Hubitat vs Node-RED response time

Fully understand and agree with the general premise. However, if a person has a LOT of devices or runs a LOT of apps and automations at high frequency, the second hub may be necessary. The Hubitat Hub is designed/built for the anticipated 90% (guess) household usage. I agree that a professional Hub should be considered as a future offering with more Lan, Zigbee and Z-wave I/O and processing capability. But that is a business-based decision for the Hubitat staff.

I :yellow_heart: :yellow_heart: :yellow_heart: my Hubitat, the Hubit Staff, and the User Community. :sunglasses: :sunglasses: :sunglasses:

3 Likes

I've been overall happy as well. While I use HA and Node-Red for my logic, control and dashboard, HE does still have good radios and all-around support. I have found the z-wave radio a little weak on it's own (ie having only 1 or 2 devices at a far distance), but with a properly built mesh it's a complete non-issue. And I think with the new c7 hub and the 700-series chip that should be much better. It actually gives the hub a leg-up on most of its competition.

My worry is, if the second hub is presented as a solution and the slowdowns continue, it's going to push those 10% away. And those 10% that are "power users" are usually the ones pushing the platform forward and developing new drivers and apps. I've seen a lot of them start to jump ship. When they're all gone, what's left is the "average" users who have already bought a hub and will never touch it again. They won't continue to buy hardware and then what? With diminishing revenues it's hard to stay operational and offer support, and it's just a downward spiral.

I'd love a "pro" hub, I'd even love the option to buy a software license and run on my own hardware. But that could quickly become a support nightmare.

I'd like to comment on this, as it comes up regularly.

We have tested this, as I would hope would be obvious. We learned two things: (1) For most automations it makes no difference what the power of the hub is; the hub's overall performance is ultimately limited by devices and mesh network performance, things that have nothing to do with the hub itself. There is no such thing as improved "Zigbee and Z-Wave I/O capability" other than breaking these networks down into multiple networks running on separate hubs. (2) We were sufficiently disappointed with the lack of improvement overall, that we were fairly convinced that anyone paying more for a more powerful hub would not be happy about the purchase.

The business case for a more powerful hub is very weak. The number of users who would benefit from it is small, a small percentage of customers. The cost to produce and market such a product would not yield desired financial returns to justify doing it. The opportunity cost to our business would be high, when we are pressed on other fronts. All in, it's a loser for us. That coupled with our experience of the actual level of improvement in performance argues pretty strongly against pursuing this.

Having said that, we do intend to pursue steps to improve hub performance. Some of these are already done, and will show up in the next release.

12 Likes

Thanks much, @bravenel, for the candor and insight! :clap: As I'm sure you're aware there has been a great deal of speculation and questioning around this in many topics. Here's hoping you've laid most of that to rest. :+1:

2 Likes

It's a perennial request. I certainly understand the thought process that leads to this desire.

There is another desire that is evident in the power user group: Being able to make Hubitat be an end-all home processor, running everything that one would want to run that controls your home. This is a very understandable desire; it's just not one that makes sense economically for a business to offer. Small audience, high investment, high support costs... If this were a great business opportunity, surely by now someone would be doing it, don't you think?

Our Hubitat Elevation product is not intended to be such a beast home processor -- it's intended to support mainstream home automation needs. Bear in mind that home automation itself is pretty much bleeding edge technology, not really suitable in its current form for mass market adoption. We've got voice controllers from two of the mega tech companies, but no home automation solution from any of them. Why is that?

4 Likes

As long as Maker API continues to be performant and supported, I have no complaints about Hubitat. :+1:

And I'm very much looking forward to getting a 700 series zwave hub, too... Once available.

5 Likes

You're one of the power users who has figured out how to solve the problem you want to solve in your home, using Hubitat Elevation.

Maker API is not going anywhere -- we are making small improvements to it from time to time, including in the upcoming 2.2.2 release.

14 Likes

Could I ask if I have the correct interpretation of those words?
Are you stating that if we have two hubs for example, running a zwave network on one, and a zigbee network on the other, would likely produce significant performance gains over one hub running both networks?

Fantastic! Bruce, I don't recall this level of assurance about Hubitat's future in any topic and I'm sure I'm not the only one to derive a lot of comfort from your words today. :+1: :+1:

1 Like

No, I'm not saying that. I'm simply pointing out that the performance of a mesh network is independent of the hub. Some people have found advantages to splitting portions of a Zigbee network off to a separate hub. Zigbee bulbs make horrible repeaters, for example. So a unified Zigbee network can have performance issues, issues that depend on the actual devices used and the mesh topology.

