Kazoo Version 5 open source

[RuhNet]

Thanks so much for the update, @Chris Burgy! That's [mostly] good news.  :)

Incidentally, [please confirm with your lawyers of course, but] there is no reason why you must stop using the Kazoo trademark: The other company who successfully registered it declared their first use of it as 2019, and 2600Hz's first use was long before that. Contrary to popular belief, trademark rights don't come from registration, but from usage. Since the other company registered "Kazoo" as a trademark for their forum or whatever service, you probably can't sue them for use of it (well, you can actually, but it's not as simple since they do have a registration), but that doesn't mean your rights to continued use of the "Kazoo" trademark is limited in any way, since it pre-dates the other company by several years. Just an FYI. 

[Chris Burgy]

We did go through the motions with our legal team (obviously we have inside counsel with a handful of attorneys and outside counsel to also consult). Unfortunately, because 2600hz let the trademark lapse a long time ago, it's a sticky situation. Is what it is and just one of those things we had to roll with the punches on. We even had marketing look at other naming conventions but that's a whole different situation with trademarks. I think everyone that knows 2600hz understands what it was and new customers don't frankly care about the Kazoo branding. They are paying for a solution from 2600hz so that is good enough for them since it's all hidden anyways. I appreciate the comment though. We wished we could've continued on with the platform name we all have known over the years. 

[RuhNet]

18 minutes ago, Chris Burgy said:

I think everyone that knows 2600hz understands what it was and new customers don't frankly care about the Kazoo branding. They are paying for a solution from 2600hz so that is good enough for them since it's all hidden anyways.

@Chris Burgy Absolutely, would be nice to keep the name, but I doubt anyone really cares. 

Minimal or no documentation is fine; we can figure it out from the source code. :-)

As far as some [many] modules not being part of the open source release, that was expected, but can you tell us if the release will be complete as far as having all components of a working system, [albeit possibly barebones]? In other words, are there any critical parts that will be missing for a standard PBX that we (the open source community) will need to add?

Also, is this release going to be a "one time dump and then move on" or will the open source public repos be the same codebase used internally for your developers' continuing work on the project (like Kazoo was prior to v4.3)?

Edited 17 hours ago by RuhNet (see edit history)

[esoare]

Cool! - Keep up the good work (in my opinion). 
Really neat to see (from someone who doesn't use OS) that you guys are sticking to your word! 

Looking forward to those apps! :D

[synogeek]

3 hours ago, Chris Burgy said:

We officially approved releasing an OS version of 5.4 as a Senior Leadership Team about 3 weeks ago. Product management is working with engineering on this as we had to split the OS and closed-source repos. As you probably know, 2600hz pivoted years ago to having closed sourced / commercially licensed components as the company couldn't survive without generating revenue and OS wasn't doing that especially compared to the fully commercially licensed competitors of 2600hz (Netsapiens, Alianza, Broadsoft, Metaswitch and RingCentral/Intermedia/etc coming from the UCaaS space with a pure reseller model). All the closed source modules will remain such with an increasing number of those ported from the Ooma side. To be transparent, Ooma's priority has been porting over our intellectual property we built (our desktop, mobile, video conferencing, chat and custom scaling fixes) for the hundreds of commercially licensed customers of 2600hz that have wanted to tap into the tens of millions of dollars we've invested over the years as the largest operator of the platform. That being said, I'd venture to guess it'll probably be in the next 4-8 weeks as engineering has been slowly chipping away at it behind the scenes over the last 2 years and they are in a state where they are generally ready. It's not going to come with much documentation. There are significant improvements in the 5.X family compared to 4.3 although some of those for scaling are tied to closed source modules that have long existed. 

It was a bit complicated when we acquired 2600hz as they had pivoted their strategy to still having open source but tying it to the new marketplace and there was still a lot of work to be done. They were pretty focused on their commercially licensed customers. That being said, we recognize the value of open source as that is where Ooma started years ago with the platform and then we wanted 2600hz to continue to survive and make money so we entered into a relationship where we could compensate them in a meaningful way to help fund all the developers (the majority of the company was and is R&D resources). Hope that helps!  

