International Business Machines Corporation

Matej Pavlovic

Marko Vukolić

Chrysoula Stathakopoulou

Selma Steinhoff

A Byzantine fault-tolerant distributed computing system of nodes configured to process client requests, executes a leader-based total order broadcast (LBTOB) protocol. The computing system concurrently executes multiple instances of the LBTOB protocol at the nodes, on respective disjoint partitions of the client requests, and the disjoint partitions are assigned to the instances using a deterministic function of a state of the system. The computing system multiplexes outputs of the executing instances in an ordered log of the client requests, and the ordered log is maintained at each of the nodes.

Scalable leader-based total order broadcast protocol for distributed computing systems

A computer manages reconfigurations of a Byzantine fault-tolerant, distributed computing system comprising a network of first nodes adhering to a given consensus protocol at a reconfiguration service. The computer services the network by receiving a request of change of status of a second node with respect to the network. The computer informs at least a subset of the first nodes of the received request. The computer obtains an approval of the request, whereby at least a subset of the first nodes collectively approve the change of status as a result of contributions processed according to the given consensus protocol. The computer updates a configuration log according to request approvals obtained by servicing the network. The computer addresses requests of clients about configurations of the network based on the updated configuration log.

Growing community of inventors

Adliswil, Switzerland

Matej Pavlovic