100% this.
Plus, it would be insane for Hubitat (or any other manufacturer) to downgrade the capabilities of their hub to support Iris V1 products, which haven't been made or sold in close to a decade!
Iris V1 products date back to 2011-2012, and were replaced by V2 products, that are ZHA compliant, in 2014-2015.
Pretty sure they’ve indicated that iris v1 compatibility is over as of the C8 hardware revision. With all due respect to iris v1 device users, it would be a waste of developer time for the rest of us.
I don’t know how relevant it is, but I’ve recently been trying to figure out how I can only use 2 hubs with Hub Mesh with close to or over 200 devices on the “source” shared to “destination hub”.
The Hub Mesh docs say:
When more than approximately 200 devices are shared in a Hub Mesh, there may be concern over the increased amount of network traffic and load on the source hub. You may therefore disable or extend the amount of time before a Full remote device sync is performed.
The think the default for the Full Sync was recently changed from 2mins to 1hr.
Not sure if it was determined to be a bad default due to general load or not.
So radios and new capabilities aside…
I’d just like to have better internals in general.
I’ve been doing this for years. I have 4 production hubs split by radio (Zigbee, Zwave, and lan) that are all meshed to a coordinator hub where rules and apps are running. My coordinator has 330 devices so I am well past the 200 device limit you mention. Memory on this hub is lower than the others but I can go weeks without having to restart it.
The default 2min was found to be contributing to memory decline during 2.3.7 beta. The old default was originally a band-aid for sync issues but since Hub Mesh was moved to use mDNS and TCP it is much more reliable that when first created, so the full sync is not really needed anymore. A bunch of people in beta reported they had it totally turned off and did not see any issues.
Thanks, that eases my concerns a bunch on that subject.
Thanks good info
Does it improve performance much to separate all radio functions from rules and apps? Currently I have one C5 hub running all rules and apps and also zigbee. Z-wave is running on two dedicated C8s placed in the house to optimize coverage. As noted above I did intend to move zigbee off but ran into too many issues with zigbee and the C8.
I have been a HE customer since Feb 2018 just after they released the original C3 hub. After moving everything from SmartThings and adding a ton more devices I found a single C3 hub couldn't handle everything so I made the decision to split things up by protocol and was using HubConnect at the time to sync everything to a coordinator hub for dashboard and Homekit and rule purposes.
Fast forward I upgraded the C4s to C7s and my OG C3 is my development hub and left the protocol split in place. I had rules running on the Z hubs where it made sense to do so - zigbee motion to turn on zigbee switch vs having to traverse to another hub.
When the C8 was announced/leaked FCC I considered combining the Zwave and Zigbee hubs and I moved all rules from my Z hubs to coordinator in preparation for that. LAN still has a bunch of apps and rules running there specific to integrations. Now both Z hubs are radio only and no rules or apps. With all the Zigbee issues that folks were having on the C8 I actually never migrated and still have the C8 sitting in the box. I have it on my todo list to migrate my Zwave hub to the C8 but just haven't gotten around to it especially since I added external antennas to my Z hubs a long while ago so the C8 doesn't buy me much.
But to answer your question memory and CPU are pretty steady on my Zigbee and Zwave hubs. When I put the hub into Night mode an app turns most everything off and splitting things up to different hubs has certainly sped that up and I don't have many missed off commands like I was getting way back on my C3 before the split. But knowing I have around 100 zwave and 115 zigbee devices a single "radio" hub will likely not work for me and I may need to always keep them separate. But again having no rules running certainly makes things snappy on it.
Interesting information. I missed that during the beta and actually still have the sync turned on though the default sync interval now is every hour vs 2 minutes.
I separated devices from apps when i had a C4 brought to its knees about every 4 days, I have since upgraded to C7”s and 8’s. I still have them split, but if I wanted to I think I could go back to a single hub. Having said that I now have a third hub for cloud integration and a fourth looking for something to do.
Changing the default from 2 minutes to 1 hour was the "quick fix" for the high rate of memory loss. While I assume that the cause of the loss was eventually fixed, I have left mine to Never as the sync really serves little use in normal operation.
Makes sense and I agree. But I sometimes have to restart my coordinator hub because it runs lower memory than my radio hubs and sometimes gets down to 100k. I get around 10 -14 days before I get super low and have to restart.
I would need to invoke a refresh a bit after the coordinator hub restarts to sync with the other hubs in case things change during the restart. Inspecting the "Sync Now" button I see there is a end point I could invoke from a rule:
hub-ip-address/device/hubMeshFullRefreshNow
I will set that up now and see if memory improves.
I don't use Hub Mesh so I could be mistaken, but I thought I read somewhere that a full sync is done automatically when a hub is rebooted. Can anyone confirm? In that case you would probably not even need to make a rule, you could just disable the sync, or set it to a very high interval.
FWIW, I had issues w/some device/automations when I turned off full sync on my C7, which was hosting Linptech mmWave presence devices used in automations. Things like lights not turning on, or not turning off.
Re-enabled 1 hour sync on C7 and things went back to normal, no other changes so it did seem like having sync off on one of my hubs (left it off on C8 where the automations were running and which held the Z-Wave light switches) did seem to cause issues for me. 
Hourly interval is the max
I setup a rule to invoke that URL I posted above 10 minutes after restart just in case.
but you do. Your thinking of it wrong you think you dont need to hubs for processing etc. However think of it as a bridge for your old crummy devices and a new hub for all the rules/ processing and all new "good" devices.
I had two C7s now 1 C7 and one C8. (spare C7 for testing) The live C7 will just live on for the purpose I brought it for which was to bridge my old crummy LL and HA 1.2 devices (and originally old 300 z-wave devices) to my "hub".
Having a "bridge" hub was much cheaper than replacing the old devices and means that all devices both old and new are happy as Larry.
dont agree. i have over 200 devices and many many rules and apps on one hub and it is working fine..
average cpu load is 7.5 %
I believe that a sync is done automatically upon initial connect, and I have experienced no issues with it being fully disabled. However, only @gopher.ny can say for sure. 
How to see average cpu load?
Is this typical for speed loading a rule? This is what prompted me to ask the question in the first place:
Download the Hubitat app
