November 10th, 2009

Open source be not proud

Posted by Dana Blankenhorn @ 5:53 am

Categories: Database Management, Development, General, Oracle, Sun Microsystems, business models, java, management, mergers & acquisitions, support

Tags: Larry Ellison, Open Source, Dana Blankenhorn

Open source is, in part, a release of ego.

When a program is proprietary, it’s yours. You own it. You can feed it or you can kill it.

Not so with open source. When software is made open source it is with the knowledge that its fate is shared among all stakeholders. The contributions that make it valuable may well come from outside, the direction of the software is no longer completely in the hands of its owner or sponsor.

Larry Ellison doesn’t understand this, and I suspect neither does Wall Street. Otherwise, why would the Street be cheering on Ellison’s suggestion that he’ll kill Sun to keep Euro-hands off mySQL?

More than the future of mySQL is now on the line. So are the futures of Java and OpenOffice, and all the other projects Sun Microsystems sponsors. Ellison thinks this fact should make the EC Competition Commissioner, Nellie Kroess, back off. He seems to think the U.S. government can make Kroess relent.

The key to why Ellison is wrong can be found in the paragraph above. It’s one word. I’ll wait…

The word is sponsors.

Open source companies don’t own the code bases that are in their charge. They seek to monetize the code, so the code can be expanded, so it will draw more committers. Acquia doesn’t own Drupal, and Automattic doesn’t own Wordpress. The code bases are, in fact, owned by the community, simply by virtue of being open source.

Ellison seems to think that if he snaps his fingers and brings down the wrath of heaven, then mySQL and Java and OpenOffice will cease to exist. This would be true if they were closed source. In that case they would be orphaned, and if no buyer were found support would disappear.

Open source does not work that way.

Sure it would be tough for these big projects to find new sponsors. But there are plenty of prospects around.

Google would have an interest in Java, as might Microsoft. IBM already has a stake in Open Office. I’m certain we can find another home for mySQL, too. Even Glassfish might well find a new home within the federal government.

Ellison’s threat to kill Sun’s open source projects if he does not get his way is an empty one. Someone would pick up what remaining pieces have value.

Open source, divorced from its sponsor, turns to software water, and would quickly flow through Ellison’s hands.

Go to an open source conference. Listen carefully to the commercial open source businesspeople you see there. They may talk about their kids and their companies, their hobbies and their passions, including a passion for the projects they control.

But they know those projects are more like their kids than their sailboats. They are responsible for the software they control. They do not own it. It’s not “my” software. It’s “our” software.

This is the attitude you must take if you’re to make a success of an open source business. This is why many in the proprietary world, like Larry Ellison, confuse it with communism, or socialism, or some other foreign -ism.

Open source be not proud. Open source code responds to whomever gives it the love of time. The parents aren’t those who gave it the DNA of capital, but those who gave it the love of hard work.

November 4th, 2009

Ramji delivers a CodePlex process

Posted by Dana Blankenhorn @ 5:23 am

Categories: Development, General, Microsoft, business models, support

Tags: Commons, Microsoft Corp., Ramji, CodePlex, Open Source, Tools & Techniques, Management, Dana Blankenhorn

Successfully pulling code out of a big company can be like pulling the teeth off a lion, without anesthesia.

Sam Ramji (right), the former Microsoft executive who remains President of CodePlex, president of the CodePlex Foundation, which surrounds the Microsoft open source repository, said the key to success is a process.

CodePlex has published a draft of its process, a Project Acceptance Guideline, and is seeking comments from the community on it. The draft describes the advantages of contribution and provides a step-by-step guide for delivering new projects to CodePlex.

Ramji told ZDNet he’s anxious to get community input into the Guideline and will take that input seriously. He wants CodePlex to become a bridge between the open source community and corporate interests.

“How do we solve the problem of corporate contribution to community projects?” he asked. “The barrier is comfort. That comes from a clearly understood process and a well understood mechanism so people see contributing as low risk.”

Ramji said the Codeplex process says “here is how you should contribute in a way that’s sustainable for you and safe for the developer. There should be derivative works with no concerns about patents.”

CodePlex contributions come from software companies and non-industry sources, Ramji said. Software companies learn, through the CodePlex process, which elements of their IP are valuable and which are more valuable in the commons.

Then there are the non-industry contributions.

“Wall Street banks have talked to me over the last few months about contributions they couldn’t get legal clearance on. CodePlex offers a template for how it can get done. We have an organization that can own the copyright, that can accept cash as well as code, and can do the community management.”