This wasn't purposeful. That site was tied to a credit card we didn't know about in the acquisition, it expired and the hosting provider deleted everything. So we had no way to recover it. We would've gladly left it up (like prior KazooCon videos on youtube). Do see my prior post on open source though. We also can't use Kazoo anymore as the trademark lapsed years ago (another neat thing we wound up discovering) and another company snagged it and wouldn't sell it to Ooma. 

@Chris Burgy 

I really appreciate the transparent explanation—it helps clarify both the upcoming OS 5.4 release and the broader direction since the pivot. It’s good to hear the open-source version is still being carried forward alongside the commercial efforts. Once the release is out, I’d be interested in learning more about how the scaling improvements and repo split might impact operators on the open-source side. Please keep me in the loop if there are any early notes or best practices you’d recommend.

One quick question: will the marketplace remain exclusively available for commercial customers, or will open-source operators also be able to purchase licenses for specific modules?

[Chris Burgy]

5 minutes ago, RuhNet said:

@Chris Burgy Absolutely, would be nice to keep the name, but I doubt anyone really cares. 

Minimal or no documentation is fine; we can figure it out from the source code. :-)

As far as some [many] modules not being part of the open source release, that was expected, but can you tell us if the release will be complete as far as having all components of a working system, [albeit possibly barebones]? In other words, are there any critical parts that will be missing for a standard PBX that we (the open source community) will need to add?

Also, is this release going to be a "one time dump and then move on" or will the open source public repos be the same codebase used internally for your developers' continuing work on the project (like Kazoo was prior to v4.3)?

Well the intent is to keep it as it was with 4.3. So the core of the platform remains open source as it was before. Some things are changing though. An example is we are porting over a management interface we built for our Ooma Enterprise platform (Voxter acquisition which was built on OS) but that'll be closed source for commercial customers since we spent a lot of development time building it. So y'all can do whatever you want with Monster UI (we will deprecate that eventually for commercial customers and we've rebuilt a good portion of that into a more modern UI). The foundational code should be the same for the repo we manage. We've largely gotten to a point where most of the development effort is around ancillary closed-source applications. 5.4, for our commercial customers, is a LTS release and Ooma is going to move there (we are on 5.3 currently) later this fall/winter. I can imagine in the future we do release further updates but they are likely to be minimal (things like certain FS, Kamilio, CouchDB, etc. updates) that constitute a dot release for us. I hope that helps. It's a great foundational platform to build off and for FWIW, we solely interact with the platform via all the APIs. We built all of our own IP all around that foundation with the use of the closed-source modules to help with scaling the platform and getting access to some of the new features (e.g. Qubicle - which is the refactored ACDC that was closed source before we acquired 2600hz). 

1 minute ago, synogeek said:

@Chris Burgy 

I really appreciate the transparent explanation—it helps clarify both the upcoming OS 5.4 release and the broader direction since the pivot. It’s good to hear the open-source version is still being carried forward alongside the commercial efforts. Once the release is out, I’d be interested in learning more about how the scaling improvements and repo split might impact operators on the open-source side. Please keep me in the loop if there are any early notes or best practices you’d recommend.

One quick question: will the marketplace remain exclusively available for commercial customers, or will open-source operators also be able to purchase licenses for specific modules?

Well the repo split was largely the same as what it was before we acquired 2600hz. The big difference is we brought over our tech that is additive. There should be no effective difference for y'all in terms of scaling. There are inherent limits (you need to get to a reasonable size) in the platform without ecallmgr but I do know of a company that decided to hire a bunch of Erlang engineers and scaled it to 300k users. It's all solvable if you want to go invest in the development talent. 

 

The marketplace was originally developed by 2600hz for two purposes: facilitating the loading of applications into commercial customers (specifically those on private cloud and global infrastructure) and also for 2600hz to attempt to monetize the open source community with add-ons. To be honest, we haven't actually spent much time on the use case that 2600hz developed to monetize the open-source community. The economics on the commercial model are pretty attractive so it'd just make more sense for customers to pivot to that model.

