Enterprise Level Open Source Applications

Their project management package is called OpenProj. Projity's revenue-creating part of their business is making a web-based project management tool available for a monthly subscription fee. The fee is a little steep... $8/month for users, and $20/month for project managers. That's $96 and $240 per year, respectively, which doesn't compare well to simply buying software. The functionality of the tool is pretty close to Microsoft Project of 10 years ago, which will serve most basic needs, but it remains to be seen how it performs with projects that have hundreds of tasks. Also, modern project management packages allow multiple project managers to roll up their projects to larger superprojects, with access control lists and other features not found in this simple web-based package. Over time though, this should improve.

OpenProj provides this software for free. Clearly they are hoping people will help develop features to make their commercial product better; after all, free labor fits most budgets. There are some caveats to the license however, as according to some people on the web you have to display the Projity logo and a link back to the website at all times. This make it "badgeware".

Many productivity applications are becoming available in the software-as-a-service model, and some are getting rolled up into Google's offerings. It wouldn't surprise me to see Google acquire Projity or a similar company to add a project management tool to their suite.

Last week Google decided to contribute MySQL code. It's rare that you see a commercial entity giving back code to the open source efforts, but rarer still to see it coming from a giant like Google that really knows what they are doing when it comes to large scale information systems. Check out the story at Information Week.

Basically, Google addded or improved features for replication and failover, which just makes MySQL that much more reliable and comparable to is proprietary counterparts. If Google can use MySQL for production databases, then it really should be acceptable for many organizations, if not most.

If you're a german company, you may want to check out Synerpy, a new open source CRM product that claims to have 40% of the functionality of the far more mature SAP suite. SAP isn't seriously threatened by upstarts like this, of course, but these open source products have a tendency to gain roots in places where SAP is too expensive. And then they grow and grow. This will be worth keeping an eye on, particularly if they make an english version.

SGI has announced the death of Irix (SGI's version of UNIX) and the MIPS platform. This means that Linux has effectively killed off yet another version of UNIX. Many years ago, the market was flooded with many different versions of UNIX, and they were considered "Open" because they used the same APIs and could be easily networked together. They were, for the most part, not binary compatible, as most of them ran on different hardware. And the openness of the architecture did not translate to open source; no, the source code was proprietary, although it was licensed out to so many companies as to be widely available.

When Linux came on the scene, free and open, I knew it would be the death knell for all the many flavors of UNIX. And for the most part, this has certainly come true. Remember Sequent's Dynix? Data General's DG/UX? Fujitsu-Siemen's Reliant? Convex Computer's ConvexOS? Interix? Ultrix? The BSD-based SunOS? OSF/1, er, Digital Unix er, Tru64 UNIX? Apple's A/UX? Unicos? QNX? Coherent? gasp!!

What happened? Intel's architecture basically won out on price, performance, and general market dominance. Linux, being free, meant that servers that used to cost hundreds of thousands of dollars were now available to any person with a $200 PC and the inclination. Following this logic, the remaining UNIX es should likewise be endangered.

HP, with its HP-UX dumped the PA-RISC architecture to switch to Itanium, is finding many customers asking why they don't just run Linux on Itanium. It's a valid point. When SGI started selling Linux on Itanium, their own UNIX (Irix) died. HP sells Linux on Itanium. The death of HP-UX could be next; they were certainly comfortable dumping Tru64 Unix.

IBM has AIX, which runs on the POWER architecture. They also sell Linux on this architecture, although most people running Linux would rather go for cheaper and faster Intel boxes. This has caused IBM to greatly reduce the price of their Power5 systems. IBM holds on to operating systems for a very long time, so AIX may survive.

Sun has Solaris, which runs on the SPARC architecture. Sun doesn't sell Linux on SPARC, nor promotes it. However, Sun also has Solaris on the Intel/AMD architecture, and sells AMD-based systems. These systems also run Windows and Linux. This puts Sun in a strange position. Their SPARC systems are becoming specialized, focused on 8-core systems that run on low power and can support gobs of threads. Perfect for their thread-thirsty J2EE application architecture. But, this greatly sacrifices single thread performance and SMP capability, which greatly limits the scalability.