In both these cases CodePlex is delivering code to the commons that might not be contributed otherwise, valuable code that can be used to build new applications.

“CodePlex is a lot of my future now,” he said, even though he has left Microsoft to become vice president of strategy at Sonoa Systems, a cloud start-up.

And the work is gratifying. “The Foundation is growing pretty quickly in terms of input from community members and corporate interests.”

October 30th, 2009

Making a foundation a real solution to the mySQL mess

Posted by Dana Blankenhorn @ 6:09 am

Categories: Database Management, Development, General, Oracle, support

Tags: Oracle Corp., MySQL, Amazon.com Inc., Open Source, Databases, Enterprise Software, Software, Data Management, Dana Blankenhorn

The mySQL mess won’t go away.

It doesn’t want to go away because mySQL is very useful. The latest demonstration? Amazon’s support of the community version of the software at its EC2 cloud.

Two forces are trying to keep the code from forking beyond all recognition.

One force is Oracle itself. You buy a unit worth $1 billion as part of a larger deal and you don’t want it dribbling through your fingers.

The other is mySQL co-founder Monty Widenius, who helped launch a group called the Open Database Alliance in May to support a new “community version” of the software.

So far only three companies have joined the ODA. A general meeting is being organized for Zurich, but it might just be a few guys sitting around a table at a bar. The group’s Twitter feed is dormant.

The problem for the ODA is as simple as do-re-me. They don’t have enough of what makes the world go around. When the check comes for the drinks in Zurich, I have to wonder who’s going to pay it, or whether they’ll make the waitron wait while those around the table hunt the cushions for enough quarter-Euros to split it.

Amazon’s move offers the chance for a reboot.

Oracle has a greater financial stake in the future of mySQL than Monty Widenius does. So does Amazon. So, very likely, do some other large players.

Instead of holding the software as semi-proprietary, which the Europeans won’t accept, or just writing a check to make it go away, which Oracle does not want to do, why not create a new foundation, on the model of Eclipse or Apache?

Those who put in the most would have the biggest say in this new group. The community would be represented, in other words, but they wouldn’t be in control. The major sponsors would be.

The mission of the new mySQL Foundation would simply be to hold the code base together, to provide a central point for updates, QA and bug fixes, to run the “official” version of the code base and plot its future direction. To run the forge, in other words.

Everyone would benefit. Oracle would get Sun, and community support. The community would get a vital development hub, well-funded. Amazon would be assured its support of mySQL, and the investment it has made in it, is not being wasted. Everyone could take the software in their preferred direction and share the results.

I guess this is all too reasonable to go anywhere. But it’s worth a shot.

October 27th, 2009

Will OpenSolaris survive Oracle?

Posted by Dana Blankenhorn @ 5:50 am

Categories: Distributions, General, Linux Server OS, Oracle, Sun Microsystems, mergers & acquisitions, support

Tags: OpenSolaris, Oracle Corp., Open Source, Dana Blankenhorn

With Apple having followed through on its promise to dump the  ZFS file system, and Oracle still preparing to take over Sun any time now, we should consider the future of the technology, and perhaps the OpenSolaris operating system it rides on.

OpenSolaris was Sun’s attempt to secure a future for what had been its proprietary Unix. It has some advantages over Linux, on which its advocates will gladly bend your ear over a couple of beers.

But there’s a curious thing about technical advantages in the age of open source. They don’t matter as much as they once did. After all, if open source can compete with proprietary products that have decades’ head start and armies of programmers behind them, how big is an open source program’s technical details?

Open source has taught some hard lessons.

It’s pretty clear that the programming of a few little features don’t amount to a hill of beans in this crazy software world. Oracle has its Unbreakable Linux program in which it has invested heavily. Does it really make sense for Oracle to keep carrying OpenSolaris, or is it time for Larry Ellison to tell it, “Here’s looking at you, kid” and just walk away?

Best movie ever? You decide at Amazon.com.

October 27th, 2009

Black Duck finds its business makes sense

Posted by Dana Blankenhorn @ 5:24 am

Categories: Development, General, Security, business models, support

Tags: Encryption, Security, Dana Blankenhorn

I am of two minds concerning the Black Duck release on encryption in open source.

On the one hand it’s interesting to know that 4,000 out of 220,000 tested (less than 2% if you’re scoring at home) contain strong encryption, the kind the U.S. still thinks of as “munitions grade.”

