326 +0.436 4.36 - 18.24
CSCO 13.77 +0.73 11.04 - 21.92
INTC 19.25 +1.50 16.26 - 36.78
ORCL 9.951 +0.401 7.251 - 20.00
AMAT 19.86 +1.87 13.295 - 27.95
NYSE
GE 29.05 +1.15 26.90 - 50.20
ELN 2.03 +0.32 1.31 - 62.85
TYC 13.22 +0.56 8.25 - 60.09
AOL 14.46 +0.40 12.04 - 53.30
NOK 15.11 +1.14 11.60 - 27.06
The "most active" lists are eye-candy which are not paricularly helpful for analysis purposes. Tyco?!?! Elan?!?!?! There shouldn't be a plus sign on the same page as these dogs, much less on the same line.
By the way, this snapshot is a good indicator of how proprietary does not necessarily equal profitable ....
RHAT 5.389
SUNW 5.326 ... unless, of course, you want to bring out the big guns...
MSFT 54.03 +1.56 47.50 - 73.15
IBM 73.65 +3.14 66.10 - 126.39
So, I think a new section with a detailed analysis of open source projects, successful and unsuccessful, would be a great addition to the site. One of the success stories could be MySQL AB. For instance, they could tell us how they landed the Yahoo Finance account!
It should be more in-depth that the "most active companies sidebar", though. For example, this was Yahoo Finance's most active sidebar for 2002-07-05:
Nasdaq
Price Change 52-wk Range
SUNW 5.326 +0.436 4.36 - 18.24
CSCO 13.77 +0.73 11.04 - 21.92
INTC 19.25 +1.50 16.26 - 36.78
ORCL 9.951 +0.401 7.251 - 20.00
AMAT 19.86 +1.87 13.295 - 27.95
NYSE
GE 29.05 +1.15 26.90 - 50.20
ELN 2.03 +0.32 1.31 - 62.85
TYC 13.22 +0.56 8.25 - 60.09
AOL 14.46 +0.40 12.04 - 53.30
NOK 15.11 +1.14 11.60 - 27.06
The "most active" lists are eye-candy which are not paricularly helpful for analysis purposes. Tyco?!?! Elan?!?!?! There shouldn't be a plus sign on the same page as these dogs, much less on the same line.
By the way, this snapshot is a good indicator of how proprietary does not necessarily equal profitable ....
RHAT 5.389
SUNW 5.326 ... unless, of course, you want to bring out the big guns...
MSFT 54.03 +1.56 47.50 - 73.15
IBM 73.65 +3.14 66.10 - 126.39
So, I think a new section with a detailed analysis of open source projects, successful and unsuccessful, would be a great addition to the site. One of the success stories could be MySQL AB. For instance, they could tell us how they landed the Yahoo Finance account!
That could be a very interesting section, where technicals (for public ones), fundamentals and press-releases could be gathered together. What would make it different from CNN Money and others? Open Source! Most of Open Source companies are small and cannot afford to buy a traffic to their press releases. Big players don't highlight how open-source help in their business. We need a place where unofficial persons can add links to technicals and fundamentals and press releases.
The problem is that much of linux "news reporting" by linuxtoday, linuxdevices, .... is basically press releases by linux companies. Big vapourware announcements always follow by quiet shutdowns.
Yeah man. How comes that MySQL, primitiveDBMS without ACID is successful, while PostgreSQL, real DBMS beating by features and quality commercial ones, is a looser. Is it because the success is for primitive products (like PHP and MySQL) or the real success of really useful open source products (PGSQL and Python) is hidden in secret? Or is it possible that Microsoft dollars are behind such irrational success specially to highlight the worst part of open source software?
MySQL and PHP are so successful because of ease of use! Also they install on the Microsoft Platform. And in fact work better on MS than MS specific products!
I mention MySQL because I remembered during the post that it was providing the data I was looking at.
PostgreSQL an object-relational DBMS which supports almost all SQL constructs, including subselects, transactions, and user-defined types and functions. MySQL now supports transactions with the Berkely and Inno DB table structures, but you still can't use subselects functions, and user-defined types.
