I thought I read somewhere in the documentation that if multiple requests are made to the DataWedge API, some may be ignored if the API is busy handling other requests. However, I can't find this anywhere now that I'm reading back through the 6.8 API documentation. I'm deciding whether our application should serialize requests to the API, or allow them to be parallel, and what a reasonable resend time gap might be for unanswered requests.
Any advice is helpful. We are currently writing for the MC3300 and the DataWedge version we have on our test unit is 6.8.50.
Simultaneous DataWedge Requests |
1 Replies
Never mind. I was basing this question off of the response data that is received when a command like 'SET_CONFIG' is sent to the API, because it returns the command ID with the response. Since there is no such response data included with request commands like 'GET_VERSION_INFO' it is not possible to safely allow concurrent requests, so I'm going to serialize everything.
As a side note, it would be helpful to add command ID to all responses, no matter what type they are. This way a more robust messaging layer could be built between client apps and your API. However, I will confess that our application doesn't appear to need this at the moment. Serializing the requests should work fine for us.