 

 

[synogeek]

19 minutes ago, Chris Burgy said:

Well the intent is to keep it as it was with 4.3. So the core of the platform remains open source as it was before. Some things are changing though. An example is we are porting over a management interface we built for our Ooma Enterprise platform (Voxter acquisition which was built on OS) but that'll be closed source for commercial customers since we spent a lot of development time building it. So y'all can do whatever you want with Monster UI (we will deprecate that eventually for commercial customers and we've rebuilt a good portion of that into a more modern UI). The foundational code should be the same for the repo we manage. We've largely gotten to a point where most of the development effort is around ancillary closed-source applications. 5.4, for our commercial customers, is a LTS release and Ooma is going to move there (we are on 5.3 currently) later this fall/winter. I can imagine in the future we do release further updates but they are likely to be minimal (things like certain FS, Kamilio, CouchDB, etc. updates) that constitute a dot release for us. I hope that helps. It's a great foundational platform to build off and for FWIW, we solely interact with the platform via all the APIs. We built all of our own IP all around that foundation with the use of the closed-source modules to help with scaling the platform and getting access to some of the new features (e.g. Qubicle - which is the refactored ACDC that was closed source before we acquired 2600hz). 

Well the repo split was largely the same as what it was before we acquired 2600hz. The big difference is we brought over our tech that is additive. There should be no effective difference for y'all in terms of scaling. There are inherent limits (you need to get to a reasonable size) in the platform without ecallmgr but I do know of a company that decided to hire a bunch of Erlang engineers and scaled it to 300k users. It's all solvable if you want to go invest in the development talent. 

 

The marketplace was originally developed by 2600hz for two purposes: facilitating the loading of applications into commercial customers (specifically those on private cloud and global infrastructure) and also for 2600hz to attempt to monetize the open source community with add-ons. To be honest, we haven't actually spent much time on the use case that 2600hz developed to monetize the open-source community. The economics on the commercial model are pretty attractive so it'd just make more sense for customers to pivot to that model.

 

 

Thank you for clarifying—that makes a lot of sense. I completely understand why the economics of the commercial model are more attractive and why it’s become the primary focus. That said, I think there’s still meaningful value in keeping a clear path for the open-source community to access licensed add-ons. For many operators, the ability to selectively license modules could serve as a stepping stone toward eventually adopting the full commercial model, rather than creating an either/or decision.

From that perspective, enabling open-source operators to participate in the marketplace—even in a limited, structured way—could strengthen the ecosystem overall, foster innovation from community contributors, and ultimately expand the funnel of future commercial customers. It seems like a win-win: Ooma continues to monetize its IP, and smaller operators gain a pathway to grow with the platform without needing to fully pivot all at once.

I’ll definitely keep an eye on 5.4’s release and appreciate your transparency throughout this conversation. Please do let me know if there are any upcoming discussions or initiatives around open-source participation in the marketplace—I’d be very interested in staying engaged.

[charles_3003]

Very interesting and good news. 

[Chris Burgy]

4 minutes ago, synogeek said:

Thank you for clarifying—that makes a lot of sense. I completely understand why the economics of the commercial model are more attractive and why it’s become the primary focus. That said, I think there’s still meaningful value in keeping a clear path for the open-source community to access licensed add-ons. For many operators, the ability to selectively license modules could serve as a stepping stone toward eventually adopting the full commercial model, rather than creating an either/or decision.

From that perspective, enabling open-source operators to participate in the marketplace—even in a limited, structured way—could strengthen the ecosystem overall, foster innovation from community contributors, and ultimately expand the funnel of future commercial customers. It seems like a win-win: Ooma continues to monetize its IP, and smaller operators gain a pathway to grow with the platform without needing to fully pivot all at once.

I’ll definitely keep an eye on 5.4’s release and appreciate your transparency throughout this conversation. Please do let me know if there are any upcoming discussions or initiatives around open-source participation in the marketplace—I’d be very interested in staying engaged.

