Approved by Oasis earlier this year, WSDM was sponsored by the leaders in systems management IBM, HP, Computer Associates, and BMC and backed by Actional, BEA Software, Dell, Fujitsu, Hitachi, Novell and Tibco.
It goes up against WS-Management, a rival spec driven by the unlikely coalition of Microsoft, Sun, Intel, AMD, and fence-sitter Dell, which has not been formally submitted to any standards group.
Both proposals provide a framework for using web services for managing IT systems. WSDM includes two major elements: MUWS (Management Using Web Services), which is the heart of the specification, and MOWS (Management of Web Services), which strays outside the core purpose by proposing how to manage web services themselves. By comparison, WS-Management is more focused, because it only addresses the use of web services to manage enterprise systems.
The WSDM roadmap examines how to get to the point where web services can be used to access, control, or automate the processes of enterprise management frameworks such as HP OpenView, IBM Tivoli, CA Unicenter, BMC Patrol and others.
According to William Vambenepe, distinguished technologist in the CTO Office of HP’s Management Software Business, using web services could provide the missing link between business policy and IT operations that have long been the holy grail of systems management.
With [web services] standard like BPEL [Business Process Execution Language], you could finally blur the line between what is a business management service and an operations service, he said.
For example, as part of an orchestrated service related to order fulfillment, the entry of an extremely large order and the need to make a quick commit for delivery might trigger the use of an advanced planning system to optimize production and distribution to fulfill the order at least cost and impact.
If the advanced planning system required additional computing resources, it could in turn trigger a service request to the IT management system to provision additional server capacity for the advanced planning application, and to check service levels for affected processes.
The roadmap listed what is available today, and what pieces still must fall into place to make web services-based systems management reality. Some of the foundations, such as SOAP, WSDL, and BPEL are approved standards (or close to it), with commercial product available or starting to trickle out.
Other key components in various places in the standards pipeline include WS-Reliable Messaging, which asserts that a message has reached its destination; WS-RF (Resource Framework), which detects the availability of resources such as compute capacity; and WS-Notification, for providing event-triggered alerts.
On the horizon, system deployment descriptors, which describe constraints for installing new applications or hardware infrastructure, are at early stages of standards development. So are the management models offering templates on how to govern resources, some of which are being developed by the Distributed Management Task Force (DMTF).
According to HP’s Vambenepe, the major missing piece is policy management. For instance, while WS-Policy is still in draft stage, it confines itself to specifying how policies are expressed in web services messages.
It does not provide any way to define policy per se. Additionally, virtualization technologies being developed by hardware and OS developers would go a long way towards simplifying systems management in general, and the use of distributed web services to commandeer resources.
Of course, all this is theoretical since the vendor community remains divided on how to sync web services and infrastructure management frameworks. For a change, the IBM-Microsoft coalition that has driven the most successful web services standards thrusts is absent here, which means it will be awhile before any of this becomes more than vision.