On the other hand there is no reason to panic, as Dr. Dobbs did. And a close look reveals this release is basically a product launch for Black Duck Export, a new feature in its “watch out, look out, over there” suite of offerings that includes warnings on copyrights and other important issues.

The image that often comes to mind when I think of Black Duck is of Daffy and his friends flying across the sky when Elmer Fudd & Co. start blasting from down below. On the other hand lawyers and spies can also use Black Duck software, so security through obscurity may be a bad move.

For the government this is an opportunity to choose its attitude regarding encryption, which has been an issue for software developers going on 20 years now. Pretending that the U.S. is the only home of this stuff is just plain silly and rules should be uniform. The encryption wars should have ended a decade ago.

October 20th, 2009

Nimsoft takes another step forward in cloud monitoring

Posted by Dana Blankenhorn @ 4:00 am

Categories: Cloud Computing, General, Infrastructure, LANs and WANs, Network Administration, Software as a Service, support

Tags: Environment, Monitoring, Nimsoft, Performance Management, Human Resources, Workforce Management, Dana Blankenhorn

Nimsoft, which specializes in monitoring tools for data centers, is entering the cloud market.

CEO Gary Read said that, as of today, all the company’s performance monitors will become available for external environments, meaning they can deal with hosted providers, managed services, SaaS or clouds.

<!– @page { margin: 0.79in } P { margin-bottom: 0.08in } –>

“The computing infrastructure of the future will no longer be hosted internally by the customer, but customers will have a combination of internal and external resources. They’ll want to see the performance and service across that entire environment.”

The initial release will cover the Amazon cloud and Rackspace, but Read said the company is taking measurements on Salesforce.com and has plans to expand to Microsoft and Google cloud offerings.

Read insisted this is not like those “cloud weather reports” we wrote about last year. <!– @page { margin: 0.79in } P { margin-bottom: 0.08in } –>

“You haven’t been able to monitor a single customer’s usage, and you haven’t been able to pull all the different threads, and combine that with an internal environment, into a single integrated view of the service delivery.

That’s the big play. It’s not about one provider. It’s about pulling all the threads together, because everything then becomes part of the same fabric of service delivery.”

Read said that, for cloud vendors, his move and that of his competitors will help in cloud adoption and make clouds common parts of computing infrastructure.

The company will be hosting a Web conference on its site today to discuss the new release and new direction. I wonder if the name “rainbow” is taken?

October 15th, 2009

Should Google spin Android into a foundation?

Posted by Dana Blankenhorn @ 7:29 am

Categories: Development, General, Google, Hardware, Linux Handheld, Standards, management, support, telecom, wireless

Tags: Google Inc., Fork, Foundation, Eclipse, Linux Foundation, Forks, Linux, Java Development Tools, Open Source, Operating Systems

Google faces a conundrum.

How does it maintain control of Android and at the same time build a community of interests in which developers can seek profit?

The easy answer is to turn the Open Handset Alliance into the Android Foundation. (Fans of the late Isaac Asimov will recognize this fellow even in French.)

Critics love to claim that Eclipse is just an IBM front, but that’s a cheap shot, based on the fact that IBM gains huge benefits from Eclipse without having to pay all the bills there.

Foundations can be a great way to organize vendors who have a common purpose but divergent business plans. The Linux Foundation is a good example of this.

But there are risks in an Android Foundation, as Symbian’s David Wood said when they were going open source a year ago.

Forks are one.

Foundations lead naturally to forks. Every vendor who sells an “enhanced” version of Eclipse tools is pushing a proprietary fork. There are dozens of Linux distros, each of which forks the code in some way to provide added value.

How much Android forking can Google stand before the value starts dribbling through its fingers? Like to see some stuck-up Microsoft search engine sitting on an Android phone? (Make your blood boil? Well I should say.)

There is, of course, another risk in going the Foundation route. It doesn’t always work. Witness LiMo, which Motorola recently abandoned for Android. Witness Moblin, which Intel gave to the Linux Foundation. Witness Symbian itself for that matter.

The difference between the OHA and a conventional software foundation is that for Android to move forward it must first be expressed in phones, in hardware. The chicken-and-egg question here yields an easy answer. It’s the chicken. An egg, the software, is pretty meaningless if it’s just sitting on a server.

This fact reduces the threat of a fork. The value of any Android handset lies in its compatibility. Without that it might as well be a Windows Mobile set.

So long as Google is the biggest investor in Android, then, it’s probably doing the right thing by avoiding the foundation model. But at some point the rest of the ecosystem needs to grow up for Google to get its investment back.

