LQ Radio ogg RSS Feed

A quick update on the LQ Radio ogg RSS feed. From day one we have supported ogg at LQ Radio. In fact, every audio release we've ever done has been released in both mp3 and ogg format. That being said, the RSS feed for podcatchers only offered the enclosure in mp3 format. About two weeks ago we started offering an ogg feed in addition to the mp3 feed. In that two weeks, a full 5% of subscribers have moved over. If you're subscribed to the mp3 feed and are able to use ogg, I'd encourage you to move over to the unencumbered format at your leisure. To clarify, we are committed to supporting both formats long term.
–jeremy
, , , , ,

Interview: Jens Axboe

A fantastic in-depth interview with Jens Axboe has been posted over at KernelTrap. If you're interested in block layer stuff, schedulers or the kernel development process in general, this is a must read. If you don't learn something from each and every KernelTrap interview that Jeremy posts, you are either a serious kernel ninja or Linus.
–jeremy

MySQL AB – a new product and an IPO

Speaking of Open Source companies that understand their market and core strengths, MySQL AB has also been on a tear. Earlier this week, the company announced its MySQL Enterprise Unlimited program. From the press release:
MySQL AB, developer of the world's most popular open source database, today unveiled a simpler way for large and growing organizations to acquire and adopt enterprise software. Designed with a customer’s perspective in mind, a one-year MySQL Enterprise Unlimited subscription offers a company-wide enterprise site agreement at the unprecedented low price of $40,000 (EUR 32,000, GBP 24,000).
“MySQL Enterprise has made it significantly easier to purchase database software and technical support for our entire organization,” said Glenn Bergeron, systems manager for Instaclick Inc., one of the first companies to take advantage of MySQL Enterprise Unlimited. “This new offering is ideal for corporate IT organizations with a growing number of projects but a tightly-fixed budget.”
MySQL Enterprise Unlimited is designed for companies with existing site licenses for Oracle, Microsoft SQL Server, Sybase and IBM DB2. Last year, a survey of the Independent Oracle User Group showed that a full third of its membership also used MySQL1. With a MySQL Enterprise Unlimited subscription, an organization can develop, manage and fully support any number of MySQL database applications — significantly reducing IT time, cost and risk.
“Due in large part to advantages in distribution and volume, open source has the ability to disrupt traditional enterprise software pricing,” said Stephen O'Grady, principal analyst for RedMonk. “MySQL is attempting to prove as much with its latest site wide agreements, which offer customers the ability to support every database across their enterprise at a fraction of the traditional cost.”

Make no mistake about it, MySQL AB could have actually raised the price of their main product and still sold more. Instead they have chosen a plan which I think has a higher upside potential long term. Matt covers some of the reason the company is able to do this:
Zack told me over email (in response to my question, “How can you possibly make money at that price?):
The reason we can make money is because:
-The software really works
-And we don't have expensive Armani-wearing Ferrari-driving sales reps closing 40K deals.

Expanding on that is a post on Zack's blog. This isn't just about Oracle, it's about disrupting the industry. That's why this is a long term play. It's game changing. A lot of proprietary companies may have raised their rates as mentioned earlier, to boost the bottom line in the short term. It all looks very nice on the balance sheet. MySQL represents what I consider the future. Vibrant sustainable companies that treat their customers like partners and not like prey.
I see that the company is now looking at an IPO. As you may have guessed by my previous comments, I'll certainly try to get in on that. As you may have also guessed, I also regularly recommend and implement MySQL and run it on a site that gets a considerable amount of traffic.
–jeremy

Red Hat's Volley on Linux Management Offering

A recent article on InternetNews.com points to some moves by Red Hat that could have a large impact on Open Source in the Enterprise. From the article:
The leading Linux distributor told internetnews.com that it is rolling out online monitoring tool this year Red Hat customers.
Rich Friedman, director of product management for the Red Hat Network, told internetnews.com the management tool is already in use with a select group of customers with expanded rollouts planned throughout the year.
“We provide a set of services that can monitor systems and we handle the background work of collecting metrics, analyzing them, sending out availability and performance alerts,” Friedman said. “We provide reporting charting and history and even more capabilities in terms of managing how you set up alerts and it's all from an online perspective.”
“What we'll be doing with RHN 5 is providing support and capabilities around virtualization management,” Friedman said. [This includes] “focusing on what we do from a Red Hat RHEL perspective and the capabilities we have and the ability to control the system from a central place.”
Red Hat is taking a different angle than a one-console approach to managing all the enterprise components. “One console to serve all purposes is not as important as managing RHEL with the best capabilities,” Friedman said. “That's where we focus.”
And it's not just Oracle with which Red Hat wants to contrast its approach. Freidman said Red Hat's goal is to integrate with the big systems management vendors and work with them in a complementary manner.
Though Red Hat's goal is not be like the big systems management vendors, it is pursuing an effort to open up RHN so that it can manage more applications.

