We have noticed that again there has been more "Operation timed out" errors for app-to-tag communication, like arm/disarm, apply configuration changes or beep. Upon investigation using network packet analysis, we have blocked several IP addresses that were flooding our server with invalid IP packets. Also we have lowered the "ping frequency" of tag managers to reduce the amount of non-essential traffic. As a result, now the latency from app to tag manager is much shorter than before (always <1s and without timeout error) and the majority of app-to-tag communication latency is now from the tag radio receiver delay, which is programmable from Wireless Settings and in trade-off with tag battery life. 

You can test the intrinsic latency from app to tag manager, by using Debug Console (after you login) where you can use the "Get" button to read tag manager register values chosen at the drop down, e.g. AFSKCTRL, RSSI, etc. note the result only appears after 2nd read.)  The delay from clicking "Get" button to the result appearing, equals to the delay of app commands like "beep" minus the actual tag radio receiver delay

UPDATE (Aug 29) We have switched to a completely new IRC server software that uses "Windows I/O Completion Port" to handle large number of connections. These timeout issues seem to be completely gone now. 

‹ Go back to the blog

Comments

Leave a comment