An EAAS client subscribed US stock data, 10K, using both rmdstestclient and EMA Java based App. One of his testing cases is to use printed timestamp to check if any delay between rmdstestclient and Java App in receiving the same streams. Hundreds of milliseconds delay has been identifed in Java App's output comparing with that of rmdstestclient. Can you have a look at attached code to see where to tune to get Java App close to rmdstestclient? Thanks.
I understand the situation and the concern, and sympathize that the custom Java code is likely creating the delay in processing that being observed.
However, this forum is dedicated for general questions and discussions of API usage.
To review the custom code, and to recommend what is likely to be a comprehensive set of changes is not possible in the framework of the discussion forum.
I can see several feasible ways to proceed to help the customer who has written the code:
I hope this information helps?
In addition to resources provided by Zoya, I would like to point out that EMA is extremely powerful and flexible API, but not the fastest one - certainly not as fast as low level ETA, which is probably what rmdstestclient is written in.
If the primary need for the client is to have a lowest latency application, then please refer them to use the ETA API.
The attached client application is a complicated one. The latency might well be in the user application code.