I’m curious for any community feedback (pro’s/con’s) of using Reverse Invoke, in particular in regards to potential constrains or issues.
We’re planning some major infrastructure/architecture changes and are considering moving all of our public functions into the corporate network. It is my understanding we could do this with the least amount of development effort by simply mirroring our Public (DMZ) hosted servers on to the servers hosted in corporate network, and converting our existing DMZ hosted servers as the reverse invoke servers.
Are there any known issues or experiences regarding performance/availability - caused directly by Reverse Invoke? Also, any limitations with particular adapters (RosettaNet, Ariba, C1, Web Services, WmDB, etc…)
Our document volume is relatively low, maximum of about 1000 per hour per server -80% HTTPS/20% HTTP, with about 30% of that volume being small documents (<100kb),>100KB), 30% being large (>300KB), and ~10% being very large (>1MB).