Python is a very clean OOP language with an XML fetish. And I say that in an ultimate compliment kind of way. PHP is a scripting language, although it certainly has support for OOP and XML (through expat or Sablotron).
Why the discrepancy in popularity?
I think the reason that MySQL is used in more instances than PostgresSQL is because of its ease of use. Most sites use a DB for login and registration. PostgreSQL is not necessay for these transactions. However, if you have a more db-intensive site, you will miss subselects and stored procedures (in the same way you would miss your thumbs!).
The same is true of PHP. You can go from HTML to JavaScript to PHP very, very quickly. And then, as you mature as a programmer, PHP will still take care of you. If you move from an OOP environment, however, Python would probably be your first choice.
Is Microsoft secretly funding PHP and MySQL? Without any proof one way or the other, I'm going to say 'no'. I might even go so far as to say you had a legitimate, if oft-repeated, point that you diluted with unsubstantiated, unlikely, irrelevant conspiracy-theorist nonsense.
So lets leave on a combative note, shall we? SAP DB is the most hardcore commercial DB and Ruby is the most elegant open-source OOP language!
PostgreSQL an object-relational DBMS which supports almost all SQL constructs, including subselects, transactions, and user-defined types and functions. MySQL now supports transactions with the Berkely and Inno DB table structures, but you still can't use subselects functions, and user-defined types.
Python is a very clean OOP language with an XML fetish. And I say that in an ultimate compliment kind of way. PHP is a scripting language, although it certainly has support for OOP and XML (through expat or Sablotron).
Why the discrepancy in popularity?
I think the reason that MySQL is used in more instances than PostgresSQL is because of its ease of use. Most sites use a DB for login and registration. PostgreSQL is not necessay for these transactions. However, if you have a more db-intensive site, you will miss subselects and stored procedures (in the same way you would miss your thumbs!).
The same is true of PHP. You can go from HTML to JavaScript to PHP very, very quickly. And then, as you mature as a programmer, PHP will still take care of you. If you move from an OOP environment, however, Python would probably be your first choice.
Is Microsoft secretly funding PHP and MySQL? Without any proof one way or the other, I'm going to say 'no'. I might even go so far as to say you had a legitimate, if oft-repeated, point that you diluted with unsubstantiated, unlikely, irrelevant conspiracy-theorist nonsense.
So lets leave on a combative note, shall we? SAP DB is the most hardcore commercial DB and Ruby is the most elegant open-source OOP language!
MySQL is like an ISAM database with some SQL query support. It is very simple and fast - copy files = copy the database. Most of my databases are 95% read and 5% write. Why get fancy? MySQL smokes out the queries.
I think most software and hardware companies are flirting with bankruptcy. As a matter of fact, there is even a US company in court for having an illegal monopoly that makes the entire software field less competitive and drove proprietary and open souce companies out of business. That company's name escapes me at the moment. However, there are success stories with new companies founded soley for the purpose of selling open source software. There are also success stories of existing companies releasing open source licensed products. But its a tough world and not all the baby turtles make it to the sea.
As far as the open source projects used in this article, once you have a big backer, you are less likely to die. I don't know this for sure, but its likely that these manufacturing companies are running TurboLinux on IBM hardware. As far as the software goes, I imagine that they will be using traditional vendors who have released their product as open source, like SAP did with SAP DB, and they'll be developing in-house products with open-source licenses. There will also be more open source software project to service this market now that there is a market to be serviced.
TurboLinux is a privately-held company, so I don't have access to their numbers. However, in 2000-2001 they went through the same rough period all IT experienced and is still experiencing. They laid off staff, streamlined process and got rid of the original C-level offices. For instance, Ernest Cicogna, former controller for IBM, is their CFO now.
TurboLinux is also good at developing partnerships, like their partnerships with IBM and HP. I'm running the IBM NetVista thin clients, which are TurboLinux powered, in a warehouse setting. I imagine they would make their way into the manufacturing plants as well.
And, of course, TurboLinux is now part of UnitedLinux.
TurboLinux had publicly anticipated profitability by the end of 2001, but I can't find data to confirm or deny.
Big shifts in purchasing procedures, like the one outlined here, will go a long way towards keeping many open source companies in the black.
