Electrical computers and digital processing systems: multicomput – Computer network managing
Reexamination Certificate
1999-04-15
2002-08-06
Burgess, Glenton B. (Department: 2153)
Electrical computers and digital processing systems: multicomput
Computer network managing
C709S224000
Reexamination Certificate
active
06430613
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to a process and a system for network and system management.
DESCRIPTION OF RELATED ART
Large companies have increasing numbers of equipment units to manage. These units, linked together by a communication network called a “Local Area Network” (RLE, LAN), are managed by a manager. In order to manage (monitor, operate, supervise, control) equipment remotely from one point, the architectural model comprising a manager, for example (ISM,
FIG. 4
) and an agent, for example of the SNMP type, is the one most commonly adopted. In this architectural model, the agents (SNMP), implemented in the equipment (ET) of the network, inform the manager of the status of each equipment unit managed. Whenever there is an equipment malfunction, an agent (SNMP) sends the manager ISM an alarm via the wide area network (WAN). In the great majority of cases, this manager manages several hundred equipment units distributed over one or more countries. The information exchanged between the manager and the equipment managed flows through a “Wide Area Network” (WAN). However, the network WAN has limited capacities and the transmission of information through this network is currently slow and uncertain. This problem is explained by the fact that the bandwidth of the network (WAN) is too small relative to the continually increasing amount of information passing through it between managers and their equipment. Local area networks often support traffic exceeding 10 megabits, while the network WAN has a bandwidth that is often less than or equal to 64 kilobits; 9600 baud is a common value. Consequently, the network (WAN) is overloaded and a lot of information is lost. Moreover, the transmission of the data is very slow and the type of transmission (in periodic small packets) is not adapted to the current modes of utilization of WANs. The processing of information by the manager is slowed down and the corrective actions to be initiated are delayed. Furthermore, in some cases, the chronology of the inflow of information to the manager is disrupted because of this excessive flow. In this case, the processing of this information can result in an incorrect interpretation of events, which can trigger inappropriate actions on the part of the manager. The cost of communications is also increased.
One solution to the problem of loss of information consists in that the manager ISM generates through the network (WAN), at a given time, a request with the format “Are you okay?” to each system being managed and the latter responds with “Yes, I'm okay?” This solution is very costly. It does not solve the problem of the overloading of the cables and further increases the flow of information through the network (WAN). Moreover, the request or the response to this request can be lost in the network (WAN).
Another solution consists of managing the flow of information using a tool such as the SM Monitor 6000 sold under the IBM trademark. This tool, strongly linked to the platform known as “System View,” is remote and it makes it possible to concentrate the alarms of a network and to perform operations from information that can he supplied by the agents of the network. But SM Monitor 6000 consumes central processing unit (CPU) resources, and takes up a lot of memory space. Moreover, many companies today need to manage a large number of small networks. SM Monitor 6000 does not have a mechanism for large-scale deployment and therefore cannot be used to manage a large number of equipment units. Furthermore, SM Monitor 6000 uses a technology that is not very portable. The SM Monitor 6000 configuration can only be used with the “System View” platform, and is unusable without this tool.
One last solution, proposed by the company “BMC Software” consists of a control module that makes it possible to supervise a set of agents known as “Patrol agents.” A Patrol agent can contain several modules, each having the function of gathering a certain type of information, such as the system information or the information of an application (for example, an Oracle database). This solution is not appropriate, because, even though the Patrol technology makes it possible to gather certain information in the equipment units, it is not designed to play a managerial role vis-a-vis the agents. The Patrol agent processes local data in one machine; it is only a source of information and it is not supplied with data originating from other agents. Moreover, the Patrol agents approach does not have the capacity to handle changes in the equipment, even though this is fundamental to operation. Furthermore, it uses up the CPU time and resources of the target systems because of the interpretive language technology.
SUMMARY OF THE INVENTION
The object of the present invention is to eliminate the drawbacks of the prior art by offering a portable network management process adapted to the management of a large number of equipment units. The process according to the invention limits and protects the administrative flow between the manager and the equipment managed by preventing the sending of unnecessary messages or the repeated sending of the same message through the network WAN. Thus, this process adapts to the saturation levels of the bandwidth and makes it possible to reduce the loss of information in the network (WAN). Moreover, as a result of this process, the information gathering frequencies can be adapted to each individual piece of information, and the data gathered by the system can be reused at any time to produce statistics, to supervise the performance of the equipment or to avoid having to search for the same information several times. Furthermore, this system learns from its environment. Thousands of equipment units are handled automatically in spite of very different contexts. The appearance and disappearance of elementary systems are handled dynamically during operation, without the intervention of the operator. Finally, the invention makes it possible to reduce the information processing load at the control level.
This object is achieved by the fact that the process for network management comprises at least one submanager (COACH) located in the containing tree between a main manager (AD) and the equipment of the network; the submanager is located in the local area network (RLE), manages a subnetwork, and comprises various modules which communicate with one another and with a main manager (AD) through a kernel (N), the modules polling the equipment of the subnetwork and receiving the alarms triggered by the agents (snmp) operating in the equipment of the subnetwork, the process being composed of several steps:
a step during which a discovery module (MD) polls all the possible equipment units (ET) of the subnetwork,
a step for a domain search by the discovery module (MD) when an equipment unit responds to the polling (SNMP),
a step during which the discovery module (MD) sends a notification to a model configuration module (MCM), telling it the Internet address (IP) of the equipment unit discovered and the domain to which the unit discovered belongs,
a step during which the model configuration module (MCM) notifies an indicator calculation module (MCI) of the indicator to be instantiated in the equipment unit and notifies an alarm filtering module (MFA) of the filter model to be instantiated in the equipment.
According to one characteristic of the invention, the process for network management is characterized in that with each new step for the discovery of the equipment of the subnetwork, the discovery module (MD) updates the databases of the kernel (N) and of the model configuration module (MCM) containing the list of the equipment units and their domains.
According to another characteristic, all of the alarms sent by the various modules are sent to the main manager (AD) via the alarm security module (MSA), said alarm being accompanied by a send message sent to the server of the alarm security module (sMSA).
According to another characteristic, the process for network management is comp
Brunet Jean
Lamberet Florence
Bull S.A.
Burgess Glenton B.
Flynn Kimberly
Kondracki Edward J.
Miles & Stockbridge P.C.
LandOfFree
Process and system for network and system management does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Process and system for network and system management, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Process and system for network and system management will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2895298