By opening up the infrastructure like this they will be allowing other companies to use it as a framework to build all kinds of compelling applications that can be brought to market much faster and for much lower costs. All the while they'll be cementing themselves as a core piece of the puzzle. This sums it up nicely:
Friedman explained that Red Hat is currently working on a platform for systems management that open source projects will be able to leverage to provide systems management out of the box. “Not to be an OpenView (HP's enterprise management product) or any one else in that arena but really as a way that you can integrate a base platform, and then provide a plug-in that will mange your product set,” Friedman said.
It's that kind of focus along with an understanding of their market and core strengths that allows Red Hat to do so well. I for one hope they are able to continue to execute as they have been. With moves like this, you have to think they well. Well done.
–jeremy

It's all in the wording

I wonder if any hardware manufacturer will take Greg up on his offer of Free Linux Driver Development. From the post:
Yes, that's right, the Linux kernel community is offering all companies free Linux driver development. No longer do you have to suffer through all of the different examples in the Linux Device Driver Kit, or pick through the thousands of example drivers in the Linux kernel source tree trying to determine which one is the closest to what you need to do.
All that is needed is some kind of specification that describes how your device works, or the email address of an engineer that is willing to answer questions every once in a while. A few sample devices might be good to have so that debugging doesn't have to be done by email, but if necessary, that can be done.
In return, you will receive a complete and working Linux driver that is added to the main Linux kernel source tree. The driver will be written by some of the members of the Linux kernel developer community (over 1500 strong and growing). This driver will then be automatically included in all Linux distributions, including the “enterprise” ones. It will be automatically kept up to date and working through all Linux kernel API changes. This driver will work with all of the different CPU types supported by Linux, the largest number of CPU types supported by any operating system ever before in the history of computing.
As for support, the driver will be supported through email by the original developers, when they can help out, and by the “enterprise” Linux distributors as part of their service agreements with their customers.

When worded like that, it's hard to understand why more vendors don't want their drivers in the kernel. We really are talking about non-cost driver development here. What's more, you have people the caliber of Greg doing the work for you. He even addresses using an NDA in his post. So why the resistance from some? I'm not sure, but I'd guess it comes down to fear, perceived lack of control and misinformation. Especially in the case of commodity hardware such as NICs and on-board anything you should be begging to have Greg write your driver. Not only that, you should be paying him. But here he is, making the offer for free. Now if you make a specialized piece of $200,000 niche equipment, I can see why you would want a dedicated team of developers. Your device is likely not mainstream enough and is sufficiently complicated that you may not find someone willing or even able to maintain your driver. For the rest of you, I'm genuinely interested – what is holding you back? What would change your mind?
–jeremy

On OpenID

OpenID is getting a lot of attention this week. At LQ, we're obviously supporters of Open Standards and Open Source, but this is something slightly different. We're also believer that you own your identity and attention data. We've supported FOAF for a long time and do what we can to make sure your data is yours. While it's not without problems that need to be addressed, I'm happy to announce that soon, we'll also be supporting OpenID. It will be interesting to see how OpenID progresses. Your “login” is something that is very important to the big vendors (think Google and Yahoo! here). Once you've signed up for a login, it makes their barrier to entry to get you to try an additional one of their products or projects that much easier. After all, all you have to do is login to your existing account. Once you use 3-4+ services, it's quite hard to leave. You're what they call sticky at that point, and that's what every vendor wants. Of course, it's also easier for companies to maintain. We learned that the hard way at LQ. For example, for a while each site we launched had a different authentication system. It was really hard to get people to use LQ Bookmarks when you had to sign up for a distinct account. After we rolled it into the main infrastructure and allowed you to use your existing LQ login procedure, usage increased by leaps and bounds. At this point, the only LQ site that still has a distinct login is the LQ Wiki. That's by design though as we want the barrier to entry there to be a low as possible – it's a wiki after all. In that vein, the LQ Wiki will be the first site to support OpenID. After we're able to get a better feel for the technology and implementation, we'll make further decisions from there. No solid ETA for support at this point, but it should definitely happen this quarter. Stay tuned.
–jeremy