As far as the open source projects used in this article, once you have a big backer, you are less likely to die. I don't know this for sure, but its likely that these manufacturing companies are running TurboLinux on IBM hardware. As far as the software goes, I imagine that they will be using traditional vendors who have released their product as open source, like SAP did with SAP DB, and they'll be developing in-house products with open-source licenses. There will also be more open source software project to service this market now that there is a market to be serviced.
TurboLinux is a privately-held company, so I don't have access to their numbers. However, in 2000-2001 they went through the same rough period all IT experienced and is still experiencing. They laid off staff, streamlined process and got rid of the original C-level offices. For instance, Ernest Cicogna, former controller for IBM, is their CFO now.
TurboLinux is also good at developing partnerships, like their partnerships with IBM and HP. I'm running the IBM NetVista thin clients, which are TurboLinux powered, in a warehouse setting. I imagine they would make their way into the manufacturing plants as well.
And, of course, TurboLinux is now part of UnitedLinux.
TurboLinux had publicly anticipated profitability by the end of 2001, but I can't find data to confirm or deny.
Big shifts in purchasing procedures, like the one outlined here, will go a long way towards keeping many open source companies in the black.
Is there any light in the tunnel for companies developing open source software or are most of them still heading towards bancruptsy?
*********
Most of them are and have been doing quite well. It's the large IPOs that are having problems. The large IPOs (Red Hat, LNUX, etc) tried to do Linux like you do proprietary software. Even though they said "services", if you look at their business model, they were way too focused on applications. The ones doing well are the smaller companies developing customer solutions for customers. They are usually smaller companies (i.e. - not publicly traded), but doing well.
*********
Most of them are and have been doing quite well. It's the large IPOs that are having problems. The large IPOs (Red Hat, LNUX, etc) tried to do Linux like you do proprietary software. Even though they said "services", if you look at their business model, they were way too focused on applications. The ones doing well are the smaller companies developing customer solutions for customers. They are usually smaller companies (i.e. - not publicly traded), but doing well.
I can't understand why you listed Red Hat.
The companies that tried to create proprietary lock-in, and are now having trouble, are the members of United Linux, i.e. SuSE, Caldera, and TurboLinux.
Red Hat, on the other hand, is as Open Source as they come -- even their install procedure is Open Source.
And, Red Hat is doing quite well, in financial terms (especially in light of the recent economic slowdown that has affected so many other companies). I would venture to guess that much of their success comes from the trust generated by Red Hat's dedication to Open Source.
In fact, contrary to your claim, Red Hat's largest growth area has been in new _support_ contracts with large companies, such as Oracle.
The companies that tried to create proprietary lock-in, and are now having trouble, are the members of United Linux, i.e. SuSE, Caldera, and TurboLinux.
Red Hat, on the other hand, is as Open Source as they come -- even their install procedure is Open Source.
And, Red Hat is doing quite well, in financial terms (especially in light of the recent economic slowdown that has affected so many other companies). I would venture to guess that much of their success comes from the trust generated by Red Hat's dedication to Open Source.
In fact, contrary to your claim, Red Hat's largest growth area has been in new _support_ contracts with large companies, such as Oracle.
My company makes plenty open source software for industrial control and we are doing fine. Of course we sell the rest of the industrial sensing and control system also. Open source helps us make the sale because it helps convince the customer that the product can be supported many years down the line. Most industrial systems have lifetimes in excess of ten years, so support and maintenance are serious issues - maybe THE ISSUE. Hence the growth of open source in this sector.
The problem is that industrial systems don't change much during its lifetime. And finding people in the open source community to support linux 2.2 ten years from now is extremely hard.
And exactly why would it be hard? The code will still all be there. All writen to posix standards.
You are a pathetic worthless troll.
Linux is no where near posix standards --- even monta vista whitepapers peg linux as 30-40% posix compliant and that's from a linux company (so they probably fudged the numbers on the high side).
Linux today is still pretty close to BSD Unix from 25 years ago - all the old commands still work! 25 years from now, I expect Linux (God knows what version!) will still run nearly all software from the Linux 2.2 era. If not, the source is available, and some minor edits will usually fix the problem. Industrial controller designers are a pretty competent bunch.
I don't normally feed trolls, but ah well...
