Integration is a core activity in IT management. There are different tools for different technologies (apps, servers, network, etc.) and diffferent functional areas (fault, performance, ticketing, etc) and these tools need to be integrated in order to be able to support IT processes that span technologies and functional areas effectively.
We call RapidOSS "an integration, automation and presentation suite for IT operations management", typically used to integrate management information from disparate management systems and present to users through a simple, unified, consistent web based user interface.
RapidOSS adds to event correlation/management approach by consolidating not only events but all information related to IT operations management from different management systems.
What do we mean by that? ?RapidOSS OpenNMS integration serves as a good example of the differences. With events, the integration typically refers to sending traps from one management system to the other. This is in fact the available integration between OpenNMS and Netcool (or Smarts), where OpenNMS sends traps to Netcool using scriptd. and a Netcool probe processes traps to create/clear events in Netcool.
Yet OpenNMS has a lot more information available that would be useful (and even required) to IT operations folks. For example, an event tells us that an interface is down or utilization is over the threshold. Now what? what device is this? Who does it belong to? Where is it located? What are the services running dependent on this interface? What is "normal" as utilization?, etc.
Just displaying all the events without any of the supporting information necessary to troubleshoot the problem (or even determine whether there is a problem) is inadequate to say the least. As a result operators end up switching from application to application to collect all this information which causes significant productivity loss.
RapidOSS aims to make all the management related information easily available to the users, so bringing the events is just a start. For example, OpenNMS integration (for now) brings inventory data (which devices, interface and services monitored by OpenNMS) into RapidOSS as well as the performance graphs available for each object. Continuing with the example above, when there is an interface down event, the users can see which device the interface is part of, the status of the other interfaces on that device, the services dependent on that interface, and the performance graphs available for that interface or device directly from the RapidOSS UI. Users can also see the event history (has this happened before? how often?) and business data such as location, owner, etc., provided that they have the information in some format (file, in-house database, et.)
In case of implementations with multiple servers and/or multiple products, users don't have to know which product/server is used to monitor which application, device, interface, etc. The relevant information is provided wherever it's needed, on demand.
This first release of RapidOSS OpenNMS plugin already integrates alarms, inventory and performance graphs from OpenNMS into RapidOSS but I think we're just scratching the surface. Our goal was to come up with just enough functionality to be able to demonstrates some of the capabilities, enable people to visualize what the integration should be like and start the discussion. There are some additional potental integration areas, what/whether we do next will depend on the feedback we get from OpenNMS community.
- Maps: RapidOSS has the capability to present the network topology as in maps. The connection information can come from anywhere. Does OpenNMS have topology data (Layer2/3)?
- Timeline: RapidOSS keeps historical events and can present the events using Simile Timeline component. Would showing OpenNMS outage data using the Timeline component make sense?
- BSM: RapidOSS provides the infrastructure to manage business services. The components (devices, interfaces, services) that are monitored by OpenNMS can be used to instrument business services in order to determine impact of problems on business services, etc.
- What else ?
What do you think? Looking forward to hear your thoughts!
