Electrical computers and digital processing systems: multicomput – Distributed data processing – Processing agent
Reexamination Certificate
1999-04-13
2002-10-15
Barot, Bharat (Department: 2154)
Electrical computers and digital processing systems: multicomput
Distributed data processing
Processing agent
C709S207000, C709S223000, C709S228000, C709S241000
Reexamination Certificate
active
06466963
ABSTRACT:
FIELD OF THE INVENTION
This invention concerns an agent system in which an agent generated by an agent server is moved consecutively among communication devices via communication lines. In this agent system, a mobile agent with an individual priority is transmitted from the system server to another system server at a destination.
BACKGROUND OF THE INVENTION
A typical agent system of the type which has recently been proliferating is shown in FIG.
8
. Agent server
50
, various communication devices
60
(personal computers, work stations, etc.) and storage device
80
, which stores a program to generate agents, are all connected via cable line
90
(or, although not pictured, they may be connected cordlessly). This line
90
is used to transfer agent
70
, which we shall discuss shortly, from one device to the next.
In such an agent system, agent server
50
generates, based on a source program, an agent
70
consisting of object code. The agent
70
, which has been generated, is transmitted to a designated communication device
60
, where the system agent
601
for that device
60
executes it.
Agent systems of the sort described above use two methods to transmit an agent
70
. In a first method, the agent is given its destination and target processing but not the route by which it is to reach the destination. In a second method, the agent is given only its target processing and does not receive either its destination or its transmission route.
Regardless of whether the agent is given both its destination and its target processing or only its target processing, certain problems are inherent in the type of agent system described above.
In the first method mentioned above, in which agent
70
is given both its destination and its target processing, when the agent is to be routed through the nodes of a number of specified devices
60
, each node will have to be specified, as there is a 1 to 1 correspondence of destinations when the target processing is executed.
Thus if the target processing cannot be accomplished, agent server
50
will have to respecify all of the destinations, which will result in dead time.
In the second method mentioned above, in which the agent is given only its target processing, it will move to an arbitrary location (searching according to a rule) and execute that processing. Once it is given its target processing, the agent will never need to have its target reassigned. However, there is no way to choose a route, so if the processing is sequential, and there is, for example, an application gateway in a fire wall, the attempt of the agent to pass through will be regarded as an access violation. The agent will be unable to continue to the nodes of the devices
60
on the other side of the firewall.
In either method mentioned above (the one in which the agent is given both a destination and the target processing or the one in which it is given only the target processing), agent
70
sets out without knowing the resource state of its destination, device
60
. For example, it does not know which regions of the device's memory are free. If it turns out that the agent is temporarily unable to use the resources of device
60
because, say, there is no free area in the memory to house it, agent
70
will not be generated in that device. It will not return to server
50
within the specified time, and server
50
will have no alternative but to execute error processing.
In order to solve the problems mentioned above, the applicant filed a patent application disclosing the following invention on Mar. 26, 1999.
In this application, the aforesaid communication devices
60
modify the routing data, if necessary, for the agent which was received previously, and restore the agent. This configuration eliminates the need to respecify destinations for the agent. It enables the agent to reach the target destination without routing a specific node. It also checks the resource state of its destination in advance before the agent is transmitted in order to avoid error processing.
The following problem is, however, inherent in the invention mentioned above. The processing time to execute the agent
70
at the designated communication device
60
with node number
3
is not considered in this configuration. As shown in
FIG. 9
, after the agent
70
is transmitted from node number
1
, if an error occurs in the device
60
with node number
3
while the agent
70
is on the way, the agent has no way to reach to the device
60
with node number
3
, and will disappear before it reaches there. This means there is no way to notify to the device
60
with node number
1
that an error occurred at node
3
. Because of this configuration, there is an inherent problem with the time required for detecting an error in the system.
Another problem in this system is that it can not prioritize the processing of agents when there are a number of agents
70
at a device such as the device
60
with node number
3
as shown in FIG.
10
. This is because the agents are processed according to the order they are input.
SUMMARY OF THE INVENTION
An object of this invention is to provide an agent system that can easily detect an error in the system and that can prioritize the agents to be processed.
In order achieve this object, each agent according to this invention has an ID and priority number to prioritize the processing of the agent in the designated communication. Once this agent is transmitted to the designated device, the agent is processed according to the priority number when the designated device has more than two agents.
The agent server has a priority inquiring means to inquire as to the priorities of other agents that are already in the designated agent before the agent server transmits the agent to the designated device.
With this invention, a priority altering means alters the priority number of an agent once the agent server receives information about the priorities of other agents in the designated device by the priority inquiring means.
This invention also has a time out detecting means to detect if a mobile agent is executed within a predetermined time period, and an error detecting means to detect an error at a designated communication device by counting the number of times the agent is resent, and to judge if there is an error in the processing.
REFERENCES:
patent: 5999183 (1999-12-01), Kilgariff et al.
patent: 5999965 (1999-12-01), Kelly
patent: 6065039 (2000-05-01), Paciorek
patent: 6085238 (2000-07-01), Yuasa et al.
patent: 6157941 (2000-12-01), Verkler et al.
patent: 6175855 (2001-01-01), Reich et al.
patent: 6175868 (2001-01-01), Lavian et al.
patent: 6195705 (2001-02-01), Leung
patent: 6233601 (2001-05-01), Walsh
patent: 6282563 (2001-08-01), Yamamoto et al.
patent: 0 634 719 (1995-01-01), None
Wong et al.; “Concordia: An Infrastructure for Collaborating Mobile Agents”; Mitsubishi Electric ITA, Horizon Systems Laboratory, Waltham, MA, 1997, pp. 86-97.
Rothermel et al; “The Architecture of the Ara Platform for Mobile Agents”; First International Workshop on Mobile Agents, 1997, 12 total pages.
Barot Bharat
Dickstein , Shapiro, Morin & Oshinsky, LLP
Omron Corporation
LandOfFree
Agent system with prioritized processing of mobile agents does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Agent system with prioritized processing of mobile agents, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Agent system with prioritized processing of mobile agents will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2940705