Method and system for demounting a chain of linked processes in a distributed operating system

Claim 1

A method in a distributed operating system for enabling demounting of a chain of linked processes, wherein each process is a resource for executing applications in said operating system, and for use by a job in order to enable execution of program code in process, the job being a phenomenon which is directed to said process in order to cause execution of a method in an object owned by the process, the job also being able to create new jobs directed to other processes or towards an own process, and wherein said process provides the job with resources and synchronizes jobs by only allowing one job at a time to execute, said method comprising the steps of: providing communication links between said processes by associating communication ports as a property of the processes, wherein the communication ports are operating system resources for enabling the processes to be linked together to form a link picture in which each port of a process has a reference to a port of another process, and wherein knowledge of the link picture is limited to each port's information formed by the references; using said distributed operating system for keeping up with processes of the link picture, and using said ports for transferring termination-related information if a corresponding process is terminated; detecting an error in a faulty one of the processes and terminating the faulty process; transferring information via said ports regarding a process or computer failure, and propagating the transferred information through a whole chain of linked processes in the link picture; and reporting the transferred information to running applications being executed in said linked processes for enabling the running applications to perform application specific recoveries.

Claim 2

A method in a distributed operating system for enabling demounting of a chain of linked processes, wherein each process is a resource for executing applications in said operating system, and for use by a job in order to enable execution of program code in process, the job being a phenomenon which is directed to said process in order to cause execution of a method in an object owned by the process, the job also being able to create new jobs directed to other processes or towards an own process, and wherein said process provides the job with resources and synchronizes jobs by only allowing one job at a time to execute, said method comprising the steps of: providing communication links between said processes by associating communication ports as a property of the processes, wherein the communication ports are operating system resources for enabling the processes to be linked together to form a link picture in which each port of a process has a reference to a port of another process, and wherein knowledge of the link picture is limited to each port's information formed by the references; using said distributed operating system for keeping up with processes of the link picture, and using said ports for transferring termination-related information if a corresponding process is terminated; detecting an error in a faulty one of the processes and terminating the faulty process; transferring information via said ports regarding a process or computer failure, and propagating the transferred information through a whole chain of linked processes in the link picture; and reporting the transferred information to running applications being executed in said linked processes for enabling the running applications to perform application specific recoveries.

Claim 7

A system for demounting a chain of linked processes in a distributed operating system, wherein each process is a resource for executing applications in said operating system, and for use by a job in order to enable execution of program code in the process, the job being a phenomenon which is directed to said process in order to cause execution of a method in an object owned by the process, the job also being able to create new jobs directed to other processes or towards an own process, and wherein said process provides the job with resources and synchronizes jobs by only allowing one job at a time to execute, the system comprising: first means for providing communication links between processes by associating communication ports as a property of the processes, wherein the communication ports are operating system resources for enabling the processes to be linked together to form a link picture in which each port of a process has a reference to a port of another process, and wherein knowledge of the link picture is limited to each port's information formed by the references; second means for enabling said distributed operating system to keep up with processes of the link picture, and to use said ports for transferring termination-related information if a corresponding process is terminated; third means for enabling said distributed operating system, and in certain cases a process itself, to detect an error in a faulty one of the processes and terminate the faulty process; fourth means for enabling said distributed operating system to transfer failure information via said ports regarding a process or computer failure, and to propagate the transferred failure information through a whole chain of linked processes in the link picture; and fifth means for enabling said distributed operating system to report said failure information to running applications being executed in the linked processes in order to enable the running applications to perform application specific recoveries.

Claim 8

A system for demounting a chain of linked processes in a distributed operating system, wherein each process is a resource for executing applications in said operating system, and for use by a job in order to enable execution of program code in the process, the job being a phenomenon which is directed to said process in order to cause execution of a method in an object owned by the process, the job also being able to create new jobs directed to other processes or towards an own process, and wherein said process provides the job with resources and synchronizes jobs by only allowing one job at a time to execute, the system comprising: first means for providing communication links between processes by associating communication ports as a property of the processes, wherein the communication ports are operating system resources for enabling the processes to be linked together to form a link picture in which each port of a process has a reference to a port of another process, and wherein knowledge of the link picture is limited to each port's information formed by the references; second means for enabling said distributed operating system to keep up with processes of the link picture, and to use said ports for transferring termination-related information if a corresponding process is terminated; third means for enabling said distributed operating system, and in certain cases a process itself, to detect an error in a faulty one of the processes and terminate the faulty process; fourth means for enabling said distributed operating system to transfer failure information via said ports regarding a process or computer failure, and to propagate the transferred failure information through a whole chain of linked processes in the link picture; and fifth means for enabling said distributed operating system to report said failure information to running applications being executed in the linked processes in order to enable the running applications to perform application specific recoveries.

Claim 16

The system according to claims 6, further comprising: a system updating function; system updating information attributes for carrying information internal to said updating function regarding type of traffic carrying on in the system; and means for transferring said information attributes in a chain of jobs for enabling determination of the version of a specific program to be executed at an execution occasion.