So how easy do you think it will be to find support for Win2k in ten years time - maybe you are hoping that MS will still be supporting it? And how many proprietary software companies last ten years exactly? With OS, at least the company can *hire* people to change the source if needs be. With closed source, they are totally stuck.
So how easy do you think it will be to find support for Win2k in ten years time - maybe you are hoping that MS will still be supporting it? And how many proprietary software companies last ten years exactly? With OS, at least the company can *hire* people to change the source if needs be. With closed source, they are totally stuck.
> And finding people in the open source community to support linux 2.2 ten years from now is extremely hard.
Well, since the source code will be available, I would say that it will be much easier than it is to find support today for Windows 2.0, Windows/286, Windows/386, Windows 3.0, Windows 3.1, Windows for Workgroups, Windows NT 3.1, or any of the multiple early variations of Windows CE.
Note that each of those versions of Windows involved API changes for the application programmers.
Note, also, that Microsoft has already dropped support for Windows 95, with support for Windows 98 and NT4 to be dropped next year:
http://www.smallbiztechtalk.com/news/archives/tips 061801-bn1.htm
> "While it may be old news to corporate IS managers, small businesses are just waking up to a harsh reality in the apparently one-horse-town world of operating systems. Microsoft is once again dangling the veiled threat of limited and withdrawn support to "encourage" (i.e. intimidate) customers into moving to either the Microsoft Windows 2000 or the soon-to-be-released Microsoft Windows XP OS platform."
Plus, if your applications are Open Source, it is very likely that you will still be able to run them on newer versions of Linux with just a recompile, or an easy porting job. Odds are that someone will have alredy done this, but, if not, you can do it yourself.
Compare that to the Windows case, where your old software is simply lost forever, because the company that wrote it is gone, or chose not to port it to newer versions of Windows.
Well, since the source code will be available, I would say that it will be much easier than it is to find support today for Windows 2.0, Windows/286, Windows/386, Windows 3.0, Windows 3.1, Windows for Workgroups, Windows NT 3.1, or any of the multiple early variations of Windows CE.
Note that each of those versions of Windows involved API changes for the application programmers.
Note, also, that Microsoft has already dropped support for Windows 95, with support for Windows 98 and NT4 to be dropped next year:
http://www.smallbiztechtalk.com/news/archives/tips 061801-bn1.htm
> "While it may be old news to corporate IS managers, small businesses are just waking up to a harsh reality in the apparently one-horse-town world of operating systems. Microsoft is once again dangling the veiled threat of limited and withdrawn support to "encourage" (i.e. intimidate) customers into moving to either the Microsoft Windows 2000 or the soon-to-be-released Microsoft Windows XP OS platform."
Plus, if your applications are Open Source, it is very likely that you will still be able to run them on newer versions of Linux with just a recompile, or an easy porting job. Odds are that someone will have alredy done this, but, if not, you can do it yourself.
Compare that to the Windows case, where your old software is simply lost forever, because the company that wrote it is gone, or chose not to port it to newer versions of Windows.
But if you stick with a POSIX certified OS in the first place, then if the first OS company goes bankrupt --- you just recompile your app for a second Posix certified OS.
You don't even have to hire people to maintain your branch of linux 2.2 ten years from now.
You don't even have to hire people to maintain your branch of linux 2.2 ten years from now.
A Linux tsunami is on the verge of happening.
Yup, it's been on the verge of happening for
5 years. It's not a tsunami, but free software
is becomming the dominate software, as it should be.
5 years. It's not a tsunami, but free software
is becomming the dominate software, as it should be.
Another excellent article from Jack! Let's all vote Jack as the top newforge writer.
Only thing missing from this article would be an analysis of the ailing US steel industry's use (or lack of use) of Linux.
Also, what does this mean for other industries? What other US industries are losing the competitive advantage by wasting money on proprietary IT?
A political tangent:
We have President Bush recently coming to the aid of the failing US Steel with the imposition of tariffs. This is the same guy who called the dogs off of Microsoft. This is the same guy who has made no efforts to support OSS and reduce governmental IT spending.
