|
|
Cisco EMF consists of a series of applications that form a "front-end" graphical user interface (GUI) to process input and a series of "back-end" server processes which maintain a model of the network, and actually carry out the interfacing to the network elements.
The standard Cisco EMF system comes with the following set of applications:
Only a brief overview of these applications are provided in this chapter. Detailed instructions for each application are provided in later chapters.
The most common form of a Cisco EMF installation includes at least one plug-in module referred to as an Element Manager. The Element Manager modules add custom GUI windows and modeling behavior to the standard Cisco EMF system in order to allow the management of a specific type of network element.
Multiple Element Managers can be installed onto a single Cisco EMF server, thereby allowing multi-device and multi-vendor management from a single system.
The extra concepts and operating instructions for a particular Element Manager are provided with that specific module.
The Cisco EMF system keeps a model of the managed network within it's database. This model is used to keep track of the current state of the various network elements and various abstractions of this network.
Some of the concepts behind this model must be understood in order for you to make full use of the features within the system.
Each element managed by Cisco EMF is modeled as an object.
An object can represent some part of the network; for example, a router or a switch. An object can also represent an abstraction of the network, that is, a site or a region; or it can represent some of the services provided by the network; for example, a PVC. An object may also be used to represent something (or someone) that interacts with the network, such as a subscriber or a customer.
Each object within Cisco EMF has an associated object class. Each class of object simply indicates a different kind of element. Examples of classes are: routers, line cards, sites, etc.
Each class of object has different data stored against it and displays different behavior.
In the Map Viewer application, the class of the object is indicated with a different icon used within the Map View browser.
The use of classes also allows powerful queries to be carried out based upon the kind of object. Examples of this type of query could be: show all events in the system from cable modems or create a group of router objects.
These attributes are associated with the object according to the granularity of object types.
A type is simply a collection of related attributes and each class usually has a number of types.
An object's class defines which types, and therefore, which attributes, it is allowed to have and which types it has by default.
An example of the association between classes and types is shown in Figure 2-2.
In the above example, a UnixWorkstation class is specified. This class of object includes two types: System and snmpManageable. The System type includes the sysDesc, sysUpTime, and sysObjectId attributes. The snmpManageable type includes the read-community and write-community attributes.
You can access Cisco EMF objects by navigating through one of the views to find the object. Each view represents a different way of containing and grouping the objects. Most Element Managers add additional views onto the standard set supplied by Cisco EMF. The standard views provided are the Physical and Network view.
The Physical view is used to show the actual physical location of an element. An example of a physical containment relationship is shown in Figure 2-3.
An object group is simply a collection of objects which are related in some way. They may all be the same type of equipment or all belong to the same customer.
Object groups can be built either manually or by building a query. Some Cisco EMF subsystems may also build object groups which may be visible and usable by the Cisco EMF user.
Object groups are accessible through the Object Group Manager application (refer to "Object Group Manager," for more information).
This section describes Cisco EMF sessions.
When you want to use Cisco EMF, log in to a Cisco EMF session. All activities you perform happen within this session.
Cisco EMF provides User Access Control which allows a system administrator to control what different users are able to do. Each user has a different login name and password, with a specific set of privileges within the system.
A standard Administrator user (admin) is available by default. The Administrator user has access to all features at all times. The Administrator user may not be edited other than to change the password.
Cisco EMF allows you to open multiple user sessions. Each user session has access to the tools displayed in the Launchpad. Any changes in a user session are reflected immediately in all other user sessions.
Each Cisco EMF application provides a set of features which can be accessed by all other applications within Cisco EMF. Using this model, the suite of applications appears to be one single integrated application.
You can log into a Cisco EMF session and use the features of any application within the Cisco EMF suite. The features run within the context of your session.
Examples of some Cisco EMF features can be found by right clicking on a network element within the Map Viewer application. The pop-up menu shows a set of features currently available to you in this network element.
The selection of available features depends upon the class of element and the user's access permissions.
For example, administration of particular managed objects should only be performed by operators who are responsible for that particular site, or for a region in which that site belongs. However, these operators may also require visibility of objects outside their own area of control.
The basic building blocks used to control user access are described below. Refer to "User Access Control," for more detailed information.
Cisco EMF user accounts can be collected by an administrator into groups. These user groups can be used to model user roles. A typical set up would involve a user group for system administrative users, or for network fault detail users, or for operators to manage a given site.
It is on the basis of these user groups that Cisco EMF applies access control. The Cisco EMF administrator configures access control by assigning access specifications to the relevant user groups.
All features offered to a user are grouped together into feature lists. The benefit of feature lists is that it is easy to give access to a related set of features by simply choosing a feature list instead of having to assign features individually.
Any given feature may appear in more than one feature list.
Access specifications connect together the user groups, the features that can be invoked by a group, and the objects upon which these features can be invoked.
A number of access specifications are provided by default with the Cisco EMF system. More access specifications can be built at the discretion of the system administrator.
The installation of an Element Manager also usually causes extra access specifications to be added to the system in order to give access to the new features provided by this module.
Each access specification may include the following components:
A Cisco EMF event represents a notification from a managed entity that a certain condition has just occurred. These events usually represent error conditions on managed elements.
