Particle Mesh update — a note from the CEO

And this is what I love about Particle, and what I will sorely miss with mesh. I can relate to all the problems you must have had trying to figure out how to write a generic toolset to fit the myriad of ways of handling the CAP connundrum in a mesh. There are so many questions which must be customized on a per-deployment basis, from the trivial (what does a successful mesh publish mean?) to the complex (in a sleepy mesh, how can I distinguish between a partitioned (e.g. dead) unit and a sleeping unit?).

I love that in front of me right now are three Argons, two Xenons, and a Boron, and they all work with the same API, allowing me to abstract out the physical nature of the cloud/mesh connection. They are all running identical code and it is completely transparent, and I don’t have to have specialized knowledge about each one’s PHY capabilities. This is a unique and beautiful thing, I’m so very sad its days are numbered.

P.S. @zach, thanks for your answer.

11 Likes

I’m going “against the grain” somewhat, but I’d like to say Thank You for your honesty.
Particle tried to push the envelope w/ 2.4GHz Mesh…it didn’t work out.
@zach did a great job explaining the reasons for this unfortunate situation.

Sub-GHz Mesh Networks are available and mature… just not with a easy method of OTA Flashing for the endpoints (which was the main benefit of Gen3, in my opinion).
Existing radios can be incorporated into the Particle Ecosystem pretty easily.

The Photon and Electron eventually became outstanding devices.
I’m optimistic and look forward to seeing the Boron’s (and Argon’s) firmware evolve likewise, now that resources can be focused.

10 Likes

Guys, i know it sucks for some of you but ease up on @zach. Particle could have done what many other companies have tried to do and just keep milking the product along trying to squeeze out every penny from it while the customer deals with all its issues never getting and real fixes.

Instead, particle realized its problems and that it was just not up to their standards to fix its issues, so rather then keep leading people along they shot that dead horse.

As engineers you all should know that anytime you use a module from anyone your always at risk of them pulling the plug so this is not new territory. Its just the way the bottle spins.

6 Likes

Unfortunate news for everyone. Pulling off the band-aid is painful, especially for those not expecting it, but hopefully those resources can be deployed in a way that is more productive for Particle and the developer community.

2 Likes

They currently have the Xenons reasonably usable and at least OTA updatable. They’re never going to get a properly working Mesh system, but Particle could do a lot more towards keeping the Xenon hardware usable in a more limited role than their original ambition. There’s no reason for them to essentially brick perfectly good hardware that already has a decent level of functionality.

And most of the time when you use modules from third parties, you’re at risk of the product being discontinued. You’re not at risk of anyone pulling the plug by flipping a switch and disabling all your already-deployed products. That Particle is willing to do so shows that using Particle at all is suicide for any company going forward.

Particle likes to brag that using their cloud simplifies IoT development. Well they’ve just amply demonstrated that using their cloud is a gigantic mistake for any company.

2 Likes

“ease up” does not work well with my investors nor my customers. As I’ve said above, this should have been developed way further before going out with big messages such as these: https://www.youtube.com/watch?v=QwifCetjoFU. and https://www.youtube.com/watch?v=54kmDEoQSP0
and https://www.youtube.com/watch?v=HJyHmMJY6as

I am fine with companies trying out new tech and bringing this out early. But that should be guided with an appropriate message without giving the impression that it is all done and it is now just up to you to make something with.

Yes I have dealt with a wide variety of components big and small going into lifetime buy. Never pleasant but manufacturers generally help you transition to a new device. I’ve had a very unpleasant experience with Motorola when they pulled an MCU and within 6 months pulled the backup MCU. So we removed Motorola as much as possible from our designs - can’t work with a party one cannot trust. Particle’s attitude in my experience is similar. Where is the P1 Zack? Be clear about that also.

2 Likes

That Particle is willing to do so shows that using Particle at all is suicide for any company going forward.

Particle likes to brag that using their cloud simplifies IoT development. Well they’ve just amply demonstrated that using their cloud is a gigantic mistake for any company.

Exactly! Particle is for hobby projects, not for companies solving real world problems.

5 Likes

I am totally agree with you @OracleJ
I’m also feeling like the same that we can’t trust anymore and frightened for 2nd gen devices.
I was super fan of Particle Team since Spark Core but now I realized that I have to find the alternative of Particle & their service.
Sorry Particle! if seem bad but I need to find the alternative for backup.