You can find links to the some software used in the steel industry through the US Dept of Energy's Office of Industrial Technologies website.
http://www.oit.doe.gov/steel/steel.shtml
http://www.oit.doe.gov/steel/steel.shtml
>>>Only thing missing from this article would be an analysis of the ailing US steel industry's use (or lack of use) of Linux.
The analysis will show that it means nothing in the big picture for an alleged 8 million dollar saving for a company with an 10 billion dollar annual revenue.
The analysis will show that it means nothing in the big picture for an alleged 8 million dollar saving for a company with an 10 billion dollar annual revenue.
Actually it will show that the manager who implemented the switch got a big promotion.
Its small minded peons like yourself who are responsible for bloated US industry. That's how you increase competiveness, 8 million here, 10 million there, it all adds up.
Add up 8 million across 20 divisions in different areas of that company's IT spending, and then you will see the real rewards.
No, it's peons like you that didn't even read the article. NKK is saying even if they deploy linux enterprise wide, the saving would be 8 million dollars --- there is no 8 million here and 10 million there.
wrong you dipsh*t.
its one article about one use of linux. you think they are required to mention every use of linux throughout the company in one article?
Go and read the nippon steel annual report --- only 1 blast furnace (the number 3 furnace at kimitsu is linux based).
>>>its one article about one use of linux. you think they are required to mention every use of linux throughout the company in one article?
Precisely --- they don't mention that the rest of the world is run by unix'es and qnx'es. That's why you don't know that your electricity comes from a QNX-based nuclear power station, the tunnel/bridge toll booths in New York city runs QNX, the traffic lights are run by QNX, the subway lines are run by QNX, the radio stations are hooked up with satellites run by QNX, the television studios uses QNX systems for automations...
One article about one company using linux on their blast furnace ain't mean anything at all to the rest of the military industrial complex.
>>>its one article about one use of linux. you think they are required to mention every use of linux throughout the company in one article?
Precisely --- they don't mention that the rest of the world is run by unix'es and qnx'es. That's why you don't know that your electricity comes from a QNX-based nuclear power station, the tunnel/bridge toll booths in New York city runs QNX, the traffic lights are run by QNX, the subway lines are run by QNX, the radio stations are hooked up with satellites run by QNX, the television studios uses QNX systems for automations...
One article about one company using linux on their blast furnace ain't mean anything at all to the rest of the military industrial complex.
> The analysis will show that it means nothing in the big picture for an alleged 8 million dollar saving for a company with an 10 billion dollar annual revenue.
So, you add the word "alleged", and compare the savings to the revenue of the entire company, and the savings look small.
Nice try, but you are fibbing, in more ways than one.
First, you should really be comparing the savings to the company's profits, not revenue. In a heavy industry like steel, most of the revenue goes to covering the costs of raw materials.
If we assume that NKK's profit is 10 percent of revenue (and that's being very generous in an industry like steel), then your $10 billion revenue becomes $1 billion profit.
Now, we see that the $8 million in savings represents a 1 percent increase in profits. That's looking better.
But wait! Let's go back to the original article, and see what it actually said:
> For example, starting in a few months, NKK will introduce a new general-purpose computing system that will handle all 118 production lines AT ITS HUGE PLANT IN THE CITY OF KAWASAKI. Based on its internal cost of ownership study, the company expects to save over $8 million a year in licenses, support and administration costs. [emphasis added]
Yes, contrary to the implication in your post, this is the savings for only one plant, in one city. Thus, your comparison of the savings to the revenue OF THE ENTIRE COMPANY is even less accurate.
The article says there are more savings to come:
> The company's Fukuyama operations are undergoing a similar changeover.
Plus, as the article states, those are only the direct, easily calculated savings for "licenses, support and administration." They didn't mention the other likely benefits, arising from increased reliability, greater flexibility, faster deployment, the ability to run the same software across multiple platforms, and so on.
But the best line in the article is the one you were hoping your distraction would keep us from noticing:
> Its initial studies suggested the firm would save as much as 40% a year in IT costs.
Yes, a 40 percent savings in IT costs.
***FORTY*** PERCENT SAVINGS!
JUST FROM USING LINUX!
If every department in the company could find a way to reduce costs by 40 percent, some might think it a good thing. But you would advise against it, of course, since, in each case, its only one department, and the savings are small compared to the revenue of the entire company.
