Network, Systems and Service Management (2)

As an example of a SOMA management agent, let us consider the monitor agent that can report to one administrator the information about the state of the whole distributed system. The agent gives the situation of each node in terms of system and application indicators (see Table 1). In addition, it can report network management information, such as the measured collision rate.
System indicators
Application indicators
CPU load
collision rate
availability of services
file system occupation
network connectivity
program versioning
swap space available
firewall state
application processes situation
daemon process situation
...
local agent states
printers status
...
...
Table 1. Some of the figures available in the SOMA management application.

SOMA makes possible to delegate specific controlling actions to agents, relieving the administrator duty: one agent, for instance, can automatically take care of software upgrading on all the nodes of one domain. Another distinguished feature of SOMA is the capacity of modifying system policies at run-time. In case of a policy modification that interests several nodes, there is no need to shutdown the system: a new agent can bring the new policy everywhere. The same run-time propagation applies to any static function.

Any administrator can access to the SOMA management environment via any Web browser. In fact, SOMA provides a user-friendly graphical interface to operate directly on the system. For example, Figure 1 shows how the administrator can control the initial configuration of the system and its modification at run-time. Any administrator is first authenticated, and then authorized to perform different operations depending on her role. The same interface permits administrators to handle new roles and administrators, to add new places and domains to the system, and to provide new resources and behavior. Figure 2 describes the GUI offered by the monitoring tool to report the state of a specific host. The administrator is given the situation of a node in terms of system and application factors, for instance, the state situation of physical resources, such as CPU and disk occupation. The application monitoring part permits to create and send new agents where requested in the system.


Figure 1. Web-based GUI for locality configuration
(click here for a larger view of the picture)


Figure 2. Web-based GUI for distributed monitoring
(click here for a larger view of the picture)
 
Page updated on
In case of problems, or if you find any bug, please contact us.