So if Google does set up an Android Foundation some time down the road, know that it’s a sign of success, and that it no longer has to push this rock up the hill all by itself.

October 1st, 2009

GroundWork Monitor 6.0 offers new view into virtual data center

Posted by Paula Rooney @ 8:19 am

Categories: Applications, FOSS, General, Microsoft, support, virtualization

Tags: Virtual Data Center, Performance, Dashboard, Microsoft Corp., JBoss, GroundWork, Data Center Virtualization, Virtualization, Performance Management, Open Source

GroundWork is a leading open source systems and network provider and recently enhanced its platform with JBoss and Microsoft System Center support.

That’s not all.  GroundWork Monitor 6.0, which shipped on September 2, offers a new “Seurat” view designed to allow administrators to access the status and performance of hundreds of hosts and services.  Version 6.0 also offers a redesigned status view that provides performance data in a single view.

The company quotes Cameron Haight, a Gartner Vice President, who comments that the Seurat View is a break from traditional topology and tree-style views, which may not be sufficient for dynamic, virtual data centers and cloud computing infrastructures.

Monitor 6.0 also features a revamped user interface based on the JBoss Portal and a new dashboard builder for both the community edition and enterprise editions.

The new user interface purports to offer finer grained, roles based access controls and a higher level of customization. The software’s new tailor made monitoring allows admins to to create and customize personal and roles based dashboards for any purpose.

Earlier this month, GroundWork announced the availability of a connector to integrate its software with Microsoft System Center. The GroundWork Connector, which is available standalone and separate from GW Monitor 6.0, pulls data from Microsoft System Center and displays it within GroundWork Enterprise Edition.

September 24th, 2009

HP supporting Oregon State Linux portal

Posted by Dana Blankenhorn @ 5:00 am

Categories: Development, Distributions, General, Hardware, Linux, Linux Server OS, education, support

Tags: Hewlett-Packard Co., Linux, UNIX, Portals, Open Source, Operating Systems, Software, Internet, Dana Blankenhorn

Hewlett-Packard is quietly lending its support to a new Oregon State support portal for Linux, Communitylinux.org.

The site is hosted by Oregon State’s Open Source Lab and, while its content is HP-centric, it is pointedly not part of the HP domain.

(This happy little penguin is supporting the Linux Foundation’s annual membership drive. Are you?)

The idea is that non-commercial Linux distributions running on HP hardware can find support, even though it’s not officially coming from HP. We’re talking about server versions of Asianux, CentOS, Debian, Fedora, OpenSUSE and Ubuntu.

All these programs can now be run on HP servers without violating their hardware warrantees.

Bdale Garbee, HP’s Linux head, indicated at LinuxCon that HP’s support over time will include loading tests and certifications it has done with Linux onto the site so the community can use the data.

It’s a small announcement, but LinuxCon is a small show.

September 17th, 2009

The forge as a customer club

Posted by Dana Blankenhorn @ 6:01 am

Categories: Development, General, LANs and WANs, Linux, Network Administration, marketing, support

Tags: Monitoring, Open Source, Dana Blankenhorn

Even after open source companies began insisting on having their own “forge” sites, programmers were the target market.

(Groundwork was a sponsor of the Linux Picn*x in 2007, where this picture was taken. Good times.)

Sites like Appcelerator.com, which I covered last year, offered extensive “atta boys” to developers, letting them put their credibility up-front before  their peers, encouraging contributions of knowledge and code.

Forge sites have continued to evolve, and the new site of Groundwork is a good example of the new trend.

Monitoringforge still welcomes developers, but it aims to be more. Specifically it aims to be a customer’s club.

To quote from the press release:

MonitoringForge is designed to appeal to IT administrators who want to compare and understand the differences between various open source monitoring tools and plugins available today, facilitating the selection of open source monitoring software over proprietary offerings.

Note that they are not just inviting current customers to the site, managers and administrators. They are also inviting prospects, promising to build a community for all those interested in the niche they serve.

That means more than offering Groundwork software and seeking tweaks and tips. It also means what amounts to editorial coverage of other monitoring packages — Nagios, NeDi, and WMI are all on its front page.

Can a vendor deliver what amounts to editorial coverage as part of a sales job and still serve the guys and gals in the corporate boiler room working on code?

September 10th, 2009

Could One Laptop Per Child be a bad thing?

