Microsoft Technology Licensing, LLC

Krishna Kannan

Vignesh Somasundaram

Robert Aron Drollinger

Joern Lindhard Mortensen

Savio Mendes De Figueiredo

Chukwunwike I J Onianwa

Marvin Prieto Adriano

Michal Kaczmarek

The disclosure herein describes a message ordering system for processing and synchronizing chat messages in partitions to maintain messaging order and load balancing in a distributed system at scale. Messages are placed in partitions based on session identifiers (IDs). Messaging order is secured by a session receiver holding a lock over particular partitions. Receivers having subscription to messages in service bus are enabled to terminate on occurrence of a predetermined event, such as threshold wait time without receiving additional messages with a selected session ID from their partitions, activity level on a node, number of receivers on a node or other metrics. Session ID locks are released after a session ends, a receiver terminates or a node crashes for failure handling. New receivers are created to handle new incoming messages for additional load balancing and/or failover.

Processing and ordering messages in partitions

The disclosed technology is generally directed to online chat management services. In one example of the technology, organization information is accessed for tenants. The organization information includes, for each organization, configuration information for the organization. Each organization belongs to a corresponding tenant. Worker service instances are scheduled to perform fills associated with the organizations, such that sufficient fills are performed for each organization according to the configuration information for the organizations in the accessed organization information. Each fill includes: causing artifacts associated with a function to be created, such that the artifacts associated with the function are configured based on the corresponding tenant. Artifact information associated with the created artifacts is received, including at least one unique identifier associated with the created artifacts for the function. Server-side caching of a cache record that includes the artifact information associated with the created artifacts for the function is performed.

Distributed multi-tenant, multi-service server-side artifact caching

A system generates a first data set by executing, in response to occurrence of an event, a data flow configured with a property of a first object of a first data processing system to map the first object to a second object of a second data processing system. The system acquires additional information, including additional properties of the first object and data used by the first and second data processing systems to communicate with other data processing systems, from the first and second data processing systems to map the first object to the second object. The system generates a second data set having a format compatible with the second data processing system based on the first data set and the additional information and sends the second data set to the second data processing system. The system maps other objects using reconfigured data flows.

System integration using configurable dataflow

Growing community of inventors

Seattle, WA, United States of America

Krishna Kannan