Back to results
US 2012/0072689 A1Published application
Method of data replication in a distributed data storage system and corresponding device
AssigneeThomson Licensing
InventorsAlexandre Van Kempen, Erwan Le Merrer, Nicolas Le Scouarnec
Life of this patent
Estimating the term…
Abstract
The present invention generally relates to distributed data storage systems. In particular, the present invention relates to optimization of the distributed storage system in terms of number of data replicas needed to ensure a desired level of data replication.
Claims6
A method of data replication in a distributed data storage system comprising network devices that are at least used as storage devices, wherein said method comprises the following steps:
- for each storage of a data item, selection of a replication set, said replication set representing a predetermined number k of storage devices in said distributed data storage system that are selected for storing of said data item, said predetermined number k representing a so-called replication level;
- for each storage device in said replication set, determination of a timeout value for a timer associated to said storage device that is started upon disconnection of said storage device from said distributed data storage system and that is stopped upon reconnection of said storage device to said distributed data storage system, said storage device being detected as being disconnected upon expiration of said timer, said timeout value being determined as a function of a distribution of downtime data measured for said storage device, said downtime data representing time durations when said storage device was unavailable for said distributed data storage system;
- when said at least one storage device is being detected as being disconnected, and if a number of connected storage devices in said replication set is inferior to a predetermined threshold, restoring of said replication level by selection of at least one second storage device new to said replication set, and by creating of a new data item from said data item on said at least one second storage device, and adding of said at least one second storage device to said replication set.
A server device for data replication in a distributed data storage system comprising network devices that are at least used as storage devices, wherein said server device comprises the following means:
- for each storage of a data item, means for selection of a replication set, said replication set representing a predetermined number k of storage devices in said distributed data storage system that are selected for storing of said data item, said predetermined number k representing a so-called replication level;
- for each storage device in said replication set, means for determination of a timeout value for a timer associated to said storage device that is started upon disconnection of said storage device from said distributed data storage system and that is stopped upon reconnection of said storage device to said distributed data storage system, said storage device being detected as being disconnected upon expiration of said timer, said timeout value being determined as a function of a distribution of downtime data measured for said storage device, said downtime data representing time durations when said storage device was unavailable for said distributed data storage system;
- when said at least one storage device is being detected as being disconnected, and if a number of connected storage devices in said replication set is inferior to a predetermined threshold, means for restoring of said replication level by selection of at least one second storage device new to said replication set, and by creating of a new data item from said data item on said at least one second storage device, and adding of said at least one second storage device to said replication set.
A computer readable non-transitory storage medium having instructions, which, when executed by a computer, performs the steps of:
- for each storage of a data item, selection of a replication set, said replication set representing a predetermined number k of storage devices in said distributed data storage system that are selected for storing of said data item, said predetermined number k representing a so-called replication level;
- for each storage device in said replication set, determination of a timeout value for a timer associated to said storage device that is started upon disconnection of said storage device from said distributed data storage system and that is stopped upon reconnection of said storage device to said distributed data storage system, said storage device being detected as being disconnected upon expiration of said timer, said timeout value being determined as a function of a distribution of downtime data measured for said storage device, said downtime data representing time durations when said storage device was unavailable for said distributed data storage system;
- when said at least one storage device is being detected as being disconnected, and if a number of connected storage devices in said replication set is inferior to a predetermined threshold, restoring of said replication level by selection of at least one second storage device new to said replication set, and by creating of a new data item from said data item on said at least one second storage device, and adding of said at least one second storage device to said replication set.
Dependencies are inferred from the claim text.