SOMA: Service-oriented management architecture
By Craig Wassenberg
,
Network World
, 03/02/2007
This vendor-written tech primer has been edited by Network World to eliminate product promotion, but readers should note it will likely favor the submitter's approach.
- Share/Email
- Tweet This
- Print
The problems network managers face in today's data center are increasingly complex. Changes in network usage models require
the remapping of traffic flows. New intrusion threats demand stronger security measures. Using virtualization to migrate workloads across physical resources requires constant reallocation of network resources.
The list goes on. In response, systems management vendors are constructing their solutions differently from the way they used
to.
The new approach to network management - and systems management generally - is based on service-oriented management architecture (SOMA). Vendors are building SOMA-based
network management appliances, agents and proxy agents that are flexible and capable of evolving as management needs evolve.
Adaptability was limited in past network management solutions, but it's largely a given with SOMA.
Under SOMA, all important management operations are implemented as services (such as for retrieving device status, for device
control, for changing configuration settings and for provisioning). Each service is a software component with a formally defined,
message-based, request-response interface. The business logic behind each interface is hidden from users. Messages are in
XML and are passed among services running within a device via a management services bus (MSB). Programming a management application
or an agent is relatively easy because all available management functionality is exposed via consistent interfaces, and most
services are highly reusable.
When new services are required, implementers can write new code or encapsulate and integrate commercial or open source legacy-management
code: Either way, the code is hidden behind formal service interfaces. Used this way, SOMA unifies what otherwise would be
disparate management solutions.
Services provide simple or complex functionality. A simple service might, for example, return a device's current temperature
settings or fan speeds, and a more complex service might perform complex diagnostics requiring the correlation of information
from multiple sensors and internal event logs. Services can cooperate with one another, and more sophisticated services can
be formed by layering atop lower-level services. A vendor can, for example, provide management agents for an entire gamut
of routers, from low-end to high-end, by picking appropriate services from a services library.
Partner Content
Blue Stripe Software
www.bluestripe.com/
Improving Application Performance Troubleshooting
Diagnosing why an application is slow is hard, at times taking days or weeks to isolate and resolve. This paper explains the challenges involved using current management tools, provides a 'wish list' for application management and analysis, and explains the need for an application system-wide approach that monitors entire applications, not components.
Download Whitepaper
Virtual Vigilance: Managing Application Performance in Virtual Environments
This paper highlights the impact of virtualization on application performance. "Managing Application Performance in Virtual Environments" states: "Best-in-Class organizations are predominately taking actions around improving visibility across both physical and virtual systems, assessing the business impact of application performance and understanding interdependencies of applications in virtualized environments."
Download Whitepaper
Application Service Requests: The Missing Link for Pragmatic ITSM
Forrester Research analyst Glenn O'Donnell and BlueStripe co-founder Vic Nyman discuss a breakthrough approach to application problem management. Learn the new approach for ITSM problem management, which provides: Rapid isolation of application slow-downs to specific components for quick problem resolution, 24/7 monitoring for proactive notification of potential issues before end users are impacted and much more.
Register for Webcast
Comment