Posted by Dana Blankenhorn @ 6:08 am

Categories: General, Hardware, Internet, education, mass market, support, telecom

Tags: Child, Laptop Computer, One Laptop Per Child Project, Pastorius, Client Hardware, Productivity, Notebooks, Hardware, Notebooks & Tablets, Dana Blankenhorn

You know the old saw about helping people rise from poverty? Give a man a fish and he eats for a day. Teach a man to fish and he eats for a lifetime — or until the fish run out.

The idea of One Laptop Per Child (OLPC) is to teach a child to fish. But is it really more like giving a child a fishing pole?

It’s true the man who offered this criticism, Tom Pastorius of Projects Abroad, is selling something. Volunteering. He is the exclusive agent of the UK-based volunteering group for the U.S. and Canada.

He says volunteers should be there to help kids who get laptops, ideally at a community computer cluster rather than leaving each child to their own devices. This reduces the child’s frustration, without hindering the learning process.

Fair enough. One volunteer per child sounds better than one laptop. But it’s not going to happen. So that means we go back to the idea of computer labs, with the hardware locked away at night to keep it safe?

Pastorius’ concept also flies in the face of recent trends in Australia, where as we noted earlier students in New South Wales all got laptops to start the year. Windows laptops.

How long do you think those will last, with their hard drives and optical drives and what are the odds the kid’s going to update their antiviral each day? How about one computer repairman per child?

Which leads to my own modest proposal. Client hardware is not the problem. OLPC has proven that hardware can be super-cheap. Taiwanese Netbook makers have taken that idea and run with it.

The real problem is connectivity. Would our aid money, and the work of volunteers, be better spent upgrading the bandwidth available to villages in the developing world? One WiFi per child.

Or perhaps one pigeon?

September 9th, 2009

Euros try multi-product open source support

Posted by Dana Blankenhorn @ 12:01 am

Categories: General, business models, support

Tags: Covalent Technologies, Credativ, Open Source, Dana Blankenhorn

Credativ, a 10-year old open source support company launched by Postgres contributor Michael Meskes, is bringing its act to America.

The Open Source Support Center aims to solve your open source problems no matter in what program the problem lies, president Joe Conway told ZDNet.

“The concept of the support center is we’ll take on support for any open source software, in a holistic way. We’re not just supporting Postgres or Debian or RedHat. We give the customer some place to call to get their problem solved.”

Conway’s description gave me a distinct feeling of deja vu. A bit of Googling found my old story on Covalent, which also tried multi-product support in the middle of this decade. Covalent was eventually bought by SpringSource, now part of VMWare.

One difference might be that Covalent specialized in Apache products, while Credativ will work with any product. Another difference is that Credativ has 10 years’ experience making its business model work. ” They’ve more than doubled their staff in Europe over the last two years,” Conway said of his strong European arm.

The U.S. operatives of Credativ, under the trademark The Open Source Support Center, will be just east of  San Diego in the suburb of El Cajon.

So can Credativ succeed where others have failed? Will companies buy open source support that is not tied to a specific product?

September 3rd, 2009

Young man yells at cloud

Posted by Dana Blankenhorn @ 5:55 am

Categories: Cloud Computing, General, Infrastructure, Internet, Network Administration, management, mass market, support

Tags: PC, Cloud, Blogger-novelist-activist Cory Doctorow, Man, Client Hardware, Desktops, Netbooks, Nettops & MIDs, Hardware, Dana Blankenhorn

Blogger-novelist-activist Cory Doctorow is out today with a screed condemning the whole idea of cloud computing.

(Here he is in Wikipedia. Doesn’t that live somewhere in the cloud?)

Here is a summary.

The Man is trying to put us down. It’s all a conspiracy to make us pay for what we could do ourselves for free. The corporate shills want to control our machines and through them, our brains.

I agree with Doctorow on many things. On other things I’m sympathetic.

On this he’s dead wrong.

Running a PC is a hassle. There are software updates, there are anti-virals and anti-spyware and registry cleaners to worry about. It can take five minutes for even a Netbook to boot up, and another five minutes to shut it down.

Hardware is not the issue. Client hardware is an incredible bargain. The issue here is software.

All the problems now endemic to Windows machines are slowly infiltrating the worlds of the Mac and Linux, too. This has to do with the size and complexity of modern operating systems, and the large number of very nasty people working overtime to break them.

Maybe, if you have just one laptop, you can deal with this expense and hassle. But even small companies may now have 10-20 or more PCs running at once. The expenses of managing clients are driving companies to the wall.