So, you add the word "alleged", and compare the savings to the revenue of the entire company, and the savings look small.
Nice try, but you are fibbing, in more ways than one.
First, you should really be comparing the savings to the company's profits, not revenue. In a heavy industry like steel, most of the revenue goes to covering the costs of raw materials.
If we assume that NKK's profit is 10 percent of revenue (and that's being very generous in an industry like steel), then your $10 billion revenue becomes $1 billion profit.
Now, we see that the $8 million in savings represents a 1 percent increase in profits. That's looking better.
But wait! Let's go back to the original article, and see what it actually said:
> For example, starting in a few months, NKK will introduce a new general-purpose computing system that will handle all 118 production lines AT ITS HUGE PLANT IN THE CITY OF KAWASAKI. Based on its internal cost of ownership study, the company expects to save over $8 million a year in licenses, support and administration costs. [emphasis added]
Yes, contrary to the implication in your post, this is the savings for only one plant, in one city. Thus, your comparison of the savings to the revenue OF THE ENTIRE COMPANY is even less accurate.
The article says there are more savings to come:
> The company's Fukuyama operations are undergoing a similar changeover.
Plus, as the article states, those are only the direct, easily calculated savings for "licenses, support and administration." They didn't mention the other likely benefits, arising from increased reliability, greater flexibility, faster deployment, the ability to run the same software across multiple platforms, and so on.
But the best line in the article is the one you were hoping your distraction would keep us from noticing:
> Its initial studies suggested the firm would save as much as 40% a year in IT costs.
Yes, a 40 percent savings in IT costs.
***FORTY*** PERCENT SAVINGS!
JUST FROM USING LINUX!
If every department in the company could find a way to reduce costs by 40 percent, some might think it a good thing. But you would advise against it, of course, since, in each case, its only one department, and the savings are small compared to the revenue of the entire company.
I used the word "alleged" because of enron type fudging of numbers. The numbers comes from press releases and unaudited sources.
There are better ways to get increased reliability, greater flexibility, faster deployment on multiple platforms --- just write your apps to be fully posix compliant on a bunch of safety/mission critical POSIX certified RTOS.
I don't question that there might be huge savings by deploying linux. But the savings comes from the IT side, switching servers from unix to linux, running linux/beowulf clusters instead of buying a cray supercomputer for thermal/stress analysis.
What I question is the extent of linux deployment on the actual factory floors and the extent Microsoft is getting hurt by it. It's like amazon.com switching from unix to linux and saving 17 million dollars, it hurt SUN, not microsoft.
There are better ways to get increased reliability, greater flexibility, faster deployment on multiple platforms --- just write your apps to be fully posix compliant on a bunch of safety/mission critical POSIX certified RTOS.
I don't question that there might be huge savings by deploying linux. But the savings comes from the IT side, switching servers from unix to linux, running linux/beowulf clusters instead of buying a cray supercomputer for thermal/stress analysis.
What I question is the extent of linux deployment on the actual factory floors and the extent Microsoft is getting hurt by it. It's like amazon.com switching from unix to linux and saving 17 million dollars, it hurt SUN, not microsoft.
What I question is the extent of linux deployment on the actual factory floors and the extent Microsoft is getting hurt by it. It's like amazon.com switching from unix to linux and saving 17 million dollars, it hurt SUN, not microsoft.[Emphasis added]
Microsoft?
Nobody mentioned Microsoft, Troll!
Microsoft?
Nobody mentioned Microsoft, Troll!
>>>Nobody mentioned Microsoft, Troll!
Microsoft is mentioned in the original article. Please read the actual article before commenting.
Microsoft is mentioned in the original article. Please read the actual article before commenting.
> It's like amazon.com switching from unix to linux and saving 17 million dollars, it hurt SUN, not microsoft.
Amazon replaced SUN servers with Intel boxes.
To run the Intel boxes they chose Linux rather than Windows.
I'd say it hurt both SUN and Microsoft.
Plus, those millions of Intel-based web servers that are running Linux or BSD, instead of Windows, have cost Microsoft a few $billion in sales.
Amazon replaced SUN servers with Intel boxes.
