Apparatus and method for efficient, coordinated, distributed execution
The present disclosure describes a system and method for efficient, coordinated, distributed execution. Processes are registered into a process registry which includes a table or list of processes. Processes which are determined to be no longer running are removed from the registry. Each process is configured to send a broadcast when it starts or stops. Each process periodically reads the registry, sorts the list, and uses it to determine the current master process based on an algorithm.
1. A method, comprising:
determining that a particular process of a plurality of processes registered with a process registry is not running based on a message received from at least one process of the plurality of processes;
in response to determining that the particular process is not running, removing the particular process from the process registry;
in response to removing the particular process from the process registry, selecting a master process from remaining processes registered with the process registry based on a length of time the selected master process would be the master process and a number of processes registered with the process registry; and
periodically selecting a new master process at predetermined time intervals.
2. The method of claim 1 , wherein selecting a master process from remaining processes registered with the process registry comprises:
selecting the master process from the remaining processes registered with the process registry based on: index=[((time+offset)/rotate period) % no. of processes].
3. The method of claim 1 , further comprising:
selecting a secondary master process from the process registry based on the master process.
4. The method of claim 1 , wherein the message received from the at least one process indicates that the at least one process is running or that the at least one process is not running.
5. The method of claim 1 , wherein determining that a particular process of a plurality of processes registered with a process registry is not running based on a message received from at least one process of the plurality of processes comprises:
determining that the message received from the at least one process does not include a message received from the particular process.
6. The method of claim 1 , wherein selecting the master process from the remaining processes registered with the process registry is performed by the plurality of processes registered with the process registry.
7. The method of claim 1 , further comprising:
registering the plurality of processes with the process registry by associating each of the plurality of processes with an index in the process registry.
8. The method of claim 1 , wherein each of the plurality of processes registered with the process registry maintains a local copy of the process registry.
9. A computer readable medium storing computer program instructions, which, when executed on a processor, cause the processor to perform operations comprising:
determining that a particular process of a plurality of processes registered with a process registry is not running based on a message received from at least one process of the plurality of processes;
in response to determining that the particular process is not running, removing the particular process from the process registry;
in response to removing the particular process from the process registry, selecting a master process from remaining processes registered with the process registry based on a length of time the selected master process would be the master process and a number of processes registered with the process registry; and
periodically selecting a new master process at predetermined time intervals.
10. The computer readable medium of claim 9 , wherein selecting a master process from remaining processes registered with the process registry comprises:
selecting the master process from the remaining processes registered with the process registry based on: index=[((time+offset)/rotate period) % no. of processes].
11. The computer readable medium of claim 9 , the operations further comprising:
selecting a secondary master process from the process registry based on the master process.
12. The computer readable medium of claim 9 , the operations further comprising:
registering the plurality of processes with the process registry by associating each of the plurality of processes with an index in the process registry.
13. An apparatus comprising:
a processor; and
a memory to store computer program instructions, the computer program instructions when executed on the processor cause the processor to perform operations comprising:
determining that a particular process of a plurality of processes registered with a process registry is not running based on a message received from at least one process of the plurality of processes;
in response to determining that the particular process is not running, removing the particular process from the process registry;
in response to removing the particular process from the process registry, selecting a master process from remaining processes registered with the process registry based on a length of time the selected master process would be the master process and a number of processes registered with the process registry; and
periodically selecting a new master process at predetermined time intervals.
14. The apparatus of claim 13 , wherein selecting a master process from remaining processes registered with the process registry comprises:
selecting the master process from the remaining processes registered with the process registry based on: index=[((time+offset)/rotate period) % no. of processes].
15. The apparatus of claim 13 , wherein the message received from the at least one process indicates that the at least one process is running or that the at least one process is not running.
16. The apparatus of claim 13 , wherein determining that a particular process of a plurality of processes registered with a process registry is not running based on a message received from at least one process of the plurality of processes comprises:
determining that the message received from the at least one process does not include a message received from the particular process.
17. The apparatus of claim 13 , wherein selecting the master process from the remaining processes registered with the process registry is performed by the plurality of processes registered with the process registry.
18. The apparatus of claim 13 , wherein each of the plurality of processes registered with the process registry maintains a local copy of the process registry.