Sure. Honestly, the open source community is a bit of an unknown black box to us since there is no way to really identify who is using it and we've haphazardly encountered folks who wound up reaching out to 2600hz because they needed scaling or some other situation they couldn't solve on their own and reached sufficient size where they wanted the closed-source components. We'd be happy to consider a case where certain modules are available to that community (and what was built for the marketplace can serve as a mechanism for that). Ultimately, like y'all, we have to support a business and investors as public company. Good news is, there were other interested parties bidding on 2600hz at the same time we were forced into this, and we have a favorable approach to all of this from our background with the company. The other parties may have well just shut it all down because they just wanted the tech for themselves. 

 

@marketingChris K -I think we might have an upcoming OS call but if we don't it'd be useful to get that back on the regular quarterly docket to discuss more of this with the OS community and what they are interested in from 2600hz and helping to educate on the closed source module (including the stuff Ooma has ported over). 

[Mooseable]

Well, if you want to demystify your OS community, feel free to reach out. We've got a strong one going with the kazoo-classic efforts as you've likely seen.

Since the way Apps work, hooking into an AMQP bus, I always saw it a missed opportunity to not just run apps on your own nodes (like CallCentre apps) and tap it into the RabbitMQ queue of self-hosted open-source customers. I happily would have paid for apps and on-charge those features to our customers.

Also, doesn't 4.3 have telemetry on by default? Or do you not collect/analyse that? (genuinely curious)

[Chris Burgy]

5 hours ago, Mooseable said:

Well, if you want to demystify your OS community, feel free to reach out. We've got a strong one going with the kazoo-classic efforts as you've likely seen.

Since the way Apps work, hooking into an AMQP bus, I always saw it a missed opportunity to not just run apps on your own nodes (like CallCentre apps) and tap it into the RabbitMQ queue of self-hosted open-source customers. I happily would have paid for apps and on-charge those features to our customers.

Also, doesn't 4.3 have telemetry on by default? Or do you not collect/analyse that? (genuinely curious)

No, there is no telemetry in the platform. That would've been nice. Your point on the apps was part of the idea 2600hz had for the marketplace to deliver apps to OS community. Although that concept does get a lot more complicated as open source users don't have access to some of the closed sourced components that are required for certain apps/scaling and also the code could theoretically deviate from baseline (you obviously can modify it as you see fit) which could cause problems. Plus how you architect your hosting infrastructure can impact all of this too. 

[synogeek]

1 hour ago, Chris Burgy said:

No, there is no telemetry in the platform. That would've been nice. Your point on the apps was part of the idea 2600hz had for the marketplace to deliver apps to OS community. Although that concept does get a lot more complicated as open source users don't have access to some of the closed sourced components that are required for certain apps/scaling and also the code could theoretically deviate from baseline (you obviously can modify it as you see fit) which could cause problems. Plus how you architect your hosting infrastructure can impact all of this too. 

I hear you on the challenges with the marketplace and compatibility, especially since open-source deployments can diverge from the baseline and lack access to certain closed components. That’s always going to make support and predictability more complicated.

One thought I wanted to raise: has there been any consideration of opening the marketplace to allow the OS community to develop and sell apps, with Ooma/2600hz taking a percentage of those sales? It could work similarly to what Apple, Google, and Microsoft have done with their app ecosystems. This way, the community helps grow the platform with innovation at the edge, while Ooma still benefits directly from marketplace revenue and keeps the ecosystem healthy.

It seems like that could create a win-win scenario: commercial customers continue to get the full value of the licensed modules, while the OS community has a structured way to contribute and innovate, with Ooma getting a share of the upside for providing the marketplace infrastructure.

[RuhNet]

1 hour ago, Chris Burgy said:

No, there is no telemetry in the platform. That would've been nice.

In v4.3 there is a telemetry module (kazoo_telemetry) which is enabled by default as part of Kazoo core, but maybe y'all haven't been collecting from it. It grabs statistics (which can be anonymized with a setting) on Kazoo services, Kazoo cluster nodes, and CouchDB cluster nodes, and sends data to https://telemetry.2600hz.org/api

Perhaps you could re-enable that endpoint and see what you get. Might be interesting. :-)