Publisher-based message data caching in a publish-subscription environment

Abstract

Embodiments of the present invention provide an approach for a broker-assisted, publisher side cache that can be utilized to reduce a volume of data (e.g., network traffic) delivered between a publisher and broker in a publication/subscription (pub/sub) environment. Specifically, in a typical embodiment, when a message is received on a broker system from a publisher system, the broker system analyzes the message for potentially repetitive data. Such data can be determined from historical messages (e.g., determining that a certain set of data appeared in multiple messages and is thus flagged in a database or the like), a set of rules (e.g., if data set 'Y' appears, flag it as potentially repetitive), etc. Regardless, once such a set of data has been identified, a unique identifier/key corresponding thereto will be sent back to the publisher system and the set of data will be stored in a (data) cache. Upon receipt, the publisher system will cache/store the unique identifier in a (key) cache. Then, when the publisher system generates a future message that would contain the same set of data, the publisher system will substitute the unique identifier (as retrieved from the key cache) for the set of data to yield an abbreviated/truncated message. The abbreviated message would be sent to the broker system, which would: analyze the message; locate the unique identifier; retrieve the corresponding set of data from the (data) cache; replace the unique identifier with the set of data to yield a completed message; and then send the completed message to relevant subscriber system(s).