This approach of offering a flexible and supported board with Lora would also eliminate the idea that the Bussiness model was not working. I like to point out that for us the mesh meant we could develop a product I was dreaming about for years. We am for consumer markets. This means you have tech people and you have people that just want a IOT device to do one thing. Paying the additional cost of a bridge is just not the way to go for the latter, at least not at the beginning. Take a look at Philips Hue that now offers a BLE and bridge free solution too. In our case a differentiator would be that our product could be bought with optional internet. I had many discussions with particle by mail and on the phone about the Bussiness model behind this. It was clear that the use case of Xenon’s posed a Bussiness model issue for Particle. So I appreciate the openness about the technical problems and I can relate to the issues with OpenThreads bandwidth in the 2.4g spectrum. But a simple solution to the bussiness problem would be to offer BLE updates to xenons that our not connected to a bridge. We would pay particle for the BLE updates for those customers that just update by bringing there phone near our product. Particle could gain and we could continue with our product line up. We are rethinking our product opensforyou.com but I would appreciate to know, like @kubark42 mentioned, why bussiness model creativity was not explored, or at least not explained to us. Engineers understand, and I actually love you for it, that you kill your darlings. I just hope you do not use my love as an engineer to obscure the bussiness model lack of creativity in the rip xenon line-up.

1 Like

Last message was a reply to one of mine so I thought I would clarify, I’m not suggesting doing LoRa at all, the low bandwidth of LoRa makes OTA very difficult. So OpenThread that Xenon and all 3rd Gen particle product is built on is based on IPv6, 6LoWPAN, IEEE 802.15.4 and the radio technology that this all runs on is available in 2.4GHz and Sub GHz bands.

What I would Love to see is the mesh stay in the product line and use short range via 2.4GHz where appropriate and have a separate daughter board with a Sub GHz* radio that talks to the 3rd Gen boards over SLIP and integrates with the OpenThread stack.

I think it doable. I was a backer of the original Spark Core Kickstarter and I’ve watched Particle do an amazing job at open source and community development around a great IoT platform. Mesh is hard let’s not give up on it, just lower the commercial expectations and keep working at it with the community.

*The bands do differ from country to country based on available licence free spectrum which is why a frequency isn’t stated, this has its own challenges but not insurmountable.

2 Likes

Although I can agree with the reasons to discontinue Mesh, this announcement effectively destroys for a great deal the badly needed trust in Particle as IoT Solution Provider. For two reasons: 1. Making big announcements over Mesh and then finding out “it does not work” means that the technical homework was done inadequately and 2. Obsoleting hardware where people including myself already have invested in does not show much compassion with clients or the situation they are put in now. Luckily we are not depending on Particle’s cloud solution-what would happen if they would pull the plug on that one?

Come one, Particle, you must do better than you are doing now!

1 Like

Sad thing, but I personally think it’s better to discontinue things that don’t (ant won’t at all) work as expected than continue them without getting reliable (and suitable) networks that match your (and your customers) desires in quality.
In my case my planed thing can also easily be done with Wifi or Cellular devices but I found the mesh-network an interesting way with slightly cheaper hardware.
But overall I think it’s better to focus on less things that can grow than “wasting” resources on stuff that will probably never work as expected.

How is the refund working? Do I get some money for the Xenons back or are they “upgraded” to Argons/Borons or something?

I’m just thinking of an easy to deploy Argon that anyone can connect to his Wifi easily :wink:.

Thank you for understanding. You summed it up - Mesh was not delivering as expected and could not deliver the quality we are aiming for.

In regards to refunds, you will receive store credit that you can use in the future.
The Boron and Argon both have extra hardware on to allow them to connect to Cellular/Wifi, so it’s not possible to upgrade your Xenons.

The way I see it is not that “it does not work”, but the amount of effort it would take to get it to where it would “just work” would be heavily disproportionate to the amount of demand there is from the market.
The fact that most customers then want a solution that’s also outside of the physical limits, for their intended purpose, doesn’t help. Mesh as a concept does work, but expecting these to cover 1000ft each isn’t feasible unfortunately.

It’s not that it isn’t doable, but more so the reward vs effort is not realistic. This is especially true when improving Mesh would mean neglecting the other core cellular and WiFi products, which have an undoubtedly bigger market share.
There’s potential in Mesh for sure, but it’s currently a lot less than the other solutions, yet would take a lot more work to get it to work remotely as well.

That’s a bit sad, but in fairness they do provide a couple of alternate uses for your product, as opposed to outright bricking them (like others might do/have done). It’s also ‘just’ the xenon that will be going away, whereas the Argon and Boron will still be first-class citizens.

For me personally, I’d rather have wifi and cellular working well and being maintained, than have all effort spent on getting Mesh to a semi-okay state which might still be inadequate for many people’s use cases.

Bittersweet, but I get it.

