We just fully released a long standing feature we’ve been calling “fastOTA”. It should significantly speed up over the air firmware updates for most, especially if you are international, or have higher latency on your connection. If your device is running 045 or later, it should be enabled automatically for you.
This means that:
The previous long magenta update on a new Photon should now take less than a minute on standard Wi-Fi.
The “blinking magenta” state that you are used to seeing when flashing user firmware to your Core or Photon may now appear as one or two single magenta flashes, due to the speed of the update.
fastOTA speeds things up by sending all the chunks of firmware at once, instead of waiting for an acknowledgement between each packet, then the device can ask for anything it missed at the end.
So hopefully this means much much quicker updates for all
thats crazy how fast they update now! i wassn’t sure if it flashed properly it was so quick, but the program is running well now all we need is to be able to set a firmware to flash as soon as the device comes online (possibly out of sleep)
I believe it works by blasting all the sectors without waiting for a response for each one from the Photon, then at the end the Photon can ask for any that it missed. Basically they took the overhead out of the protocol in the normal case and added overhead in the case of failures.
Ive been using Particle since the Spark Core Kickstarter and I’ve ended up doing all my development localhost upload by USB for speed. I’ve got a project at the moment that requires developing via wifi and it’s fantastic. It takes longer to reboot than it does update the code.
I just wanted to add that we rolled this out for the P1 and for all product creators as well today! If you create a new product it might take up to 4 hours before it ‘switches on’, but eventually that delay will be eliminated. Thanks everybody for their patience, we used the time to do extra testing before releasing it fully.
Good question! I think it would be nice to share that functionality with the local cloud certainly! So far it hasn’t been urgent since I imagine local cloud implementations are typically hosted either locally, or on servers close to the devices in question, which should give them very low latency. But definitely something that I think should be on the roadmap. I don’t have a timeline for that feature yet, but I’ll post more when I know more!
We’re starting to dream up some big upgrades for the local cloud as well, but if I had to guess those are probably slated for Q2-Q3 2016.