Apple has MacOS, a derivative of NextStep, but they are so focus on form over function and style over scalability that no one considers them a serious player in the server space. They practically hide the fact that they're a UNIX derivative. Still, I don't see Apple switching to Linux, as they are the opposite of open; they're more proprietary than Windows, since they control the software AND the hardware.

And... that's about it. That's all that's left of UNIX today. Every year, Linux continues to improve, and it continues to gain wide acceptance in large enterprises. Other UNIX versions will become increasingly irrelevant over time.

It is true that PHP has an easy learning curve. Wayne Duquaine from Grandview Systems says that IBM will be coming out with PHP for the mainframe environment in the next few months, and that his company will be making a plugin for it so it can run under CICS.

IBM is most likely partnering with Zend to create a version of PHP and Zend Core for z/OS. Zend has already done this for Linux and AIX (IBM's UNIX).

Mainframe developers should certainly learn skills beyond COBOL, and PHP is a good choice. It's a very powerful language, used in a lot of places, and it's very easy to learn. There may be better job opportunities for Java developers, but J2EE is an overly complex language, and a lot harder to learn than PHP. Actually in some respects, very large Sun servers are the new mainframe. In the 1990s, Sun servers were starting to replace jobs on the mainframe. Today, Linux servers are replacing Sun servers. Could it be that in 10-15 years, Java developers and their overly complex J2EE will be looked upon with contempt, much like COBOL is today?

SplendidCRM is an open source CRM offering available for Linux and Windows. By using the Linux .NET tool Mono, they were able to build the same product on both systems. Actually, they built the product on Windows, using the rich tools available for .NET development, and were able to copy the code over to Linux by leveraging Mono. On the backend, they support SQL Server, Oracle, DB/2, and MySQL.

All they need now is for a tool that will allow SQL Server to run (efficiently) on Linux. Now that would be interesting.

SplendidCRM is a young company based in the Raleigh/RTP area. They are also a Microsoft Certified Partner, and I'm sure Microsoft will tout them as an example of open source applications on the Windows platform. But of course, they won't tout them too much, for fear of competing with their own CRM offering in the Microsoft Dynamics line.

MySQL continues to gain credibility as its client list grows. As reported by Enterprise Open Source Magazine, the list of "web 2.0" sites that use MySQL on their backend now includes YouTube, Flickr, Habbo Hotel, Linden Labs, CyWorld, Technorati, Facebook, FeedBurner, Feedster, Wikipedia, Digg, LiveJournal, Mixi.jp, SimpleStar, PhotoBucket, 37signals, del.icio.us, Trulia, Neopets, and Zimbra. Wow. That's quite a list. And that's in addition to big "web 1.0" sites Craigslist, Google, and Yahoo! that also use MySQL. Of course, one always has to take these claims with a grain of salt; it is not obvious how much MySQL is used in these companies. In some instances, it might not even by the primary database. Google, for example, has an in-house proprietary system (the Google File System) to manage most of their data and indexes.

But there can be no doubt that MySQL is being heavily used by some major names now. The LAMP approach (Linux, Apache, MySQL, PHP) is a very common stack for young startups, as it delivers a lot of power for very little cost. And the number of people with skillsets in this stack is increasing all the time.

EnterpriseDB has really come a long way since starting up not too long ago. Now they have gained some serious credibility with a partnership with Sun Microsystems. Recently, Sun decided to start including PostgreSQL software with Solaris 10, as of the Solaris 10 06/06 release. Evidently, Sun felt they did not have enough resources internally to support the product, so they have selected EnterpriseDB to handle the overflow of tech support cases, as well as provide training and consulting services. Makes sense. Why EnterpriseDB over Greenplum, though? Maybe EnterpriseDB, being the newer company, decided to be the lower cost provider. Or maybe EnterpriseDB and Sun have a closer relationship than what has been revealed so far. Hard to say, although you have to figure this move has to cause some discomfort to Sun's relationship with Oracle. Of course, Oracle lately has been pushing Linux over Solaris, so maybe EnterpriseDB and their "drop-in replacement" for Oracle are a good choice to counter.

