System, method, and recording medium for distributed probabilistic eidetic querying, rollback, and replay
AssigneeInternational Business Machines Corporation
InventorsBong Jun Ko, Christian Makaya, Jorge J. Ortiz, Swati Rallapalli, Dinesh C. Verma, Xiping Wang
Life of this patent
Estimating the term…
Abstract
A provenance method, system, and non-transitory computer readable medium for a plurality of eidetic systems having logs, include crawling the logs of each node of a plurality of nodes of the eidetic systems to cluster segments across the logs of temporally correlated events into clustered segments and analyzing the correlated segments to interleave an order of processes in the logs and assign a probability to the order of the processes occurring.
Claims18
A provenance system for a plurality of eidetic systems having logs, the system comprising:
- a processor;
- and
- a memory, the memory storing instructions to cause the processor to perform:
- crawling the logs of each node of a plurality of nodes of the eidetic systems to cluster segments across the logs of temporally correlated events into clustered segments;
- and
- analyzing the correlated segments to interleave an order of processes in the logs and assign a probability to the order of the processes occurring while keeping the logs of each node of the plurality of nodes separate to their respective eidetic system,
- wherein each node of the plurality of nodes of the eidetic systems independently logs only local events, and
- wherein, during the analyzing of the correlated segments, the order of processes in the logs are interleaved using a clock-drift analysis between the provenance system and each of the plurality of eidetic systems,
- further comprising preventing an exchange of the logs of each of the plurality of eidetic systems to another eidetic system of the plurality of eidetic systems during the crawling by conditioning an output of the logs and, through observation, updating a probability tree for each set of events in the clustered segments by serializing these sets for traversal.
A provenance method for a plurality of eidetic systems having logs, the method comprising:
- crawling the logs of each node of a plurality of nodes of the eidetic systems to cluster segments across the logs of temporally correlated events into clustered segments;
- and
- analyzing the correlated segments to interleave an order of processes in the logs and assign a probability to the order of the processes occurring while keeping the logs of each node of the plurality of nodes separate to their respective eidetic system,
- wherein each node of the plurality of nodes of the eidetic systems independently logs only local events, and
- wherein, during the analyzing of the correlated segments, the order of processes in the logs are interleaved using a clock-drift analysis between the provenance system and each of the plurality of eidetic systems,
- further comprising preventing an exchange of the logs of each of the plurality of eidetic systems to another eidetic system of the plurality of eidetic systems during the crawling by conditioning an output of the logs and, through observation, updating a probability tree for each set of events in the clustered segments by serializing these sets for traversal.
A non-transitory computer-readable recording medium recording a provenance program for a plurality of eidetic systems having logs, the program causing a computer to perform:
- crawling the logs of each node of a plurality of nodes of the eidetic systems to cluster segments across the logs of temporally correlated events into clustered segments;
- and
- analyzing the correlated segments to interleave an order of processes in the logs and assign a probability to the order of the processes occurring while keeping the logs of each node of the plurality of nodes separate to their respective eidetic system,
- wherein each node of the plurality of nodes of the eidetic systems independently logs only local events, and
- wherein, during the analyzing of the correlated segments, the order of processes in the logs are interleaved using a clock-drift analysis between the provenance system and each of the plurality of eidetic systems,
- further comprising preventing an exchange of the logs of each of the plurality of eidetic systems to another eidetic system of the plurality of eidetic systems during the crawling by conditioning an output of the logs and, through observation, updating a probability tree for each set of events in the clustered segments by serializing these sets for traversal.
Dependencies are inferred from the claim text.