Open Management Consortium to Drive Tool Integration
Published: May 15, 2006
by Timothy Prickett Morgan
Last week, after The Linux Beacon went to press, a group of software companies who sponsor a diverse collection of open source software tools that aim to make the managing of Linux servers--and in some cases, other platforms, too--a lot easier ganged up and formed a consortium that they hope will make open source tools work better together and, therefore, make open source platforms like Linux more appealing to businesses of all sizes.
According to Mark Hinkle, vice president of strategy at Emu Software, one of the founding companies behind the Open Management Consortium, this consortium is a little bit different from the typical consortia we see founded in the IT industry all the time. "A lot of consortia are designed with grander plans and are started with a big agenda," says Hinkle. "But the first thing we want to do is coordinate and collaborate. Then, after we figure out how to work together, we can talk about putting together an agenda. If we started out with an agenda, the problem is we would only end up with three quarters of the software stack we need."
The founding open source systems management software projects behind the OMC and their related sponsors include:
In many minds, open source is perhaps a necessary condition for success in the modern software world, but it is by no means sufficient. Being open source has been an important aspect of the success of the Linux platform, which has become the number three server operating system in the world and will soon be the only one that is exhibiting growth--Unix is flat to down, and Windows is starting to flatten, others are in decline. But perhaps more importantly, Linux became the de facto platform for running the Apache Web server during the dot-com boom a decade ago and then, as Web programming became more sophisticated, Perl, PHP, and Python were integrated with Apache on the front end and the MySQL database as a backend data store, to create the LAMP stack. That stack is the reason why Red Hat went public in 1999 and why Novell bought SUSE in 2003. These companies are still focused on building out that stack and services to support it. Why? Because that is where the money is.
What the founder members of the OMC want to do is a similar thing: Create an open source systems management software stack, one that is as flexible and as standardized as the LAMP stack has been. "The genius of the LAMP stack is that all four levels are replaceable," explains Hinkle. If you want to run Apache and the three Ps of development on Windows or Unix, go right ahead, it is supported. If you want to replace MySQL with another database, you can do that. If you want to use another Web server, or plug in a Java development environment, feel free. "We're going to push a central systems management stack, but we also want the ability to plug something else in stack and not break it."
And for good reason, since there is some overlap in what the software created by the members of the OMC do. What the OMC members want to do is create a stack of software that works together seamlessly that can be used to:
- Provision and patch servers
- Configure and manage applications that ride on top of the servers
- Monitor the complete software stack
- Wrap the whole shebang in tight security
Hinkle says that in the week since the announcement, hundreds of companies (those behind open source projects, those who make systems and servers, and those who use all this stuff) have signed up to participate in the consortium. And now, the founding members have to sift through all of that interest to find out who is really going to contribute actively, and if so, in what way. After that is done, the key players in the consortium will probably share their respective product roadmaps and then try to figure out how to better integrate their tools.
A few obvious standards that these players might agree to come immediately to mind. For instance, the way that these various tools pass messages around might be standardized, or the formats they store their internal data in might be standardized so one tool can read the configuration and system information created by the other.
While this is all a necessary first step, in the long run, OMC will likely be the forge for a sophisticated systems management toolset. The members of the OMC took great pains to not position themselves against IBM's Tivoli, CA's Unicenter, and Hewlett-Packard's OpenView tools, which are the big three in systems management. "Replacement is not the goal--but integration with existing tools is," said Hinkle. While this is true, small and medium businesses don't use these big system management frameworks. They pay administrators to use all kinds of tools and write scripts and baby-sit machines. And what small companies in particular need is a set of cheap, supported, and easy-to-use system management tools that provide some of the functionality of Tivoli, Unicenter, or OpenView, but are not overkill. If the members of the OMC can work together to integrate their tools--and to share in the spoils that will undoubtedly result from such integration--then there is a good chance SMBs will finally get decent system management tools.