Back in January of 2005, Pervasive decided to market PostgreSQL. This seemed like a good idea at the time, as PostgreSQL has steadily been gaining adherents over the past few years. This week though, Pervasive decided they could not add anything meaningful to the PostgreSQL world, and have decided to pull out. Apparently, the great amount of support available from the open source community exceeded anything Pervasive as prepared to offer.

This is a little disturbing, as PostgreSQL still needs commercial-grade support. Perhaps the support coming from Sun Microsystems, GreenPlum, EnterpriseDB, and others was too competitive. Sun, in particular, has started to include PostgreSQL in Solaris. Certainly, we never saw PostgreSQL making any impact on Pervasive's quarterly reports. Hopefully, PostgreSQL will attract new champions.

Compiere is making some very nice progress. They recently secured $6M in new funding from their venture capital backers, New Enterprise Associates. Being a small, 10-person company, $6M definitely gives them a lot of breathing room. They are also relocating their headquarters from Portland to Silicon Valley; it's not clear why they're doing this, as Portland has quite a large open source community. But maybe the pool of available talent is fully tapped there, and they are seeking the mythical endless spring of talent that people expect in the valley.

It should also be noted that Compiere's website is looking a lot better than last year; their product offerings are maturing, and getting more and more competitive all the time. Like SugarCRM, being free makes their offering very attractive to a lot of small companies, which opens up opportunity for implementation and integration consultants. It wouldn't surprise me at all if some commercial company takes an interest in partnering with, or acquiring Compiere before long.

RedHat appears to be buying JBoss for peanuts. This acquisition takes them well past the operating system space, as JBoss is one of the leading application servers out there. Hell, some surveys actually put it in first place, ahead of IBM's Websphere and BEA's WebLogic. (We use WebLogic where I work.) This is huge!



Recently, Oracle bought InnoDB and Sleepycat, a couple of open-source database technology companies, and it was said that Oracle was hoping to purchase JBoss and Zend as well. And there was another opinion article in InfoWorld this morning saying Novell should buy JBoss. Sorry, Novell!!

Zend is still out there, though. Zend is the leading company for PHP tools and support.

Now that RedHat owns JBoss, will it cease to be free? Will the source code fork into a free version and an "enterprise" version? Hmmm...

It sure is interesting that all these open source technology companies are getting snapped up now. I suspect it's because they can't earn enough money to stay independent.

Although some people foolishly believe that Oracle means no harm to MySQL through its purchases of InnoBase and Sleepycat, it is pretty clear that they're trying to constrain MySQL's ability to penetrate Oracle's markets. First Oracle bought InnoBase, MySQL's transaction-oriented database underpinings. (Typically MySQL uses either ISAM, which is flat-files, or InnoBase.) Sleepycat, with its Berkeley DB, was considered to be a good alternative if InnoDB went away; so last month, Oracle bought Sleepycat as well. One has to assume that Oracle approached MySQL AB earlier about a purchase and they turned them down; so Oracle is buying MySQL AB's suppliers instead.

Meanwhile, Oracle is trying to purcase Zend, the leading developers of PHP tools, and JBoss, the very popular java application server. Why is Oracle buying open source companies? Clearly, these open source companies pose a threat to Oracle's business model. Why pay hundreds of thousands of dollars for Oracle licenses if you can get similar, functioning, acceptable software for free? Especially as the open source products improve and become equal to, or better than, their proprietary competitors. Oracle is definitely not going to sit idly by and watch this happen, so they are taking action. Buying these companies is not a permanent solution, as the open source community can merely fork the code and start a new organization. These efforts take a lot of time, however, and Oracle will continue to rack up license sales while the fear, uncertainty, and doubt continues over MySQL's future.