2 Likes

Oh that was more meant as physical replacement of the devices, that they are different on the hardware side is quite obvious :wink:, because of that there were the quotes around “upgraded” sorry for not being exactly :smile:.

I have been a professional software developer for over thirty years but with no formal training in electronics or engineering. I am frantically building a series of products to monitor and manage flooding along the Great Lakes watershed; my mission is huge and time sensitive. But before I sold my project to investors, or even myself, I invested hundreds of hours working with Particle products to understand the benefits and limitations of the technology.

As a hardware novice (IMHO) I feel my achievements with particle products have been amazing so far. I started with cellular Electron, then wi-fi Argon and lately cell/mesh Boron with Xenon’s as routers and end points. I have spent a great deal of time exploring and quantifying each device type and I was only weeks away from defining my ideal hardware spec. However, as I narrowed down these technologies with scaling in mind, I found the exact same limitations to the Mesh approach that Zach has explained fairly well. The Mesh devices have been an excellent tool for guarding my home and my business property but the process of scaling it to vast numbers with ease of use was just not quite happening for reasons stated in this thread.

Zach’s announcement therefore does not come as a complete surprise and in many ways I am grateful the company has chosen to proactively pull the plug on what could have been years of frustration for all. I can’t tell you how many hours I have spent on software platforms that have gone into the waste bin. As an Apple devote, even before the Macintosh, I can tell you that Apple has abandoned and wasted far more of my time and work than Particle has or ever will; all software and hardware is an evolution.

With all due respect to some on this thread, I have had tremendous support from Particle despite currently being a low volume user. From one on one phone consultations to the Particle Community resources, I cannot fault Particle as its support has been helpful. By contrast, try calling Apple or Adobe and see how that works out for ya!

I very recently ordered and I am receiving this very day the last 8 Xenons available from Adafruit.com so yes I do feel the pain but I am not willing to “throw the baby out with the bath water”. The company has supplied a way forward and if I get another year from these devices I am doing better than many of the tech products I usually buy.

Particle is not the only company doing IoT products and as a business person I am bound to explore all possibilities - doing otherwise would be a disservice to myself and my investors. But I would like to see Particle succeed and its products evolve and refine further. I want Particle to focus on making a ‘ready for prime time’ device that I can build into an IoT product and I need it to be rock solid and scalable; if Mesh is not that solution so be it - I deeply hope we can move on…

14 Likes

@zach Yeah, thank you very much for luring us into this, how bloody annoying! I am (maybe “was”) a big fan of Particle/Spark. In fact, I was an early adopter, starting with the first Spark Cores and supporting the projects already on Kickstarter. I was very excited about the mesh functionality in the 3rd Gen devices. In fact, I got them as soon as they came out and started to decommission one by one the various Photons with which I had built up my home automation system and replaced them with Xenons, Argons, and Borons. The mesh functionality was replacing and greatly facilitating the device-to-device communications (without going through the Particle cloud) which previously I had implemented with Xbee modules connected to the Photons. Also, the new devices didn’t require additional hardware (Xbee), which made expanding the system much cheaper. What we are now left with is a huge disappointment. It gives me the impression that the decision to embark with the Mesh technology was taken prematurely: a year ago, Mesh was THE big thing and now you are simply pulling the rug. Well, true: all Argons will be slightly better Photons, and the Borons slightly better Electrons, but the Xenons, deprived of their Mesh functionality will just be downgraded the status of a generic Arduino clone and not even being able to use them in the Particle cloud is so annoying! Of course, I am just a hobby enthusiast, not a corporate client, so why should you care? Yeah, you sent me today a generous $20 credit note: thank you very much. If you look in my account, I have 13 3rd gen devices of which 8 are Xenons, and this is not counting several Xenons and other Borons I still have not registered with the cloud. What am I supposed to do with the generous gift? Have faith in the future of Particle devices and shell out my $20 to buy more devices that tomorrow will be crap when fickle Particle realises again that they have gotten things wrong? Not sure about this one…

1 Like

Particle’s fleet and product management is excellent, we have seen great value using it for Photon based products. If your management tools could work with other microcontrollers ( ESP32 for example) and decouple from hardware, you would see a lot of companies flock to your management tools with the latest surge of IOT device management. Each product has different hardware needs and a Photon or Gen 3 device won’t fit every application , but your management tools would be very powerful for more devices.

3 Likes

While I understand your sentiment, your Xenons won’t be killed off as such.
As long you stick with device OS <= 1.6.0 your devices will still be able to talk to eachother even after sunsetting mesh and your gateway device will still be able to communicate with the cloud.
So not all is lost.

2 Likes

Yeah, true, but there is not going to be any way to add a new mesh point, or maybe replace a defective device, very useful.