Interesting operating system technology tainted by stolen money

Posted on July 14th, 2009 by bile
Tags: , , , , , , , , , , , , , , 3 Comments »…

Every enterprise wants to harden its servers and increase uptime, but security updates often require reboots. Companies that want to please their customers need a better way to apply software updates. One potential solution for Linux servers is Ksplice, which can seamlessly apply live updates while the system is running.

The underlying technology behind Ksplice is highly sophisticated. To generate a live update, it compares compiled object code from before and after a source patch is applied, a technique that the developers refer to as “pre-post differencing.” They take advantage of the -ffunction-sections and -fdata-sections options of the C compiler to eliminate some variance between the pre- and post-object code.

To determine where the symbols reside in memory, they use a method that they describe as run-pre matching, which compares the “pre” object code to the code that is running in memory. This is done with a special Ksplice kernel module. The live updates generated by Ksplice inject new functions into memory while the kernel is running and modify the old functions so that their path of execution will be redirected to the new versions.

According to a research paper published by the developers, the live update process will disrupt system operation for a mere 0.7 milliseconds. The system state will be left completely intact through the process, meaning that that the overall impact of the live update should not be perceivable.

A majority of kernel security patches can be applied through Ksplice without requiring intervention. Patches that make semantic changes to kernel data structures, however, will need to be accompanied by some custom code to aid the update. In tests, the researchers found that 88 percent of the critical security patches issued for the x86 Linux kernel recently could be applied by Ksplice without requiring additional custom code.

The company announced on Monday that it has received a $100,000 Small Business Innovation Research (SBIR) grant from the National Science Foundation to fund further development of Ksplice technology.

“We think that the Ksplice technology represents an opportunity to finally conquer the software update challenge that exists in every computing system “from the server software stack to communications equipment to storage appliances,” said Ksplice CTO Tim Abbott in a statement. “We are pleased that the National Science Foundation recognizes the potential of this technology and has decided to support our company.”

This technique is hardly new. There have been hot kernel fixes for decades. It is nice to see it come to Linux though. Unfortunately the NSF has decided to spend taxpayer money on this adventure rather then allowing the marketplace decide whether it’s a necessary feature or if it’s implemented well enough.

You can’t have it both ways: FSF not happy with Amazon’s usage of FOSS

Posted on June 19th, 2009 by bile
Tags: , , , , , , , , , , , , , , , , , , , , , , , , , , , 2 Comments »…

As my colleague Michelle Menga is reporting, Amazon is now making new source code available for its Amazon Kindle. Basically what it represents is, Amazon’s responsibility to make the GPL licenced source code that is used in the Kindle available to others.

That’s part of the GPL license and Amazon is doing its part.

Digging into the code that Amazon is now making available, provides some really interesting insight into the underlying structure of the Kindle.

For one, Kindle (at least the DX) is using a modified Linux 2.6.22 kernel. This is a kernel that originally was released by Linus Torvalds in 2007. Is it a surprise that the Kindle is Linux powered? (not really).

Where there is LInux there are always some key Linux tools. In the Kindle’s case that’s the GCC 4.1.2 release for code compilation. In GCC terms that’s now an older release (originally out in 2006), so I would hope that Amazon moves to the newer GCC 4.4 over time as it could yield some performance gains for them.

Amazon is also using BusyBox (how can you not if you’re running embedded?), so it’s a good thing they’ve released that code – BusyBox has been active in recent years by way of the Software Freedom Law Center (SFLC) in making sure that vendors that use their code actually comply with the GPL.

That doesn’t necessarily mean that those that back the GPL are entirely thrilled with Amazon. In fact the Free Software Foundation (FSF), actually refers to the Kindle (somewhat less than politely) as the ‘Swindle’.

“It’s good that Amazon is complying with the licenses and not behaving illegally, but this is hardly something praiseworthy,” John Sullivan operations manager at the FSF blogged. “Amazon benefited from the freedoms passed on to them by other free software authors, and that benefit comes with an obligation to convey that same freedom to their users — to share alike.”

