rand($thoughts);

Infoworld’s Paul Krill has an interesting article about the server operating system becoming a two-horse race between Linux and Windows in which both OSes will grow at the expense of others (i.e. Linux).  Market research we did at IBM over 7 years ago indicated that Linux would grow at the expense of Unix more so than Windows.  So, I’m not surprised that a Gartner survey from a Linux-oriented conference indicated a 3-to-1 ratio of migrations to Linux will be at the expense of Unix versus Windows.

Another interesting finding from the Gartner survey:

“Linux was ranked by 39 percent of respondents as the OS expected to have the most growth in their datacenters during the next five years. Windows was a close second, ranked as the OS with the most growth potential by 35 percent of respondents at the Linux-oriented conference.”

The fact that 35% of respondents at a Linux-oriented conference, reported growth of their Windows OS install base is telling.  As enterprises get more comfortable with Clouds, offering an environment to deploy Windows and Linux workloads is a no-brainer. I’m certain that Microsoft realizes this and will announce Linux OS support when the Microsoft Cloud goes live ;-)

Organizers of the first Open Video Conference (June 19-20 in NYC) are requesting proposals for panels, presentations and workshops.  What is “Open Video”?  I asked the same question a few minutes ago!  I found out that:

“Email, blogs, and other staples of the open web rely on ubiquitous and interoperable technologies that have low barriers to entry; they are massively decentralized and resistant to censorship or regulation. Video, meanwhile, relies on centralized distribution and proprietary technologies which can threaten cultural discourse and innovation.

Open Video is the growing movement for transparency, interoperability, and participation in online video. These qualities provide more fertile ground for bottom-up innovation and greater protection for free speech online. Many organizations are already taking steps to change the nature of video on the web: Mozilla is moving to support open video formats in Firefox, the Participatory Culture Foundation promotes open source and standards in video publishing and distribution, and Wikipedia has increased its focus on the open Theora codec.

Yet Open Video is more than just having a functional open source video codec. It’s all the legal and social norms surrounding online video. It’s the ability to attach the license of your choice to videos you publish. It’s about media consolidation, aggregation, and decentralization. It’s about fair use. “

The submission deadline is March 19th.  The conference is a co-production of the Yale Law School Information Society Project, the Participatory Culture Foundation, Kaltura, and iCommons.

Ingres announced availability of the Ingres Icebreaker Enterprise Content Management software appliance today.  The software appliance delivers “just enough operating system” (JeOS) from rPath along with the Ingress Database and Alfresco’s ECM software.

This is an excellent move on the part of Ingres, Alfresco and rPath.  While some disagree with me, I’m beginning to think that customers will increasingly shift from purchasing individual support agreements to purchasing from consolidators.  I previously blogged about OpenLogic as an example of a support consolidator.  Now we’re seeing Ingres play the role of a solution-based consolidator.  Instead of acquiring support for Alfresco, Ingres Database and a Linux operating system, and having to deal with 3 vendors when a bug arises, customers get one product, with one point of contact. Delivering this software appliance on top of a blade server is the next step towards making it easier for enterprises to purchase open source solutions.

What do you think? Does this software appliance make you more willing to try the ECM solution?

Fast Company recently published a list of the “world’s most innovative companies”. Funny enough, the #1 company is really not a “company”; it’s “Team Obama”. Bygones.

IT vendors like Google (#2), Apple (#4), Cisco (#5), Intel (#6), Amazon (#9), HP (#12), IBM (#19), and Microsoft (#34) are in the top 50. However, there isn’t one open source vendor listed. The closest oss vendor is Sun. Sun made the 2008 list, and hence is included in the “33 companies from last year’s Fast Company 50 that didn’t make the list this time but deserve watching”.

Thoughts?

I just read this interesting Forbes article (via Ostatic) about individuals contributing more to open source than institutions.  First off, the article is not talking about whether individuals or OSS institutions (aka vendors) create open source code.  Everyone accepts that key open source projects are no longer the purview of a few interested souls on evenings and weekends.  Rather, these projects are vendor driven and the majority (~99.99%) of content created by the project is done so by an employee of the vendor behind the given project.

The Forbes article talks about contributions that come into projects from outside the vendor.  Most of these contributions will be bug fixes, new feature request or in rare cases, actual code for new functions.  Based on info from Alfresco and Eclipse, Forbes asks why this user contributions come from individuals and not corporate intuitions.  Matt is quoted:

“Asay explained that huge companies and large systems integrators were using Alfresco for large-scale development projects in which tens of millions of dollars were being invested. But while the individual users of Alfresco regularly send back contributions of bug fixes and feature suggestions, the company rarely hears from institutional users, whether they used the licensed version or not.”