Maybe at some point in the future, Oracle will develop these open source products into their own businesses, using the service-oriented business model (selling the support intead of the license). And maybe Oracle will become a leader in open source efforts. After all, Oracle's real competition is not MySQL, it's Microsoft SQL Server and IBM DB/2 and Sybase ASE. But given that Oracle's license fees are the highest in the industry, I would say they have the most to lose from these efforts, and it should be someone like IBM that is pushing the open source model. IBM is far more comfortable selling services, as well as hardware to run the databases... they are best positioned to do it right. So I find it more likely that Oracle will not pursue the open source options as a real business. They'll just keep buying the companies to keep the open source community on the defensive.

There is a fascinating article over at eWeek where they have interviewed E*Trade's VP of Architecture, Lee Thompson. It describes what happened when E*Trade decided to convert from clunky old Sun Enterprise 4500 systems to newer 1U-based Linux boxes. They achieved a large cost savings over the Sun systems, which is not at all surprising given how expensive the E4500s were. It would be interesting to see how their current Linux systems measure up to the new boxes Sun has, such as the new Sparc IV systems, the dual-core Opteron systems, and the Sparc T1 systems coming out this spring. The interview also talks about how E*Trade is adopting open-source software development methodologies, where few people have the rights to commit code, which reduces the problems encountered when too many people are updating the same section of code. It's good stuff, and rare to see real tech people talk about their proprietary systems in public.

A few startups and organizations are trying to push open source technologies by declaring support for a set of interoperable products, or "stack". Open source software is free, but supporting it can be a challenge. By endorsing and selling support for a specific stack, companies create a revenue model around the free products, and clients get a set of products they can be assured will work together cleanly. Initially, the most common combination of open source products was LAMP -- Linux, Apache, MySQL, and PHP. But then products like ImageMagick and other tools entered the picture, and some of these products created conflicts. Sometimes different products would require different versions of other products. Next thing you know, you've got four different versions of a compiler on your system, and the stack still won't compile cleanly. So companies like SpikeSource were created to address this issue.

Now it appears that the Government Services Administration is creating its own stack, Open Source eGov Reference Architecture, or OSERA (a product of the GSA's Office of the CIO). The appeal of OSERA is that it automatically gets a very large client: the GSA. This could lead to adoption of OSERA in thousands of government departments and related organizations across the country. Sometimes, a government-adopted standard becomes a mainstream de facto standard. For example, there is some cachet in products that are "milspec", or made to military specifications. While other "stack" support companies battle it out over what makes the perfect stack, the GSA could be building a very large user base. Keep an eye on OSERA, chances are it may have a significant impact on the stack in the years to come.

It will also greatly affect, for better or worse, the companies and products that are included and excluded from the OSERA stack. Gnome or KDE? PHP or Python? Struts or Rails? Definitely worth watching.

Oracle's purchase of Innobase was very suspicious, to say the least, and many feared it was an attack on MySQL, which depends on Innobase. MySQL's reaction was muted, and consisted of mumblings about how they could easily replace the Innobase data store with an alternative. Others suggested that Innobase itself was based on open source, and MySQL could simply fork the source and make their own version.

Today however, others are suggesting that Oracle would actually want to support MySQL, as the open source database is more of a threat to Microsoft SQL Server and Sybase ASE than it is to Oracle. Both Sybase and Microsoft compete on the low end, offering an affordable alternative to Oracle, which currently is far too expensive for most companies to use. Most companies also do not need the extensive feature set that Oracle offers, and find the alternatives to be more than good enough for their needs. MySQL has been chipping away at market share from the lower end, so it makes sense that MySQL is more of a threat to Microsoft and Sybase than it is to Oracle and IBM.

This is not to say that Oracle will end up buying MySQL AB; that just seems far too unlikely. MySQL values their freedom and runs a much different business model than Oracle does. Also MySQL would lose its credibility and coolness if it joined with Oracle, and people would start to use PostgreSQL more.