For completeness, and for others who aren’t aware: when using the Solace broker as the JNDI lookup (rather than a separate external JNDI provider, as is more typical in production deployments) the Solace JMS API makes two client connections: one specifically for the JNDI connection factory lookup, and then a 2nd one for the actual data / messaging connection. You can see this if looking at the broker’s event.log. It sounds like @arm was trying to figure out a way to ensure the initial JNDI connection (1st one) uses the compressed port 55003, not the default uncompressed 55555.
BTW, @arm: why are you using the compressed port? Connecting over the WAN? Lots of easily compressible text payloads? Just curious, as using compression increases latency and CPU utilization, but can improve data throughput by being more efficient at network bandwidth.