The problem is that institutions are not yet familiar with how to contribute code outside their enterprise while retaining their IP rights and minimizing any real or perceived risk.  Gartner suggests that these companies need an “Open Source Policy“.  I haven’t read the Gartner article but I suspect that a part of this policy would discuss how an employee can interact with, and contribute to, an open source project.  Working at a large organization, I know the importance of policies (and yes, IBM has a policy and process for interacting with open source projects).

Without these policies, I suspect that employees at these institutions using open source are simply contributing to the projects “on their own time”.  That Alfresco gets bug fixes from individuals but not institutional users is proof of this to me.  Alfresco is not software that you or I would use for personal purposes.  When an “individual” is contributing fixes to Alfresco, that individual is very likely part of an institution that is using Alfresco in some fashion during his/her day job.

Institutions would be more willing to contribute to open source projects if there were a handful of “standard” contribution policy templates that institutions could adopt.

What do you think?  Does your company have a contribution policy?

OpenLogic shared some of their year end stats with us that I found interesting.  OpenLogic reported that the following projects drove the most customer interest:

• Tomcat
• Apache HTTP Server
• JBoss
• MySQL
• PostgreSQL

OpenLogic also reported that their average customer acquires support for an average of 66 open source projects. They also reported a significant uptick in interest for open source software support during the second half of 2008:

• OpenLogic saw a 2x increase in the number OpenLogic enterprise customers in 2008 (to a total of 75 enterprise customer); over half of these new customers coming on board in the second half of 2008.
• OpenLogic received more than triple the number of inbound sales leads in the second half of 2008 over the first half of 2008.  More than half of all new deals came from inbound leads from enterprises actively shopping for open source support.

I find it very interesting that customers turn to OpenLogic for support of JBoss, MySQL or PostgreSQL, and even Tomcat for that matter, when the vendors behind these projects offer their own support or enterprise editions.  This goes back to the argument that SpringSource’s Rod Johnson was making about customers only acquiring support from vendors who have substantive control over the code in question.  But if the average customer is getting support for 66 projects, that customer definitely wants a one-stop-shop.  Managing 66 vendor relationships is nobody’s idea of fun!

I get the sense that having access to the developer of the project is no longer the #1 buying criteria for some customers seeking OSS support.  This was bound to happen with the maturity/quality of open source projects and customer’s familiarity with these projects.  Think of the 15% of customers that purchase support and decide not to renew in year 2 because they didn’t make sufficient use of the support offering.  Play this out over several products and it’s easy to see the value of a combined support offering like OpenLogic’s offerings.

So does this mean that the MySQL/Sun’s of the world are going to lose revenue at the hands of the OpenLogic’s of the world?  Nope.  Customers do value the relationship with the vendors behind key open source projects.  I emphasize “key” because out of the 10s or 100s of OSS products in use at a customer, a handful are more critical than the rest.  For example, I want to be sure that the HTTP server or database stays up as these products impact customer experience.  If I keep getting an error with Ant, it’s annoying for my developers, but doesn’t impact the customer experience.  There will always be different use cases and customers have differing needs.  It behooves the open source ecosystem to address these needs as broadly as possible.

Peter Galli over at Port25 is reporting:

“Microsoft and Red Hat announced this morning that they have recently signed agreements to test and validate their server operating systems running on one another’s hypervisors.”

Co-opetition being as it is in the IT industry, this is an excellent move for both vendors. Supporting customer choice continues to be a winning strategy:

“”This means that those customers with valid support agreements will be able to run these validated configurations and receive joint technical support for running Windows Server on Red Hat Enterprise virtualization, and for running Red Hat Enterprise Linux on Windows Server 2008 Hyper-V or Hyper-V Server 2008,” Neil says.

So, while Microsoft and Red Hat will continue to compete, customers have asked us to work together on technical support for server virtualization. These agreements respond to that request by giving them a new level of integration between Red Hat Enterprise Linux and Windows Server for their heterogeneous IT environments.”

It really is great to see Microsoft opening up to opportunities like this.  Whatever its history with open source, Microsoft keeps showing that they can evolve their thinking to remain relevant to its customer base.

Very cool – now when will this happen ;-)

Earlier this year, Matt wrote that, in 2009, companies will balk at long and costly implementation cycles for enterprise ERP products and turn to open source ERP, especially in the current economy.  I’m not completely convinced that enterprises spending on “18-month, multimillion-dollar ERP rollouts” will turn to open source ERP.  A company was looking at a multi-million dollar Oracle or SAP implementation is probably large enough (read: risk adverse enough) to stay the course.  Projects may not progress as quickly as previously planned.

