Abstract:
A computer implemented method and system for uniform, consistent, stateless, and deterministic consistent hashing for fixed size partitions. The method comprises receiving server information for a plurality of servers in a server pool; ordering the server information for the plurality of servers; calculating a mean number of partitions per server in the server pool; calculating a first maximum allowed number of partitions for a server by applying a floor function to the calculated mean number of partitions per server in the server pool; in the order of the ordered server information, assigning partitions to servers based on a hashing function, using the first maximum allowed number of partitions per server; calculating a second maximum allowed number of partitions for a server by applying a ceiling function to the calculated mean number of partitions per server in the server pool; in the order of the ordered server information, assigning partitions to servers based on a hashing function, using the second maximum allowed number of partitions per server; and assigning unassigned partitions wherein assigning comprises: for each unassigned partition, mapping the partition to a server with the least number of assigned partitions at a time of the mapping.
Abstract:
Network analysis including dynamic smoothing, isolating non-optimally performing systems and downstream aware transactional workflows is described herein. Each of the implementations, separately or together, is able to improve network functionality.
Abstract:
Network analysis including dynamic smoothing, isolating non-optimally performing systems and downstream aware transactional workflows is described herein. Each of the implementations, separately or together, is able to improve network functionality.
Abstract:
A computer implemented method and system for uniform, consistent, stateless, and deterministic consistent hashing for fixed size partitions. The method comprises receiving server information for a plurality of servers in a server pool; ordering the server information for the plurality of servers; calculating a mean number of partitions per server in the server pool; calculating a first maximum allowed number of partitions for a server by applying a floor function to the calculated mean number of partitions per server in the server pool; in the order of the ordered server information, assigning partitions to servers based on a hashing function, using the first maximum allowed number of partitions per server; calculating a second maximum allowed number of partitions for a server by applying a ceiling function to the calculated mean number of partitions per server in the server pool; in the order of the ordered server information, assigning partitions to servers based on a hashing function, using the second maximum allowed number of partitions per server; and assigning unassigned partitions wherein assigning comprises: for each unassigned partition, mapping the partition to a server with the least number of assigned partitions at a time of the mapping.
Abstract:
Message processing, forwarding and filtering are able to be implemented based on attribute-values extracted from a message. By utilizing attributes within the body of the message instead of header information, control of the messages is more dynamic.