To run the Intel boxes they chose Linux rather than Windows.
I'd say it hurt both SUN and Microsoft.
Plus, those millions of Intel-based web servers that are running Linux or BSD, instead of Windows, have cost Microsoft a few $billion in sales.
NKK's annual revenue is something like 10 billion dollars, they spent more on stamps than the alleged 8 million dollar savings by switching to linux.
It's like people keep on mentioning that amazon.com made money for the first time ever by saving 17 million dollar because they switch to linux (from unix) --- for a company that has billion+ dollar revenue. Guess what, amazon.com lost money again in the latest quarter.
It's like people keep on mentioning that amazon.com made money for the first time ever by saving 17 million dollar because they switch to linux (from unix) --- for a company that has billion+ dollar revenue. Guess what, amazon.com lost money again in the latest quarter.
These companies have a choice:
(1) Not convert to Linux spend millions
(2) Convert to Linux save millions
Why exactly should they choose (1)?
No serious commentators on the industry suggested that Amazon made money by switching to Linux. Whose arse did you pull that out of? And how is it relevant?
The actual cost of the software licensing fees is very much irrevelent in safety critical and industrial manufacturing. Much of these systems are very high priced unix'es and qnx'es.
Why?
Because uptimes for heavy duty unix'es and qnx'es are measured in years, and linux is not at that level yet. Once you turn off the blast furnace, it takes several days to get it back up to steel melting temperatures. Meanwhile you got a few hundred union workers getting paid 25 dollars an hour for just sitting around and doing nothing.
Why?
Because uptimes for heavy duty unix'es and qnx'es are measured in years, and linux is not at that level yet. Once you turn off the blast furnace, it takes several days to get it back up to steel melting temperatures. Meanwhile you got a few hundred union workers getting paid 25 dollars an hour for just sitting around and doing nothing.
Very informative.
I suggest you let the Japanese stell companies know about this. Once they hear your valuable opinion, I'm sure that they will stop their Linux deployments.
On the other hand, they might know you for the dip sh*t troll that you are.
The japanese companies do know about this. That's why only 1 company and 1 blast furnace is linux-based. Most of the linux systems in steel companies are linux clusters performing thermal and stress analysis. Even in this newsforge article, the NKK is said to be introducing a GENERAL purpose computing system --- not exactly in the realms of RTOS controlled safety critical systems.
Linux typically compares itself to MS and operates in a similar space. You're way out of your league here. Believe me when I tell you that you don't want to start calling people who tout the stability of QNX and zSeries running AIX 'dips**t trolls'.
The remarkable part of this article is not that there is a huge migration from the steel community from QNX to Linux because of blue screens of death, downtimes and viruses. There isn't a migration bcs there isn't a problem. That Linux would even be considered an alternative to QNX should be considered flattering.
You would do yourself a favor by looking into QNX, which is an OS you probably aren't familiar with. Look at where its deployed and the metrics that accompany its stability. Linux is a very stable desktop system, particularly when not running a DE. Its not as stable as BSD, but more stable than most MS offerings. But QNX doesn't crash. AS/400s don't crash. zSeries don't crash. Not 'rarely crash', not 'have been running for 320 days', not 'I resart X and everything's fine'. But, 'put it on a sattelite and shoot it to Europa because this thing won't crash until the sattelite does' stable.
And licensing issues are truly the least of your worries in a nuclear reactor, waste management system, sattelites, medical equipment, etc.
Nothing in the prior post was innacurate, much less 'dips**t trolling'.
The remarkable part of this article is not that there is a huge migration from the steel community from QNX to Linux because of blue screens of death, downtimes and viruses. There isn't a migration bcs there isn't a problem. That Linux would even be considered an alternative to QNX should be considered flattering.
You would do yourself a favor by looking into QNX, which is an OS you probably aren't familiar with. Look at where its deployed and the metrics that accompany its stability. Linux is a very stable desktop system, particularly when not running a DE. Its not as stable as BSD, but more stable than most MS offerings. But QNX doesn't crash. AS/400s don't crash. zSeries don't crash. Not 'rarely crash', not 'have been running for 320 days', not 'I resart X and everything's fine'. But, 'put it on a sattelite and shoot it to Europa because this thing won't crash until the sattelite does' stable.