The idea of the cloud is to abstract this complexity, take it out of the hands of users and put it in the hands of experts. One set of experts can handle the hassles of thousands of users, and hundreds of companies, for less than those users and companies are paying now.

Or as Mark Twain once said, “Put all your eggs in one basket and watch that basket.”

As to the broader market, Mr. Doctorow has one kid, a toddler. Congratulations. I have three PCs, my dear wife runs three, my daughter has two and my son has a desktop for gaming with a terabyte of storage.

We use both wired and wireless networking to keep it all together. Several times a year I lose a day of work scrabbling around on my knees, under a desk, trying to check wiring. I have a repair guy on speed dial. I’m not that unusual, and becoming less so all the time.

The complexity and vulnerability of PCs means you can’t just run one. I always have my Netbook on standby for emergencies. If malware infects me, or the cable goes out, I can be at a local coffee shop within minutes.

I no longer even trust my PC for really important stuff. That goes on a USB stick. You can get a 32 Gigabyte stick these days for about $70. Next week it will be less. Wear it around your neck, plug it into the cloud anywhere and, if everything works right, there you are.

Now Doctorow has valid concerns. Networks are not yet built to handle massive use of clouds. The legal environment for cloud users is, well, cloudy.

Fact is these are early days. In PC time it’s 1978. In Internet time it’s still 1994.

What 30 years of experience tells me is that condemning the future when you don’t know what it looks like is never a wise move.

August 8th, 2009

What Asus wants in a Linux

Posted by Dana Blankenhorn @ 6:06 am

Categories: General, Google, Hardware, Linux, Linux Laptop, Microsoft, Strategy, mass market, support

Tags: ASUS, Netbook, Netbook War, Linux, Netbooks, Nettops & MIDs, UNIX, Operating Systems, Open Source, Software, Hardware

Sell-through and no hassles.

What Taiwanese OEMs want from Linux, in other words, is just what they get from Microsoft.

1. Sell-through. Support that will get the product off the shelf. As I’ve said before there is a price lower than free. Taiwan demands a Linux that will pay that price. Intel can, Ubuntu can’t.
2. No Hassles. Asus does not want its gear coming back. The high return rate of last year’s Linux Netbooks not only doomed those distros but any distro from a firm too small to provide enough after-sales support to keep the units in the field.

You can summarize all this in one word — support.

Taiwan OEMs like Asus would love to get out of the Microsoft box, because Microsoft controls the relationship and takes most of the money. But if it can’t get support from the software vendor, on the ground, the effort is not worth it.

The question that must be asked, then, looking at reports from Germany’s Electronista that Asus will deliver a Moblin-based netbook this fall, is whether Intel is going to put serious dollars behind the effort, or just point to the Linux Foundation (the official sponsor of Moblin) when the going gets tough.

Those same questions pertain to the chances of a Google Chrome OS netbook, which Germany’s Network News is discussing this weekend. The netbook wars are being played with money, big money, bigger money with every turn of the wheel. Those software vendors who don’t have the ante don’t get into the game.

Linux’ only hope in cracking the Taiwanese netbook market is with a Linux whose owner is willing to pony up the big bucks for both pre-sales and post-sales support.

August 5th, 2009

Microsoft admits the obvious but not the truth

Posted by Dana Blankenhorn @ 5:44 am

Categories: Cloud Computing, General, Google, Hardware, Microsoft, business models, mass market, support

Tags: Microsoft Corp., Open Source, Dana Blankenhorn

In its latest 10-K report Microsoft admits what we already knew. Open source has run away with its lunch money. (Picture from MeallDubh.org, the blog of John McCreesh.)

A whole bunch of companies are identified as villains. Springsource. Canonical. RedHat. Google.

What Microsoft does not admit here is the truth. Look at that list of “competitors” again. Other than Google — which gets is money from ads — are any of them more than a a bug heading for the Microsoft windshield?

What threatens Microsoft is not an open source company, or a collection of open source companies, but the open source model.

Microsoft can maintain its share only by offering subsidized software as an alternative to free. But it can’t maintain its margins. It’s going to have a tough time maintaining profitability.

But Microsoft has an even bigger problem than the open source business model.

The big sea change over the last years, the force Microsoft can’t keep up with, is the cheap client. A Netbook may sell with a $3 copy of Windows XP, but how much software will the owner of a $300 product buy?

Fact is, even when the support is online, Windows costs money to support. It costs money to create patches and updates, money for servers to push those patches out. It’s money that has to be spent, because without it Microsoft clients get infected with malware and become useless.