On the other hand, there is an absolutely large opportunity for Open Source ERP in the SMB market.  Companies that have 50-2000 employees often need more than Quickbooks, MS Excel and roll-your-own apps.  This is where Openbravo and Compiere have a role to play.  Along these lines, I spoke to Sujee Saparamandu, the CEO and founder of OrangeHRM, an HR management software vendor.

Think of OrangeHRM as the PeopleSoft of open source ERP software. OrangeHRM has over 200,000 downloads, but more importantly, over 700 paying customers.  These customers have an average employee size of 50-100 and the largest customer has 2,000 employees.  OrangeHRM appears to be business model agnostic.  OrangeHRM offers their product under the GPL license which leverages the open source distribution model. From here, customers can choose to pay for support as is common with a support-based oss business model.  Or, customers can choose to pay to access a hosted version, thereby leveraging the SaaS business model that other oss ERP vendors utilize.  And lastly, OrangeHRM sells proprietary add-ons around the open source OrangeHRM package, which we know as the Open Core business model.

I told Sujee that it’s a good idea for a company like his to offer multiple ways for customers to gain value from OrangeHRM.  Over time, one of these models may become the major revenue driver and require special focus.  But for now, broader choice is a better choice.

Matt reports that MySQL’s Marten Mickos is leaving Sun.

It seems that the bureaucracy of a larger company didn’t fit well with Marten’s goals.  I’ve seen it happen several times with acquisitions at IBM.  And yet, some of these have been extremely successful.

Yes, this is a blow to Sun and their open source street cred.  But it’s not insurmountable.

Marc Fleury and a long list of key executives left Red Hat after the acquisition of JBoss.  And yet JBoss is still performing well for Red Hat.  As Marten mentions, he is:

“now stepping aside to let the strong managers of the group take over and continue the ambitious business ramp-up.”

This is exactly what happened at Red Hat (under Sacha’s leadership), and I’m expecting it’ll happen at Sun.

Having said all this, losing Marten is unfortunate.  He was a rational and highly pragmatic thinker in the open source arena.  His insights have shaped my thinking.  The open source movement is better off for Marten’s efforts.

Best wishes to him and to MySQL/Sun.

Matt Aslett had a great post about defining an “open source vendor” earlier this week.  The post was in response to OpenNMS CEO Tarus Balog, who didn’t agree with Matt’s use of the term “open source vendor” to describe companies with “open core” licensing strategies.  Tarus argued:

“Hyperic IQ is commercial software. There is no open version, and it’s published under a closed license (at least from what I can gather). Both Hyperic and JasperSoft are ‘open core’ companies – commercial software companies who use open source to market their software. So it seems a little ironic that you would preface this news, where they don’t even make the pretense of having a ‘community’ version, by calling these companies ‘open source vendors’.”

To which Matt replied:

“In writing these reports I made the decision that for me an “open source vendor” is one that is reliant on open source software in order for its business strategy to function. By that definition we would include open source purists such as OpenNMS but also the likes of Hyperic and JasperSoft, which would have very different commercial business strategies were their products not based on open source.”

Matt is clear that this is his definition of an open source vendor because, at this time, there isn’t a universally approved definition.

To which Tarus replied:

“It’s seems from your definition both Google and Amazon would qualify as “open source vendors” since their business model would not be possible at all without open source.”

I started this post thinking I would take Matt’s definition and alter it slightly to ensure that companies like Google and Amazon are not consider open source vendors.  And here’s where I ended up:

An “open source vendor” is one that develops, contributes to and distributes open source licensed products, which are integral to driving its revenue.

By this definition, Google and Amazon are not open source vendors.  Open source software is not integral to driving Google or Amazon’s revenue.  It is absolutely critical to Google and Amazon’s business models and cost structure, but not revenue.

For all of Microsoft’s open source actions, they do not distribute open source licensed products that are integral to its revenue.  Contrast this with vendors that utilize the open core business model, such as Hyperic or JasperSoft.  Hyperic’s revenue from Hyperic HQ Enterprise or Hyperic IQ is expressly linked to the customers they’ve secured with their open source licensed Hyperic HQ Open Source product.

The problem with this definition is how it applies to companies such as IBM, HP and Accenture.  They all contribute to and distribute Linux and other open source in order to generate revenues from servers or implementation services.   As such, they would be considered “open source vendors” by my definition.  Sigh…I guess I shouldn’t apply for that job at Webster’s.

What do you think; can you improve on the definition?