And licensing issues are truly the least of your worries in a nuclear reactor, waste management system, sattelites, medical equipment, etc.
Nothing in the prior post was innacurate, much less 'dips**t trolling'.
> That Linux would even be considered an alternative to QNX should be considered flattering.
What the heck are you talking about? The article doesn't even mention QNX, nor is it talking about the types of applications that require QNX.
Having said that, I wouldn't underestimate the reliability that is possible with Linux. I ran Debian Stable with IceWM, on my desktop, for two years without a single crash of the OS or the GUI. Now obviously I was lucky where my PC hardware was concerned, and I would not suggest using Linux to control a nuclear reactor. But still, Linux is very reliable, and can provide an inexpensive way to automate many functions in the factory.
What the heck are you talking about? The article doesn't even mention QNX, nor is it talking about the types of applications that require QNX.
Having said that, I wouldn't underestimate the reliability that is possible with Linux. I ran Debian Stable with IceWM, on my desktop, for two years without a single crash of the OS or the GUI. Now obviously I was lucky where my PC hardware was concerned, and I would not suggest using Linux to control a nuclear reactor. But still, Linux is very reliable, and can provide an inexpensive way to automate many functions in the factory.
you dont turn off a blast furnace, once you do you have to reline it. The reliability of Unix v linux is a myth, i use Solaris and HP every day, in extreamly heavy overload, like a database, yes Unix might win, but in reasonable duty use Linux is little if any worse. Simple answer is the industrial process is a constant one so you can size your hardware to meet it, while database is peeky. Stick to admin not process control which you obviously know diddly squat about.
regards
Thing
regards
Thing
Companies act sometime to protect themselves, even if this yields no return.
Having the source for their applications grants them additional safety against a provider closing or worst.
Of course, they probably already have the source for their critical apps (which probably is not open). Nonetheless this has other implications...
If I´m not wrong, NKK is the owner of the prestigious Kawasaki motorcycles -- I once heard they did them just to reap marketing recoinassance.
Having the source for their applications grants them additional safety against a provider closing or worst.
Of course, they probably already have the source for their critical apps (which probably is not open). Nonetheless this has other implications...
If I´m not wrong, NKK is the owner of the prestigious Kawasaki motorcycles -- I once heard they did them just to reap marketing recoinassance.
Re:Not everything is money, there´s power, too
Posted by: Anonymous Coward on July 06, 2002 12:42 PM
That's why they have source code escrow services.
Re:Not everything is money, there´s power, too
Posted by: Anonymous Coward on July 08, 2002 10:40 AM
> That's why they have source code escrow services.
Oh good. So now that Microsoft has stopped supporting Windows 95, my company can simply ask for a copy of the code, and change it to suit our needs? Sure, and pigs can fly.
Putting code into escrow will protect you when an honest supplier goes out of business.
But it won't protect you from lock-in and extortion (forced upgrades, and other manipulation) by a dishonest company like Microsoft.
As the original poster suggests, it's much better to be in control of your own destiny, and Open Source gives you that.
Oh good. So now that Microsoft has stopped supporting Windows 95, my company can simply ask for a copy of the code, and change it to suit our needs? Sure, and pigs can fly.
Putting code into escrow will protect you when an honest supplier goes out of business.
But it won't protect you from lock-in and extortion (forced upgrades, and other manipulation) by a dishonest company like Microsoft.
As the original poster suggests, it's much better to be in control of your own destiny, and Open Source gives you that.
Re:Not everything is money, there´s power, too
Posted by: Anonymous Coward on July 08, 2002 12:58 PM
No, writing your apps to be fully posix compliant gives your better control to your own destiny. If the first OS company raises the price, then you switch to another OS.
> NKK used modified white box PCs running Linux and open software. Its initial studies suggested the firm would save as much as 40% a year in IT costs.
Forty percent of total IT costs? Per year?
Now that's impressive. It's enough to make other companies take notice, and perform studies of their own.
Forty percent of total IT costs? Per year?
Now that's impressive. It's enough to make other companies take notice, and perform studies of their own.
Was just wondering.
Posted by: Anonymous Coward on July 05, 2002 07:24 PM#