Unleash the Borg: OS/400 Gets Autonomic Tooling

by Alex Woodie

Your dream of building an army of self-aware, self-healing iSeries servers is drawing nigh: IBM recently added key autonomic runtime capabilities to its prized OS/400 server. With the release of Autonomic Toolkit 2.0, IBM added OS/400 and Solaris to the list of operating systems that support the development of autonomic systems that can automatically take action to counter specific problems and events.

IBM launched its autonomic computing initiative several years ago as an integral component of its on-demand strategy to deliver computing as a utility. The goal of this initiative is to build self-reactive computer systems that automatically adapt to changes in themselves and their environments, in much the same way that the human autonomic nervous system continuously makes the right adjustments in our organs and bodies without requiring any conscious thought or effort on our part.

For years, systems management vendors have been helping IT shops to reduce or eliminate the need to have humans administer to their computers 24/7. You don't need to look any further than Help/Systems and its Robot suite of OS/400 tools for an example of how some AS/400 shops have been able to go "lights out" in their data center operations. But IBM's cross-platform undertaking to create self-configuring, self-healing, self-optimizing, and self-protecting systems is quite a bit more ambitious in its scope.

Much of IBM's autonomic technology is available in a series of freely downloadable components called the Autonomic Toolkit. There are 12 pieces to the toolkit, and they can be downloaded from IBM's Web site.

Let's take a closer look at the Autonomic Toolkit's major components.

Autonomic Management Engine is the core runtime component that IBM brought to OS/400 V5R3 and Solaris 8 with Version 2.0 of the toolkit (it also supports Windows Server 2003, AIX, and Linux on X86 and zSeries). The AME forms the crux of an Autonomic Toolkit implementation, as it is in charge of monitoring system resources, sending events to the "Integrated Solutions Console," and taking corrective action. The AME gains the intelligence needed to take action through "resource models" that contain specific metrics, events, thresholds, and parameters used to determine the health of the system and which corrective actions to take.

Integrated Solutions Console is a WebSphere Portal-based system that now runs on OS/400, AIX, Solaris, Linux, and Windows, and allows administrators to interact with autonomic runtime components and to induce failures to train the AMEs how to act.

Generic Log Adapter (GLA) Runtime and Rule Sets: The GLA takes event log files from OS/400, AIX, Solaris, Linux, and Windows systems and converts them into the Common Base Event data format. (See below for more information about the Common Base Event format.)

GLA Configuration Editor and Log and Trace Analyzer (LTA) plug-ins: The GLA Configuration Editor is an Eclipse plug-in that works with the GLA Runtime to generate rule sets for different kinds of log files. The LTA plug-in is for viewing and correlating log files for debugging autonomic systems, particularly multi-tiered systems.

Remote Agent Controller runs on all supported platforms and works with the Log and Trace Analyzer to import any type of log file remotely from any platform, making it easier to analyze a log file from a remote machine.

Resource Model Builder is a Windows application used to define the rules that your Autonomic Management Engine will follow in carrying out your autonomic will when you're not there.

Eclipse Software Development Kit includes the Eclipse development framework and assorted Java development tools, and runs on Windows or Linux PCs.

In addition to these major components, IBM provides several downloadable "scenarios" that provide users with guidelines and assistance in deploying autonomic systems, including the problem determination scenario, and several solution installation and deployment scenarios, using installation scripts from InstallShield and InstallAnywhere.

Two OS/400 PTFs, SF99269 and MF34019, are also required before iSeries shops start playing with the toolkit.

COMMON BASE EVENT

Before anybody can start building autonomic capabilities into their systems, they need to have a good understanding of how their computers communicate when they detect a problem. However, one of the stumbling blocks to creating a general, cross-platform autonomic methodology is the wide variety of message log formats in use around the world.

To alleviate the differences in message log formats, IBM researchers came up with a standardized, XML-based message format called the Common Base Event. This really should be considered one of the crowning achievements of IBM's autonomic computing initiative and deserves some attention and discussion.

The Common Base Event format is the distillation of analysis of hundreds or thousands of different messages types from IBM's systems, as well as those of its competitors. IBM boiled all these message types down into 12 key ingredients, or "situation types," which include starting and stopping events; requesting the status of events; initiating connections between components; reporting a services state of availability; and several more. IBM also came up with the Generic Log Adapter to provide a way to translate between the various message log formats and its Common Base Event format. For an in-depth explanation on the Common Base Event data format, see the article "Standardize Messages with the Common Base Event Model," on IBM's developerWorks Web site.

So there you have it, a series of tools for helping your OS/400, Windows, Linux, and Unix servers learn to become self-sufficient. Do you hear that? Your iSeries is calling, and it wants you to download Autonomic Toolkit Version 2.0 right now. After all, resistance is futile.