International Business Machines Corporation

Shlomit I Shachor

Vita Bortnikov

Ilya Shnayderman

Gregory Chockler

Steven Duane Clay

Micha Gideon Moffie

Alexey Roytman

Omer Y Boehm

Zvi Cahana

Natalia Razinkov

Dmitri Perelman

David H Carman

A method comprising: obtaining multiple glossary terms each comprising one or more words; generating multiple fuzzy tokens from each word of each of the glossary terms; calculating a similarity score for each of the fuzzy tokens, the similarity score denoting a similarity between the respective fuzzy token and its respective word; obtaining multiple input terms to be matched with the multiple glossary terms; separating each of the input terms into multiple input tokens; generating multiple n-grams from each of the input tokens; comparing the n-grams with the fuzzy tokens, to output a list of matching n-grams and fuzzy tokens; based on the list of matching n-grams and fuzzy tokens, identifying, from the glossary terms, candidate glossary term matches for each of the input terms; and calculating one or more scores that quantify the match between each of the candidate glossary term matches and its respective input term.

Fuzzy matching of obscure texts with meaningful terms included in a glossary

A computer determines whether it has received user input or a node within a replica set has reached a capacity threshold. Based on receiving user input or determining that a node within a replica set has reached a capacity threshold, creating a snapshot of the data stored in the replica set and partitioning the data based on the created snapshot. The computer then initializing nodes within a new replica set and moves a partition from the original replica set to the new replica set before deleting the other partition from the old replica set.

Scaling quorum based replication systems

An approach is provided in which a replicated memory system replicates a set of data between multiple replicas, whereby each of the replicas includes a sequence update number corresponding to an update state of its respective set of data. At least one of the replicas identifies a number of replicas to which it connects and, in turn, the replicas select one of replicas as a leader replica based at least on the selected leader replica's sequence update number and the number of replicas to which the selected leader connects.

Choosing a leader in a replicated memory system

A processor-implemented method, for continuing operation of a quorum based system is provided. The method detects a loss of quorum. A plurality of speculative configurations is created, whereby each speculative configuration is isolated from other speculative configurations in the quorum based system. Each speculative configuration continues to order requests during the creation of speculative configurations. The method selects and starts one of the plurality of speculative configurations as a new operational configuration. Ordered requests continue to the new operational configuration. The original configuration of the quorum based system is restarted in response to the plurality of speculative configurations not being isolated.

Continuing operation of a quorum based system after failures

Using a quorum of fully updated replicas to designate a leader replica. A quorum of fully updated replicas designate a leader replica through 'prepared' messages, allowing for restoration of failed replicas without a split-brain condition.

Repopulating failed replicas through modified consensus recovery

A method for selecting a leader node among a plurality of network nodes, comprising: providing a current configuration of selected nodes in replicated state machine based system, wherein a first node is set for handling commands received from clients; executing a consensus protocol by the selected nodes under the current configuration; identifying at least one fault indicative event of the first node; calculating a suggested configuration of selected nodes, wherein a second node is set for handling the commands; informing each member of the suggested configuration and the first node of the suggested configuration; executing the consensus protocol in parallel under both the suggested configuration and the current configuration; and when detecting that the first node is faulty, setting the second node for handling the commands under the current configuration in place of the first node and reconfiguring the current configuration to become the suggested configuration.

Efficient fail-over in replicated systems

A method for continuously updating a set of replicas. The method comprises storing a plurality of replicas of data in a current configuration of members from a plurality of nodes, receiving a reconfiguration command by a member of said current configuration, selecting at least one estimated configuration from said plurality of nodes, receiving at least one write command by a member of said current configuration, disseminating said at least one write command to each member of said at least one estimated configuration and validating said at least one estimated configuration. At least one estimated configuration processes at least one of the write commands before the validating is completed.

Growing community of inventors

Yokneam Eilit, Israel

Shlomit I Shachor