By contrast Linux is modular. A Linux client, whether a netbook or an Android phone, needs to run only those modules necessary to the function it is performing at that moment. Over time open source is just cheaper to support. And by linking clients to the cloud you centralize that support load, even monetize it.

So you have competitors who can live on less food than you need, hardware evolving toward forms that must find a lower-cost form of support, and a possible breakthrough in business models that you can’t seem to touch.

The truth is that Microsoft Windows, and Microsoft itself, have become dinosaurs in a mammalian world. To compete Microsoft must evolve.

July 30th, 2009

Ending DNS abuse with European open source

Posted by Dana Blankenhorn @ 12:34 pm

Categories: Development, Infrastructure, Security, identity, support

Tags: DNS, Open Source, Sponsors, Domain Names, Networking, Security, Internet, Dana Blankenhorn

A collection of European Internet insiders have announced OpenDNSSEC, a project aimed at managing the security of domain names on the Internet.

The group notes that DNS caches are no longer secure, and a white paper says the automated checks will make the creation of secure zones automatic, combining DNS records and digital signatures.

Sponsors are in Europe include the English registry Nominet,  NLnet Labs of the Netherlands, the Internet Infrastructure Foundation .SE in Sweden, the Swedish Kirei AB consultancy, SIDN, which maintains the .NL Netherlands domain, and SURFnet, which handles the same country’s university network, and English DNS consultant John Dickinson.

Secure domain name abuse is one of the main tools hackers have for getting past security systems, and making domains tougher to forge is something that is devoutly to be wished. The announcement of OpenDNSSEC follows an Internet Engineering Task Force meeting in Sweden.

The poisoning of DNS cache has become commonplace since Dan Kaminsky demonstrated how the DNS security model is flawed two years ago.

June 24th, 2009

Reductive to service Puppet open source configman tools

Posted by Paula Rooney @ 11:46 am

Categories: Applications, Cloud Computing, Development, Distributions, Enterprise Policy, FOSS, GPL, General, Infrastructure, Linux, Linux Server OS, Red Hat, business models, support, venture capital, virtualization

Tags: Open Source, Configuration Management, Tool, Reductive, Puppet 0.25, Paula Rooney

Key founders of Puppet have incorporated and received $2 million in venture capital funding to advance the open source configuration management software project.

Reductive Labs,  which has evolved from the same named consulting firm founded in 2003, will provide training, service and support for Puppet, the next generation open source infrastructure automation framework which is reportedly gaining strength and numbers of users.

Reductive has formed partnerships with Red Hat, Fedora and Canonical and has about 20 paying customers. Puppet currently supports Linux, Unix and Macintosh environments.

Puppet, which was first made available under the GPL in 2005,  is a configuration management framework that enables customers to write policies about how web servers should be configured, how database servers should be configured and how mail servers should be configured,” said Andrew Shafer, chief strategy officer for Reductive Labs, which will be headquartered in Portland, Oregon. “Puppet lets you write policies, enforce them and automate them on an ongoing basis and operating system installation through patches and upgrades.”

Shafer said it’s important to have a robust policy-based configuration framework that can significantly speed up deployment of corporate servers. He noted that policy-based tools are valuable because few servers are configured in the exact same way in any corporation.

He pointed out that configuration management becomes even more critical as virtualization and cloud computing take off.

“With virtualization, your hardware headache eases but with thousands of virtual machines you’ve multiplied your configuration management complexities,” said Shafer. “People are bringing up thousands of [virtual] machines with EC2 [cloud] and configuration management complexity is further magnified. Bringing up a test infrastructure or a deployment infrastructure becomes a much easier proposition than trying to manage it in other ways.”

One senior systems engineer at Digg.com was able to rebuild 60 [virtual] machines from scratch in two hours [using Puppet] that would have taken two full days of work if done manually. “And I was largely a spectator,” said that engineer, Paul Lathrop, of Digg. “Now that’s automation.”

“And if he needed to build 600 machines, it wouldn’t have taken much longer,” because of the policy-based configuration management approach, Shafer said.

Its biggest competitor is amorphous: thousands of unique scripts system administrators write for their own environments, Shafer said. There are some model-based configuration management frameworks developed by BMC’s BladeLogic and HP’s Opsware but nothing in the open source space that compares to Puppet, Shafer said.

Reductive has no plans to commercialize the framework into a product per se and will focus exclusively on the services side of the business. Puppet 0.25 is currently in beta testing and represents a huge step forward: three times the speed in one third of the memory footprint of the current 0.24 series.

