Abstract: A configuration for a key management system is provided for managing cryptographic keys in cryptosystems. The configuration includes the use of database replication to improve the reliability, accessibility, and partition tolerance of the key management system. The configuration also includes the use of database sharding and distributed networks to improve the functionality of the key management system. From a logical level, the system can appear multi-master because client software can connect with any compute node in a plurality of compute nodes and perform key management operations on the entire key database from that compute node. From a physical level, the system can be a master-slave configuration with each database shard replication group having a single master shard instance and one or more slave shard instances. In some examples, the present disclosure provides better redundancy, load balancing, availability, and partition tolerance for key management systems.

Abstract: A configuration for a key management system is provided for managing cryptographic keys in cryptosystems. The configuration includes the use of database replication to improve the reliability, accessibility, and partition tolerance of the key management system. The configuration also includes the use of database sharding and distributed networks to improve the functionality of the key management system. From a logical level, the system can appear multi-master because client software can connect with any compute node in a plurality of compute nodes and perform key management operations on the entire key database from that compute node. From a physical level, the system can be a master-slave configuration with each database shard replication group having a single master shard instance and one or more slave shard instances. In some examples, the present disclosure provides better redundancy, load balancing, availability, and partition tolerance for key management systems.

Abstract: A method, system, and computer-program product for handling network partitions in a network is provided. For example, a method can include providing a first compute node and a second compute node on the network. In such an example, the first compute node can include a master instance of a cryptographic database. In addition, the second compute node can include a first synchronous slave instance associated with the master instance. The method can further include identifying a network partition event on the network. The method can further include providing an asynchronous slave instance associated with the master instance and changing the asynchronous slave instance to a second synchronous slave instance in response to identifying the network partition event.

Abstract: A configuration for a key management system is provided for managing cryptographic keys in cryptosystems. The configuration includes the use of database replication to improve the reliability, accessibility, and partition tolerance of the key management system. The configuration also includes the use of database sharding and distributed networks to improve the functionality of the key management system. From a logical level, the system can appear multi-master because client software can connect with any compute node in a plurality of compute nodes and perform key management operations on the entire key database from that compute node. From a physical level, the system can be a master-slave configuration with each database shard replication group having a single master shard instance and one or more slave shard instances. In some examples, the present disclosure provides better redundancy, load balancing, availability, and partition tolerance for key management systems.

Abstract: A method, system, and computer-program product for handling network partitions in a network is provided. For example, a method can include providing a first compute node and a second compute node on the network. In such an example, the first compute node can include a master instance of a cryptographic database. In addition, the second compute node can include a first synchronous slave instance associated with the master instance. The method can further include identifying a network partition event on the network. The method can further include providing an asynchronous slave instance associated with the master instance and changing the asynchronous slave instance to a second synchronous slave instance in response to identifying the network partition event.

Abstract: A configuration for a key management system is provided for managing cryptographic keys in cryptosystems. The configuration includes the use of database replication to improve the reliability, accessibility, and partition tolerance of the key management system. The configuration also includes the use of database sharding and distributed networks to improve the functionality of the key management system. From a logical level, the system can appear multi-master because client software can connect with any compute node in a plurality of compute nodes and perform key management operations on the entire key database from that compute node. From a physical level, the system can be a master-slave configuration with each database shard replication group having a single master shard instance and one or more slave shard instances. In some examples, the present disclosure provides better redundancy, load balancing, availability, and partition tolerance for key management systems.