I suppose the main reason that i am looking for alternative ways to sending updated data is to reduce the overhead as much as possible for each of the updates. As i mentioned, only 1/12th of the data that i am writing to the socket is actually relevant information to the update... This is a large overhead! The question relates to whether much of the routing and contextual business information can be part of the initial message from which the client could then use for all updates that follow in the ActiveMQInputStream.
To add more information to the reason of the question, I am looking to stream a large number of updates to remote clients. Here, bandwidth is of high significance... I am looking for the most efficient way to distribute many small updates and provide the highest possible throughput to the client. One way is to increase the bandwidth, but this is typically out of my control. Alternatively i can aggressively reduce the amount of data sent over the wire for each update. Reusing routing and context information is just one way that i could think of to reduce the amount of data required for each update... I am not sure of the underlying implementation, perhaps JMS Streams are simply mapped to multiple JMS Messages anyway (is that what you are suggestion?) and there is absolutely no benefit to gain from my suggestion... I was looking at the JMS Stream - namely the ActiveMQInputStream and ActiveMQOutputStream as some sort of stream forwarding feature that is created between a producer and consumer of a specific message... Am i way off base??? -- View this message in context: http://www.nabble.com/Use-misuse-of-JMS-Streams-tf2037623.html#a5608253 Sent from the ActiveMQ - User forum at Nabble.com.