This isn’t about all supporters of FOSS but the many who are anti DRM.

For those of you… you are inconsistent. DRM is based on copyright laws. Intellectual property is both the justification for government enforced DRM (like the DMCA) and free and open source licenses. If you can use the government to force Amazon to abide by the usage rules set by authors of the Kindle’s software then Amazon can use the government to force you to obey the rules regarding the hardware and media they provide you.

People like Richard Stallman don’t understand freedom in a consistent way. They want the ability to do what they like with the physical and digital things in their possession but use the threat of violence to make others unable to do the same. Intellectual property is not actual property and can not be owned. It is inalienable. Nontransferable. Scarcity only applies to it’s ability to be transferred and not itself. To threat or actually aggress against someone in order to keep a monopoly on an idea is just as illegitimate and ridiculous as waging a war on a tactic.

Trying to have it both ways is intellectually dishonest and antithesis to the rule of law or a free society. If you use the guns of government to create this artificial monopoly power you will forever be fighting for control over it.

Dean Baker suggests government spending on FOSS to help stimulate the economy

Posted on January 14th, 2009 by bile
Tags: , , , , , , , , , , , , , , , , , , , , , , 1 Comment »…

5) Funding for Writers/Artists/Creative Workers

In the New Deal there was both a federal arts project and a federal writers project. These programs employed thousands of young artists and writers. A creative stimulus package can extend this idea for the Internet Age. Suppose that President Obama made $10 billion a year available for state and local governments to support various types of creative and artistic work. This could include music, movies, writing books, even journalism. The one condition for support is that all material be made freely available in the public domain. (Better yet, it could have copyleft protection.)

This funding would be sufficient to employ 200,000 people a year at an average of $50,000 each. This would put an enormous amount of creative work in the public domain that people all over the world could download at zero cost. In the first year or two, we could have this program administered through public agencies, but in later years we can have people choose for themselves which work they want to support through a tax credit. The cost would be approximately $10 billion a year.

6) Funding for the Development of Open Software

In the same vein, the government can spend $2 billion a year to develop open source software. This money can be used to further develop and simplify open source operating systems such as Linux, as well other forms of free software. The payoffs from this spending would be enormous. Imagine that every computer buyer in the world would be able to get a computer for which the operating system was free, as was almost all the software that they would ever use.

This would surely save consumers an average of at least $200 per computer. With sales at close to 20 million a year, the savings in the United States alone could easily exceed the cost of supporting software development. Adding in the benefits (and presumably some contributions) from the rest of the world, we will be way ahead by going the route of publicly funded open software open software. The cost would be $2 billion a year.

Slashdot makes mention that he predicted the housing collapse and current economic issues. Well la-de-da. That’s like congratulating a 12 year old for being able to tell the difference between a blue circle and red square. Baker is hardly an economist I have much faith in. His work, The Conservative Nanny State, was an abomination in regards to solutions for proclaimed problems. He sees some of the problems… some of the truths in the modern conservative movement but just wants to rearrange things how he’d like to see them rather then looking for economically viable and efficient solutions. That is supporting private property and the free market.

This story particularly annoys me because being that I’m both an ancap and a big supporter of FOSS this guy was so off in his understanding of the FOSS culture and human tendency in general. In chapter 4, Bill Gates — Welfare Mom, How Government Patent and Copyright Monopolies Enrich the Rich and Distort the Economy, he says:

Copyrights on computer software vastly increase the cost of computers, and make many software applications quite costly, when they could be transferred at no cost over the Internet. Can there be mechanisms for publicly financing software development that allow new software to be distributed cheaply or free? There would be enormous gains to the economy if software were freely available and the price of computers was no longer inflated by royalties on operating systems and applications. The savings from eliminating copyright and patent protection on software would be in the neighborhood of $70 billion to $100 billion a year. In addition, the process of software development would almost certainly be more efficient if all software was placed in the public domain where anyone was free to work on it. As with publicly financed prescription drug research, a large body of public domain software would eliminate the incentive to make duplicative programs and applications that were not qualitatively better than existing programs and applications.

