Since it seems everyone has moved on from the Panda II and WIZNet, I took the code and updated it for NetMF 4.2, the G120 and ENC28. While I was at it, I also made a new SmartEnergyGroups site. So now, every 15 seconds it sends an update to both ThingSpeak and SEG.
You can see the SEG dashboard here:
https://smartenergygroups.com/groups/network_shootout_part_deux
The ThingSpeak channel is still here:
I wonder how well it will do. The Panda II was still working fine after at least a week.
Today I got an email from Smart Energy Groups saying that this device had gone offline. Sure enough, it had stopped transmitting. I was able to break into the running program and it was stuck waiting to connect. This is likely the long standing infinite timeout on connect bug. I did not implement the workaround for this issue, so this error is expected. Before it hit this issue, it had completed 11422 iterations. Since each iteration includes two internet updates, it had completed 22844 updates with only 10 errors.
So, I would say there does not seem to be an inherent problem with networking in 4.2, even using WebRequest instead of a roll your own socket implementation. This is consistent with my experience with 4.1 and 4.2 using a socket based implementation. Once I removed the errors in my code and solved the issue with timeouts in the http code in the WIZNet driver and the connect timeout, the system was quite reliable.
I don’t know what happened in the first shootout, but I might try that code and see if I get the same results with it.