Network, Systems and Service Management (1)

The MA approach for Network, Systems and Service Management is able to avoid any centralization point and provides better fault tolerance and scalability than centralized C/S scheme. Several administrators can be concurrently active and even cooperate to obtain a single administration goal. It is easy to generate/destroy agents and to replicate them in case of a large number of nodes in the locality. In SOMA management applications, agents act on behalf of administrators and fulfil administration needs by moving and executing on different nodes. Any administrator can implement her policy by using agents.

The SOMA management environment provides a rich set of already defined agents for systems management. In addition, it is easy to tailor new agents to new specific administration needs to delegate the automation of new management tasks. The following list gives a few examples of already implemented functionality of agents in SOMA:

agents can monitor the state of the distributed system;
agents can help in the configuration of any new or reinserted node;

agents can easily be used in the control and coordination of replicated resources;

agents can be in charge of the shutdown of the whole system and can also guarantee a minimal survival service level;
agents can regulate and improve the access to different databases by taking into account both the traffic level
and the locality of the queries;

agents can dynamically install new communication protocols for new applications.

 

 
Page updated on
In case of problems, or if you find any bug, please contact us.