This guy has obviously not looked into FOSS at all. There is more duplication in FOSS then commercial software. Everyone has their own itch to scratch and their itch often isn’t working on some other guy’s code in some other guys coding style in some other language. FOSS generally is less featureful with regards to GUI apps especially with particularly esoteric features.  Even complex applications like operating systems are all over the place. Dozens of mainstream Linux distros, OS X, AIX, OpenVMS, Solaris, Windows, BeOS… all the embedded systems. As for quality… he’s also wrong. Look at AIX, OpenVMS, Solaris, zOS, etc. They are still light years ahead of ground up FOSS. And all many of the major advancements in Linux have come from closed source venders. And most of the major FOSS contributors are employed by software shops that are or were closed. And lets not forget that most of the money spent on software is in fact support and not per unit costs. Companies pay millions for support… not for software. Don’t forget too that the few hundred per home computer you see is the shelf price. They discount heavily for preinstalled systems. A copy of Vista on a $400 desktop is not the $130 shelf price.

And even though there are lots of free (beer and speech) apps which do the same job as many you’d pay for… millions and millions of customers still pick up Apple and Microsoft products.

Clearly he doesn’t understand what he’s talking about.

I have a commentary on The Conservative Nanny State from when this blog first started that I never finished. Perhaps I could dig it up and finish it off.

Big Blue bring Big Brother to Chicago

Posted on September 27th, 2007 by bile
Categories and Tags: Uncategorized, , , , , , , , , , , , , , ,…

A video surveillance program, similar to the one that has proven to be so “effective” in London, is coming to Chicago with the help of Big Brother Blue, IBM. The cameras, which will reportedly cost less than current city-wide surveillance methods, will also be linked to intelligent software.

For instance, the system will be able to send out an alert if it locates a stolen car or even a missing child. It’s too bad this seems just a little too smart, if you ask us. Like, who gets to determine what the network tracks? Sure, right now it’s just Amber Alerts and stolen cars, but what happens when it starts following people who don’t particularly like the current administration or have ideas that are deemed “too” radical? While we love our technology and the feeling of being safe and protected, we also like doing as we please without being watched.

But does it run Linux? Was Apple on to something back in 1984?

Response to a Russel Shaw’s piece on Ron Paul supporters

Posted on August 6th, 2007 by bile
Categories and Tags: Uncategorized, , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,…

I affirm, I really was going to drop this Ron Paul rant I pursued in my earlier post.

But some facts in this morning’s post by my highly respected Washington, D.C.-centric colleague Declan McCullagh entitled Ron Paul: The Internet’s Favorite Candidate” made me realize there are some lessons that still need to be taught to all you Libertarian-leaning fans of the Republican Presidential candidate.

Read More…

Hibernation with kexec in Linux

Posted on July 12th, 2007 by bile
Categories and Tags: Uncategorized, , ,

Offering a potential alternative to the existing suspend and restore implementations in the Linux Kernel, Ying Huang posted a patch utilizing kexec, “kexec based hibernation has some potential advantages over uswsusp and suspend2. ” He listed two such potential advantages, “the hibernation image size can exceed half of memory size easily,” and, “the hibernation image can be written to and read from almost anywhere, such as a USB disk [or] NFS.” He described the feature implemented by his patch as “jumping from a kexeced kernel to the original kernel“, allowing someone to first boot from one kernel, then to kexec another crashdump kernel in reserved memory and run from it for a while, and finally to “jump back” to the original kernel.

Sounds like an interesting way to do hibernation. I’ve not used any of the suspend technologies as I’ve little need for them but I’m hoping to pick up a subnotebook (maybe the Foleo) and in that case I’d want some suspend feature.