2 Likes

I'm still confused given the discussion above how I can get faster motion lighting using NR than with RM if hub cpu performance is not capping performance and it's really mostly about the mesh. I'm using the exact same mesh (eg. Motion sensors on HE -> NR -> HE -> Hue) and it's noticeably quicker (with a level playing field in terms of everything else going on on the hub of course). It feels like RM needs a significant optimisation or a faster processor driving it. But what do I know :man_shrugging:

Either way, I love my HE (despite the various issues I've had) and in combination with NR it's now dynamite :grinning:

1 Like

I've given up trying to figure out that part. I can't see the app source code, I can't see the OS, and I can't get in the java vm so any guessing as to why it is a little slower than doing it w/Maker API + Node-RED would be exactly that - guessing. Which doesn't help me improve anything.

That said, I am 100% comfortable in using external logic engines, and I have no issues continuing to do so. I am going to have my docker swarm and NAS+VMs anyway for other uses, so it isn't like I had to 'stand up' any extra infrastructure to do it that way.

2 Likes

I understand that and agree with your approach. I for one don’t use ZigBee or Z-Wave much by deliberate choice and lack of need. But I do use HE a lot as it is very easy to code integrations for which, for me, are utilities or IP based I/O typically. Aside from the deficient UDP support it’s very capable.

So for my usage I’m not I/O bound and processing and memory are desirable. But I’m a very small minority % base of your users

Lol this was me. When I got my HE I quickly managed to get everything running on it locally. I quickly bounced off of the performance issues many have seen and was quite dissapointed. It wasn't till I started offloading everything back to my pi that my HE settled in and became reliable again.

I was annoyed for quite awhile then realized I just needed to adjust my expectations. HE is a awesome little hub. Just not the one hub to rule them all that I was hoping for.

1 Like

Because they have not found a way to completely control/monetize the market the way they want..... yet.

I suspect that the technology is a secondary consideration - they "could" play nice with existing standards and go from there but have chosen not to. They certainly have the resources.

Yes, but why is that? Not for lack of seeing an opportunity certainly. They've had years to ponder this. This is a difficult problem, one that is not amenable to a nice centralized software solution. There is no way to get consumers over the hump of figuring out how to automate their homes. Someone has to do that for them. Just like we all do it for ourselves.

Why did Google walk away from Revolve? Why have Nest, Ring, Echo, Google Home come no further than point designs? Not for lack of resources...

Not that anyone asked, but I agree with 100.0% of that.

Revolv seemed like such a cool hub from the hardwire side of things. Oh well. :frowning: While I would be happy for the owners if Google threw a bazillion $$$ at Hubitat, I would be bummed for me because I would expect it to go the Revolv route.

I dunno but I wouldn't count Amazon/Ring out on that score - their eco-system tied in with (and thanks to) the Echo platform is growing albeit from the "home security" angle.

Google has a history of cancelling things both hw and sw (fun website - killedbygoogle.com) - so that seems more a symptom of their culture - not sure those tech platforms could be expanded to do what they wanted to do short term and profitably.. just a guess.

The problem for all of them is since there is a huge number of inexpensive semi-interoperable devices (good and bad) on the market to attract consumers - wifi/zigbee/z-wave/bluetooth/etc... things are diluted and fractured. China has jumped in with their offerings and even Samsung cannot be counted out. Easy to have a presence maybe but hard to dominate.

All I'm saying is I'm not convinced the issue is completely technological.

This says a lot right here and should be emphasized. It does say a lot about the people running the business. Lots of respect for this sentence.

I'm in the same boat here as @Angus_M. If the hub has enough power, how does moving the automations off the hub and then sending it back using the same mesh networks bring a speed increase? The answer may be unknown at this point, and I'm sure staff is looking at it regularly. I've seen theories, some make much more sense than others. If it's as simple as an app such as RM is "heavy", one would think throwing more power at it would help it.

2 Likes

I’ve given up trying. I went as far as creating my own groovy applications.

I personally think it’s a combination of the hub and apps. More apps, more rules, you will see small slow downs off and on plus inconsistency. At least for me.

I’ve built the hub from scratch several times. With improvements but not to my satisfaction.

I’m done trying. Wasted over a year on this. The hub is better served as a gateway for me.

1 Like