Distributed information synchronization
View Patent ↗Processing a prepared update is disclosed. A prepared update associated with a request that has been used by the sender to update a local version of a data associated with the sender is received from a sender. Based at least in part on an identifier included in the prepared update, a selected data handler is selected among a plurality of data handlers. The selected data handler is used to update a centralized version of the data at least in part by using the received prepared update. The centralized version of the data has been previously updated using a plurality of prepared updates received from a plurality of senders. The updated centralized version of the data is sent to update the local version of the data associated with the sender.
1. A system for processing a prepared update, comprising:
a communication interface configured to receive from a sender, a prepared update associated with a request that has been used by the sender to update a local version of a data associated with the sender; and
a processor configured to:
based at least in part on an identifier included in the prepared update, select a selected data handler among a plurality of data handlers;
use the selected data handler to update a centralized version of the data at least in part by using the received prepared update, wherein the centralized version of the data has been previously updated using a plurality of prepared updates received from a plurality of senders; and
send the updated centralized version of the data to update the local version of the data associated with the sender.
2. The system of claim 1 , wherein the prepared update has been addressed to the system using a machine identifier obtained from a directory server that maps the identifier included in the prepared update to the machine identifier.
3. The system of claim 1 , wherein the received prepared update had been merged with another prepared update in a sending data structure before being received by the communication interface.
4. The system of claim 1 , wherein the identifier included in the prepared update includes a hash value.
5. The system of claim 1 , wherein the data includes a mean value and a variance value.
6. The system of claim 1 , wherein the plurality of data handlers is hosted on a single machine.
7. The system of claim 1 , wherein the processor is further configured to store the received prepared update in a data structure, and using the selected data handler to update the centralized version of the data includes obtaining the received prepared update from the data structure.
8. The system of claim 7 , wherein one or more prepared updates in the data structure are processed at a determined periodic interval.
9. The system of claim 1 , wherein before the received prepared update is used to update the centralized version of the data, the received prepared update is merged with another received prepared update that includes a same identifier as the identifier included in the received prepared update.
10. The system of claim 1 , wherein the prepared update includes a differential value that represents an incremental update value to the centralized version of the data.
11. The system of claim 1 , wherein using the selected data handler to update the centralized version of the data includes obtaining the centralized version of the data using the identifier.
12. The system of claim 1 , wherein using the selected data handler to update the centralized version of the data includes computing a new mean value and a new variance value using the centralized version of the data and the received prepared update.
13. The system of claim 12 , wherein computing the new mean value and the new variance value includes using a parameter provided by a model version server that provided the parameter in response to being provided a model version identifier included in the prepared update.
14. The system of claim 1 , wherein using the selected data handler to update the centralized version of the data includes adjusting a variance value included in the data by using a dynamics adjustment factor based on a time value.
15. The system of claim 1 , wherein the processor is further configured to store the updated centralized version of the data to a backup storage.
16. The system of claim 1 , wherein sending the updated centralized version of the data to update the local version of the data associated with the sender includes storing the updated centralized version of the data in a storage accessible by the sender.
17. The system of claim 1 , wherein the updated centralized version of the data is associated with a version identifier based on a time when the centralized version of the data was updated.
18. The system of claim 1 , wherein the data is associated with a Bayesian inference algorithm.
19. A method for processing a prepared update, comprising:
receiving from a sender, a prepared update associated with a request that has been used by the sender to update a local version of a data associated with the sender;
based at least in part on an identifier included in the prepared update, using a processor to select a selected data handler among a plurality of data handlers;
using the selected data handler to update a centralized version of the data at least in part by using the received prepared update, wherein the centralized version of the data has been previously updated using a plurality of prepared updates received from a plurality of senders; and
sending the updated centralized version of the data to update the local version of the data associated with the sender.
20. The method of claim 19 , further comprising storing the received prepared update in a data structure, wherein using the selected data handler to update the centralized version of the data includes obtaining the received prepared update from the data structure.
21. The method of claim 19 , wherein using the selected data handler to update the centralized version of the data includes computing a new mean value and a new variance value using the centralized version of the data and the received prepared update.
22. The method of claim 19 , wherein using the selected data handler to update the centralized version of the data includes adjusting a variance value included in the data by using a dynamics adjustment factor based on a time value.
23. A computer program product for processing a prepared update, the computer program product being embodied in a non-transitory computer readable storage medium and comprising computer instructions for:
receiving from a sender, a prepared update associated with a request that has been used by the sender to update a local version of a data associated with the sender;
based at least in part on an identifier included in the prepared update, selecting a selected data handler among a plurality of data handlers;
using the selected data handler to update a centralized version of the data at least in part by using the received prepared update, wherein the centralized version of the data has been previously updated using a plurality of prepared updates received from a plurality of senders; and
sending the updated centralized version of the data to update the local version of the data associated with the sender.
24. The computer program product of claim 23 , further comprising computer instructions for storing the received prepared update in a data structure, wherein using the selected data handler to update the centralized version of the data includes obtaining the received prepared update from the data structure.
25. The computer program product of claim 23 , wherein using the selected data handler to update the centralized version of the data includes computing a new mean value and a new variance value using the centralized version of the data and the received prepared update.
26. The computer program product of claim 23 , wherein using the selected data handler to update the centralized version of the data includes adjusting a variance value included in the data by using a dynamics adjustment factor based on a time value.