Adobe to release PDF to ISO

It's clear that the benefits of Open Standards are starting to be realized in the mainstream. While the PDF spec has been available for a long time, Adobe is now going to submit it as an ISO standard. From the press release:
Adobe Systems Incorporated (Nasdaq:ADBE) today announced that it intends to release the full Portable Document Format (PDF) 1.7 specification to AIIM, the Enterprise Content Management Association, for the purpose of publication by the International Organization for Standardization (ISO).
PDF has become a de facto global standard for more secure and dependable information exchange since Adobe published the complete PDF specification in 1993. Both government and private industry have come to rely on PDF for the volumes of electronic records that need to be more securely and reliably shared, managed, and in some cases preserved for generations. Since 1995 Adobe has participated in various working groups that develop technical specifications for publication by ISO and worked within the ISO process to deliver specialized subsets of PDF as standards for specific industries and functions. Today, PDF for Archive (PDF/A) and PDF for Exchange (PDF/X) are ISO standards, and PDF for Engineering (PDF/E) and PDF for Universal Access (PDF/UA) are proposed standards. Additionally, PDF for Healthcare (PDF/H) is an AIIM proposed Best Practice Guide. AIIM serves as the administrator for PDF/A, PDF/E, PDF/UA and PDF/H.
“Today’s announcement is the next logical step in the evolution of PDF from de facto standard to a formal, de jure standard,” said Kevin Lynch, senior vice president and chief software architect at Adobe. “By releasing the full PDF specification for ISO standardization, we are reinforcing our commitment to openness. As governments and organizations increasingly request open formats, maintenance of the PDF specification by an external and participatory organization will help continue to drive innovation and expand the rich PDF ecosystem that has evolved over the past 15 years.”

Some would argue this is coming a little late, but it's still good to see IMHO. Adobe is finally letting go of one of its crown jewels. While the spec has been previously available, licensing restraints mean we have a ton of free (gratis and libre) viewers but less writers. We should see more of a balance now (although some good libre writing options are already available).
Moving on to the impetus for this move. I think Adobe has a vision for the future of the web and it realizes the role Open Standards will play if they want to be a legitimate player in that future. More and more Government agencies are now requiring the document formats they use be standards based. The full PDF spec can now say that it is. You also have the fact that Microsoft is going to start pushing XPS, (XML Paper Specification, also called Metro) which is a direct PDF competitor. PDF being an ISO spec means that instead of Adobe being the perceived gatekeeper, anyone can propose additions/modifications. This means an ecosystem can build up. Of course, Adobe has a ton of expertise built up in this domain so they will remain the dominant player for a long time to come. This should be good for Open Standards as it will prove to some skeptics that the value and money aren't created by locking down a spec, but by opening it up.
–jeremy
, , , ,

GoDaddy pulls security site after MySpace complaints

Not directly Linux or Open Source related, but scary if you use GoDaddy, who is currently the largest registrar. They recently summarily suspended seclists.org, run by Open Source nmap author Fyodor. The problem here is that it was not as a result of DMCA takedown notice or a court order. It was simply at the request of MySpace. Apparently they left a voicemail for Fyodor and then suspended the domain 52 seconds later. No real verification, no chance for a defense and certainly before what I would consider due diligence. The best part is that the offending page, one of 250,000+ pages at the domain, was simply an archive of a mailing list! If you use GoDaddy I would seriously consider moving your domain. It's not because of this incident either. Everyone makes mistakes and if their response to this was something along the lines of “it won't happen again”, I'd say lesson learned. Their actual response:
When asked if GoDaddy would remove the registration for a news site like CNET News.com, if a reader posted illegal information in a discussion forum and editors could not be immediately reached over a holiday, Jones replied: “I don't know…It's a case-by-case basis.”
If you're a GoDaddy customer, is that a policy that you're comfortable with?
–jeremy
Disclosure: Although I don't really advertise it anywhere, I am an authorized reseller for multiple registrars – none of them GoDaddy
, ,

The meaning of FREE

A recent blog post by Mark Shuttleworth has sparked quite a bit of debate. His assertion:
We have to work together to keep free software freely available. It will be a failure if the world moves from paying for shrink-wrapped Windows to paying for shrink-wrapped Linux.
As free software becomes more successful and more pervasive there will be an increasing desire on the part of companies to make it more proprietary. We've already seen that with Red Hat and Novell, which essentially offer free software on proprietary terms – their “really free” editions are not certified, carry no support and receive no systematic security patching. In other words – they’re beta or test versions. If you want the best that free software can deliver, a rock solid, widely certified, secure platform, from either of those companies then you have to pay, and you pay the same price whether you are Goldman Sachs or a startup in Rio de Janeiro.
That’s not the vision we all share of what free software can achieve.
With Ubuntu, our vision is to make the very best of free software freely available, globally. To the extent we make short-term compromises, for drivers or firmware along the way, we see those as bugs, and ones that will be closed over time.