Each event is associated with the object for which it provides notification. Therefore, an object can have a number of events related to itself at any one time.
When managing devices over SNMP, Cisco EMF receives traps from the managed devices. These traps are converted into Cisco EMF events which are then raised against the managed devices which sent the traps.
An Element Manager component installed onto the standard Cisco EMF system may raise events into the system to indicate various error conditions on objects under the Element Manager's control. An example of this would be the loss of communications with an SNMP agent.
An external system may raise events into the Cisco EMF system through a custom integration channel.
The default information stored against all Cisco EMF events includes:
Descriptions of event state and severity are given below.
The event state indicates whether the event is acknowledged or unacknowledged and active or cleared.
When a new event is received by the system, its state is active/unacknowledged. You may acknowledge the event, which indicates to other users that the event is being handled. Once the event has been dealt with, you may clear the event. When you cannot clear an event due to an existing problem, it can be returned to the unacknowledged state and subsequently acknowledged or cleared by another user.
When an event is in the unacknowledged or acknowledged state, it is counted as being active, and therefore, it is still affecting the state of the object upon which it was raised.
Events continue to be stored within the system after they are cleared in order to maintain an event history for an element. These events can be viewed and manipulated in the same way as any other event.
Each event has a severity, indicating the importance of the event, and is identified with a corresponding color as follows:
The source domain identifies where an event was generated. In Cisco EMF, the source domain can be one of the following:
This is the management domain of SNMP trap information. The SNMP MIB specific information typically defines the equipment type generating a trap.
Events stored within the system can be viewed and manipulated using the Event Browser. Detailed information about the application can be found in "Event Browser."
Events can be cleared either directly, by moving the event to a cleared state, or by the system clearing the event as a result of clear correlation.
When you select to clear an event, a Clearing Event is generated.
A user generated clearing event contains the name of the user who performed the clear, and the date and time at which it was performed. It is possible for you to enter text to be associated with the clearing event.
When multiple events are being cleared, you can elect to use the same clearing information for all events.
The Event Browser displays an additional gadget for clearing an event.
The Cisco EMF system can be configured to automatically clear an event as the result of another event entering the system.
This configuration is performed on a per Element Manager basis and is provided with any Element Manager package you install.
A common example of a clear correlation is where the arrival of one SNMP trap signifies a previous SNMP trap is no longer valid. For example, a link up trap indicates a previous link down trap is no longer outstanding on a device.
In order to make the identification of potential problems easy, the Cisco EMF system propagates the alarm state of objects upwards through each object view.
In real terms, this means that if an object receives an event, then not only does it change color to reflect its new state, but all parent objects within a view, also change color, to reflect the most severe alarm on any of the children. The example in the following diagram shows a typical physical view of the network. The line cards are contained within the chassis, the chassis within a bay, the bay within a site, etc.
If a minor alarm was received on Port B, then it, and all of the objects up to the region, turn yellow to indicate a potential minor problem.
If a critical alarm was then received on Port A, then it, and all of the objects up to the region, turn red to indicate a potential critical problem.
If the critical alarm is then cleared, the icons return to yellow.
A map is a graphical representation of related objects being managed by Cisco EMF. Each object within each view has a separate map. In other words, the same object in different views could have a different map in each view.
You can view different levels of complexity in your managed network by looking at high level maps or by navigating to submaps which represent lower levels of detail in the managed network.
Each map is a two-dimensional canvas that displays nodes and links, each of which represents a Cisco EMF object. Refer to "Map Viewer," for detailed information.
Nodes and links are graphical symbols which represent managed objects in the network or abstracted objects. A node appears as an icon on a map, a link appears as a line connecting two nodes. Links can only exist as a connection between two nodes.
Each node and link clearly indicates the current operational and administrative state of the object it represents.
State indication is covered in "Map Viewer."
A background image can be applied to any map within the system. Normally, this would be a geographical map, a floor plan, or some form of schematic. Nodes and links can then be placed on top of this image as desired.
As you traverse through the views within the system using the Map Viewer, you see the map representation for the level being displayed on screen.
In many cases, what you see is all of the child objects laid out in a grid pattern within the display. This is referred to as a transient map, as it is actually generated by the map sub-system "on the fly" and is not stored within the Cisco EMF database.
If you wish to make manual changes to a transient map (for example, to place the objects at particular locations or to add a background map) you simply must open the map for editing. From then on, it is stored persistently within the Cisco EMF database.
Deployment is the term used within Cisco EMF to mean the addition of objects to the Cisco EMF network model.
You can add objects to the model in the following ways:
The Deployment Wizard is the graphical interface used to prepare the Cisco EMF Server for the deployment and subsequent management of new network elements. Cisco EMF uses a wizard style interface because it is particularly suitable for the act of deployment; a sequential series of tasks which require input at each stage. Refer to "Deployment Wizard," for detailed information.
Usually the first screen presented in the wizard is a list of the possible deployment scenarios you can perform. Depending upon which scenario you choose, you are then presented with a series of screens asking various questions regarding the objects you wish to deploy.
Auto Discovery allows you to use ICMP pings and SNMP requests to automatically find and add objects to the object model. Refer to "Auto Discovery," for detailed information.
Posted: Fri Jan 7 15:20:29 PST 2000
Copyright 2000©Cisco Systems Inc.