Read the rest of this entry »

May 11th, 2009

Education lessons for open source

Posted by Dana Blankenhorn @ 7:58 am

Categories: General, Government, Implementations, Infrastructure, education, support

Tags: Education, Open Source, Dana Blankenhorn

One of my favorite ZDNet blog posts from last week was Christopher Dawson’s “How much does open source cost schools.”

The post riffs off a Tech & Learning piece by Randy Orwin, an open source advocate in Washington state, who concluded that “free is not free” and large-scale open soruce implementations are difficult.

I have been working on the question of technology in education for 21 years, since my first child was born. I have concluded that education offers great lessons for all enterprises, especially as it has begun using open source.

I found enterprise technology in education was a dead loss before the game became Internet access. Client technology changes too fast, I learned. The moment teachers were trained on “multimedia” their tools and skills were obsolete. Only by moving to online resources that upgrade themselves could the investment begin to make sense.

Education, in other words, is an enterprise buyer.

What enterprise buyers want are desktops and applications offering workforce productivity and plenty of back-end Internet capacity. Schools and offices aren’t that different.

So the lessons Dawson offers can be applied generally.

1. Open source is a make-or-buy decision. If you go with open source make certain you have the expertise on-staff to make things work.
2. Centralize operations and automate the pushing and updating of applications.
3. Before choosing an open source solution make certain you’re prepared to be part of an open source community.

I would add that the solution you choose may also drive your best-and-brightest. In a conventional enterprise this is your IT shop. In a school setting this is your computer club geeks. If your system is running Linux, that’s what they will learn.

The difference between a school and a conventional enterprise is what happens to that expertise. In most businesses it sticks around. In school it graduates.

So what do you want your smartest kids to become, users or doers? That should offset some of the implied “costs” of open source.

May 6th, 2009

Nagios fork warning to Oracle

Posted by Dana Blankenhorn @ 7:52 am

Categories: Applications, Development, General, Implementations, Oracle, Strategy, java, support

Tags: Oracle Corp., Network, Fork, Nagios, Matt Asay, Open Source, Dana Blankenhorn

A fork of the Nagios network monitoring tool called Icinga has officially launched, with the first stable version due October 28.

Matt Asay says this illustrates the health of the open source movement. 

Nagios is a 10 year-old project and those involved with the fork say they include members of its community advisory board and makers of Nagios add-ons.

But there is meaning here for more than Nagios. There is great meaning here for Oracle.

As Oracle prepares to take possession of open source projects like OpenOffice, Java and mySQL, the Nagios fork is a warning that open source code can’t be suppressed.

Whatever Nagios’ managers did to cause the fork, it can’t be nearly as bad as the actions Oracle has been suspected of in its approach to open source competition.

Forks of these larger projects would doubtless move even more quickly, and gain more support from both developers and users, than Icinga, which is used by network managers and not the general user community.

The open source warning to Oracle is clear and was probably put best by Tina Turner, above. You better be good to me.

May 4th, 2009

How far can Springsource scale?

Posted by Dana Blankenhorn @ 6:32 am

Categories: Applications, General, management, mergers & acquisitions, support

Tags: IBM Corp., Open Source, Dana Blankenhorn

I have a long-time bias in favor of Springsource.

Their management has always impressed me with their professionalism. So I was not surprised this weekend to learn they are buying Hyperic, or to quote the press release ” substantially all of the assets of Hyperic.”

Good management doesn’t overpay in a downturn.

Our own Matt Asay then put up an intriguing headline. “SpringSource acquires Hyperic, takes on IBM and Microsoft.”

This led me to some intriguing questions:

1. Are they ready to play that game?
2. Will they ever be?
3. Will any open source pure play ever be?

I read to the end of Matt’s piece and he didn’t back down. IBM and Microsoft can “build, run, and manage applications” and now so can Springsource.

Really? I buy build, I can even get into run, but manage? Anyone who has read The MysticalMythical Man Month knows that as projects scale, project management must scale.

Open source is powerful, in part, because it effectively breaks systems down into pieces and puts small groups in charge of each piece. Open source as a business reflects this reality. Most manage just one project, some manage a few, very few run several.

Moreover, IBM and Microsoft don’t just run their own projects, but customers’ projects. Their biggest projects. Project management of large systems is completely different from supporting software. Can any open source company, even Springsource, scale to deliver that?

And if one did would those who believe in open source be happy with them? Or would we immediately start throwing bricks through their virtual windows, shouting how unclean they were?