Basically, he's calling the two main Enterprise Linux distributors proprietary companies. Greg DeKoenigsberg, the Community Development Manager for Red Hat (aka quite active in the Fedora community) fires back:
You can call Red Hat “proprietary” all you want. That doesn't make it true.
The difference between Red Hat Enterprise Linux bits and CentOS bits is virtually nil; we make all of our source RPMs available to anyone who wants them. (Ask Novell if they do the same. I'll save you the time: they don't.) What is “proprietary” is the brand, and the quality of service you are entitled to receive by being a paying customer. Oh, and the notion that “the price is the same whether you're Goldman Sachs or a startup in Rio de Janeiro” is ludicrous.
If you want to duplicate that quality of service for Ubuntu users, that's noble and admirable. That's competition. But leave the FUD at the door, dude. You're better than that. Aren't you?

The discussion gets going from there. My opinion lies somewhere between the two, although that's not surprising considering they're basically diametrically opposed. I think calling Red Hat a proprietary company is a bit disingenuous. The bits are 100% freely available and Red Hat is extremely active in contributing code back upstream. I don't think there is any danger of there becoming a shrink-wrapped only Linux situation. Matt seems to think that Mark doesn't “get” enterprise Linux. I think it might be something else though, and that's that Canonical has an entirely different business model. The Red Hat business model has turned into something along the lines of (yes, these are *gross* oversimplification meant to illustrate a point) “If you want Linux for free, we have a release that is primarily aimed at developers that you are free (in every sense of the word) to do what you please with. The maintenance window is really short and we make no guarantees, but here it is – enjoy. If you'd like a solid and stable version of Linux that comes with support, ISV certifications and long term maintenance, we also have RHEL but it costs”. I've express my desire to have a maintenance without support version of RHEL (for a fee mind you, just a low one) in the past, but that may never happen. If you start with Fedora now and end up wanting official Red Hat support, you basically have to start 100% from scratch with RHEL. The Canonical version seems to be more along the lines of “Feel free to download any version of the product we have, be it LTS or not. If you'd like support, we'd be happy to provide it for a price.” For those not familiar with Ubuntu, LTS is “(Long Term Support”. This does give companies the added advantage of being able to try the official product while maintaining the ability to easily move to the supported product. That is nice, but if you want anything in the line of ISV certification at this time, Ubuntu doesn't have much to offer you (and that is important).
The one great thing that stands out to me here is that, due to the way that Linux is licensed both companies get a chance to prove their models. What's even better, they are doing so with essentially the same bits. Everyone in the community has the potential to win! No vendor lock-in on either side, so comparisons to “Microsoft of Linux” are non sequitur. See how great choice really is? The one final comment I'll add is that Red Hat has done a good job of proving their model works. Can Canonical do the same? Will they?
–jeremy
, , , ,

Will Sun License OpenSolaris Under GPLv3?

A recent eWeek article intimates that Sun will dual license OpenSolaris under the GPLv3 once it's officially released. Rich Green points out that no decisions will be made until after the GPLv3 is officially released, but indicates that Sun is in discussions with the community regarding the detailed terms of GPLv3. He also notes that he's happy with the current progress. That makes a lot of sense and I'd certainly not expect Sun to commit to a license that isn't even released yet. Putting that aside for a moment, let's assume that Sun does like the final terms of the GPLv3 (WARNING: there is no way to know whether that will be the case, this is just speculation based on that assumption). Dual licensing OpenSolaris under the GPLv3 would be an interesting move with a lot of up side for Sun. OpenSolaris is currently a bit of a step child to some in the Open Source community due to the CDDL. Releasing under the GPLv3 would allow Sun to use a license with some street cred while also keeping the code out of the Linux kernel. This is because Linus has made it extremely clear that the kernel will remain GPLv2 only. I've already mentioned that the GPLv2/v3 split has a chance to create a small fissure in parts of the community. We're used to a large number of licenses, but the GPL is a very dominant one.
However, a move of this nature could attract significantly more developer attention for OpenSolaris. I've seen speculation that Sun may encourage some Linux developers to dual license their code under the GPLv2 and GPLv3, but that makes little sense to me. To make it back upstream into OpenSolaris proper, the code would also need to be licensed under the CDDL. Note that is my understanding of how things would need to work. I am not a lawyer and a three license scenario is far behind what I'm able to properly grok. Back to actual code, I've already said how much I like DTrace and Sun is doing some interesting things elsewhere. The ability to increase collaboration and share code should be beneficial to all, even if kernel-to-kernel code flow is prohibited. Much of the early Linux growth came at the expense of Solaris and the other proprietary UNIX variants (not Windows, as many seem to think) so it's hard to gage what impact this may have on commercial and enterprise Linux uptake. For now, it's a moot point. Once the GPLv3 is actually released however, it will be something to keep an eye on.
(Note: I had typed a much longer version of this post that contained some additional thoughts and ideas. Firefox crashed as I was hitting submit and this is a quick reconstruction of my thoughts. Hopefully I've hit on all the salient points of my original post.)
–jeremy
, , , , ,