Slashdot Mirror
IBM VP Talks About Another $1 Billion for Linux Development (Video)
Brad McCredie is an IBM VP, and head of IBM's Power Systems development. (He's also one of the mere few hundred IBM Fellows that have been named in the past 50 years.) He pointed out in his keynote at this year's LinuxCon gathering that IBM has been adopting and supporting Linux (and associated software, like Apache) in various ways for the past decade and a half. Famously, the company promised to support Linux to the tune of a billion dollars in 2001, and McCredie renewed the promise on Tuesday. I sat down to talk with him about just how they'll go about spending the next billion dollars on Linux development; when a company has more than $200 billion in market capitalization, there are lots of ways to spread it around. Spending on hardware is one way, and McCredie also talked about the recently announced OpenPower consortium, which ties directly into the ongoing Linux push.
A billion here, a billion there, and pretty soon you're talking about real money.
The money will be spent on outsourcing it. IBM making this announcement is like Walmart claiming that they will spend 50 million on American businesses. What that will go into is food so that they can take out Kroger and Safeway. As somebody that once worked for IBM, I now say that they are a pure trash company.
IBM gonna make it rain all over GNU!
...from everywhere on the globe where labor is cheap. Of course, you gets what you pays for, but that won't stop Big Blue, no sir!
Why pay your own employees in the server group, you can just ask them all to take a week's leave so you can donate money to Linux....
IBM should use that 1 billion to offer better hardware support and "desktop" linux, moving marketshare into linux is sure to be the best investment if you can gain marketshare which you wont without desktop linux
A lot of it's going to be spent in France, which is famous for non-cheap labor ... (And on things like Watson, which has teams in New York, Austin, and India at least -- I think in some other places, too.)
jrnl: http://tinyurl.com/c2l8yr / foes: http://tinyurl.com/ckjno5
You get what you pay for doesn't apply here at all. First of all, code needs to get accepted into the kernel. If you knew how that works, then you would know that garbage isn't getting into it. The people will have to be highly skilled and know their stuff. Even if they spend ten billion on "cheap labor" then there won't be a single line of code that actually makes it into the kernel.
Guns don't kill people; Physics kills people! - John Lithgow as Dick Solomon on Third Rock From The Sun
There is correctly priced labor, and overpriced labor. And then there are bitter labor that cannot face the competition. I don't post often, here, but honestly, I'm getting a bit tired of whiners with an exaggerated sense of entitlement. Decent programmers still earn enough dollars to buy cheap services from other labor groups that experience a lot more competition than the coders.
The way data centers are going, the "operating system" is migrating off the server farm. The trend is toward servers with a minimal OS that run a single application, probably in a VM. File systems are on other machines on a storage area network. Control of the server farm is on machines separate from the servers. Control machines tell server VMs what to run, what they can connect to (part of "software defined networking"), and their identity for security purposes. Logging, monitoring, and problem reporting is handled by machines other than the workers. Development takes place in a different environment than production. That's what Amazon AWS is like right now. That's what IBM needs to provide to their corporate customers.
Data center computing may not need Linux at all on the production machines. The more specialized machines which run the support systems of the cloud may use it, but they aren't the performance-critical machines. They're the security-critical machines. What we may need is a high-security OS for cloud support machines, accepting some loss in performance.
A lot of it's going to be spent in France, which is famous for non-cheap labor
Unlike US, where experienced programmers are so much cheaper than the French ones.
Ezekiel 23:20
Sorry I don't have enough bandwidth to even watch the adverts on the front without it stopping and starting all the time. I guess it looks fine on the CEOs desk, though, so that's all that matters.
when a company has more than $200 billion in market capitalization, there are lots of ways to spread it around.
What does this even mean? Does Timothy actually think market cap is money that IBM has on hand? How exactly would IBM "spread around" their market cap value?
It's okay. Yes, they've made some fairly incredible contributions to the LINUX ecosystem (LVM comes screaming to mind), but as recent as that was it was still a very different IBM from the one that's doing this now. I have seen the GDF and IBM's current business model up close and personal; I'm just saying I don't trust 'em anymore and I don't perceive that this is going to be nearly as important an event as people here seem to think.
Userspace Code is the Future!
IBM is pretty skilled
IBM stopped making stuff years ago. It buys companies that make stuff. When Brad talks about spending a billion on Linux he means purchasing a billion dollars worth of companies that do stuff in or for Linux.
IBM should just scale back the Power series to the desktop. Just give us some nice RISCs. Now that their AMD acquisitions aren't panning out, they can use AMD's personal to get it done. They can even use the ATI patents to integrate graphics so they really are well positioned for the project.
Or maybe they can get their hands on MIPS. ImgTech doesn't seem to be doing anything useful with them. And coming around from both the low and the high-end will be quite devastating for Intel. Even pick up ImgTech along the ride.
Seriously, just spend it on desktop and mobile R&D. Linux is moving along fine right now and marketing your current line of 30000$ servers won't change anything.
desktop margins are way too low, which is why IBM sold off all that business.
IBM should just scale back the Power series to the desktop.
Apple tried that already, with the PowerPC 970/970MP (which were POWER cores shoved into a desktop chip.) It didn't work. POWER is just too, well, POWER-hungry for the desktop; Apple had to resort to an unreliable liquid cooling system and still couldn't keep pace.
The best thing that IBM could do is make it possible to view Youtube (and other) video under Linux. Adobe ceased support of Flash on Linux quite awhile back. The Linux community responded by saying "Flash will die with Windows!" which, of course, will never, never happen.
Anyway, Linux limps along w/o YouTube/Flash video and sucks, sucks, sucks. Using Linux today is like being a blind man with a white cane. Is this the freedom that the Internet was supposed to give us?
There are people who apparently live in an alternate universe who claim that Linux runs Flash and that they can watch all YouTube videos, but they are full of donkey doo. Until Linux users can view Youtube, life sucks.
Being a Scientologist and driving past an accident is not like being anyone else.
Seriously, two minutes of ads before screening the video? Fuck off
How hard would it be for IBM to port its AIX and OS/2 COBOL compiler to Linux and open source it? A complete, robust, mature COBOL compiler is the one missing piece in Linux. If companies could port legacy code to Linux using an IBM COBOL compiler, they would be able to leave legacy UNIX behind. There are expensive commercial COBOL compilers out there (MicroFocus, Fujitsu) but the cost of these is "enterprise" nutty expensive. A true open-source COBOL compiler would be the final, missing piece. For $1B, IBM could port their COBOL parser to Linux and integrate it into LLVM, and still have a few hundred thousand left over for other things. Could not be all that hard to port their COBOL parser and have it generate LLVM intermediate code - I mean, someone actually implemented C++ 2011, right? Porting a front-end to LLVM couldn't be as hard as writing C++ 2011 from scratch?
Anyhow, people have asked IBM for a Linux COBOL compiler since the 1990s and IBM has ignored them.
Thjis money is dedicated to liniux on power (pSeries hardware) which nobody with half a brain would use. This is just self-aggrandizement and will not benefit the linux community at all. Just another worthless sound byte from a dying dinosaur.
They've exited desktops, but they could bring back Power workstations. But they do have to have a range of CPUs, with power consumption low enough for a tablet right up to the POWER8 line. If they can populate the low end, they could bring back some interesting workstation designs.
I'm not sure if I'm grasping the full significance of this post. But, I like it.