Riak Bug Tracker

classic Classic list List threaded Threaded
9 messages Options
Reply | Threaded
Open this post in threaded view
|

Riak Bug Tracker

Stephen Day-2
Hey all:

First of all, I'd like to pat you on the back for a job well done with Riak. I tried it out on a small multi-node cluster and was very impressed by the ease of install, sensible defaults and sensible configuration. However, there are a couple of questions I have in regard to this project:

1. Is there a publicly available bugtracker? This feature seems to be disabled on your bitbucket repository. Generally, I will use the filed bugs and resolutions as a way to gauge the quality of the project (its more about response and resolution quality than quantity). It also shows a degree of transparency from the maintainers. Also, if I run into any issues when evaluating, I will gladly submit a detailed bug report, and maybe even a patch.

2. Is there an HTTP interface for a bulk commit? CouchDB has this and it makes adding a large number of documents for an initial migration very simple. Riak seems to be "fast enough" for reads, but it seems to be quite slow for writes. I would expect much of this is from HTTP overhead on doc per request stores. (Apologies if this is in the documentation and I missed it ;).

3. Why are features that come in other projects as standard (web monitoring, site-to-site replication) bundled in an EnterpriseDB edition? This is a huge risk when considering Riak for adoption, especially when its not clear what future features might not make it into the open source version, especially if a larger company purchases Basho. For instance, if there is a great community patch for a monitoring interface or site-to-site replication or any other feature that you want to monetize, it seems doubtful that you would consider accepting it. I understand that you need to make your money, but I have to say, this model raises concerns, even if your intentions are benevolent. That said, hypothetically, I would be more interested in support, if it could be considered without having to weigh it against arbitrary bundling decisions.

Thanks for your time and hope for you all to keep up the good work.

Steve

_______________________________________________
riak-users mailing list
[hidden email]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
Reply | Threaded
Open this post in threaded view
|

Re: Riak Bug Tracker

Preston Marshall

On Mar 12, 2010, at 3:33 PM, Stephen Day wrote:

Hey all:

First of all, I'd like to pat you on the back for a job well done with Riak. I tried it out on a small multi-node cluster and was very impressed by the ease of install, sensible defaults and sensible configuration. However, there are a couple of questions I have in regard to this project:

1. Is there a publicly available bugtracker? This feature seems to be disabled on your bitbucket repository. Generally, I will use the filed bugs and resolutions as a way to gauge the quality of the project (its more about response and resolution quality than quantity). It also shows a degree of transparency from the maintainers. Also, if I run into any issues when evaluating, I will gladly submit a detailed bug report, and maybe even a patch.
http://issues.basho.com/

2. Is there an HTTP interface for a bulk commit? CouchDB has this and it makes adding a large number of documents for an initial migration very simple. Riak seems to be "fast enough" for reads, but it seems to be quite slow for writes. I would expect much of this is from HTTP overhead on doc per request stores. (Apologies if this is in the documentation and I missed it ;).
I have seen some other questions about this, not sure if it exists.  If your writes are slow, make sure you are using InnoDB.

3. Why are features that come in other projects as standard (web monitoring, site-to-site replication) bundled in an EnterpriseDB edition? This is a huge risk when considering Riak for adoption, especially when its not clear what future features might not make it into the open source version, especially if a larger company purchases Basho. For instance, if there is a great community patch for a monitoring interface or site-to-site replication or any other feature that you want to monetize, it seems doubtful that you would consider accepting it. I understand that you need to make your money, but I have to say, this model raises concerns, even if your intentions are benevolent. That said, hypothetically, I would be more interested in support, if it could be considered without having to weigh it against arbitrary bundling decisions.

Thanks for your time and hope for you all to keep up the good work.

Steve
_______________________________________________
riak-users mailing list
[hidden email]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com


_______________________________________________
riak-users mailing list
[hidden email]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com

smime.p7s (6K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Riak Bug Tracker

Sean Cribbs-2
In reply to this post by Stephen Day-2

On Mar 12, 2010, at 4:33 PM, Stephen Day wrote:


1. Is there a publicly available bugtracker? This feature seems to be disabled on your bitbucket repository. Generally, I will use the filed bugs and resolutions as a way to gauge the quality of the project (its more about response and resolution quality than quantity). It also shows a degree of transparency from the maintainers. Also, if I run into any issues when evaluating, I will gladly submit a detailed bug report, and maybe even a patch.


We have a bugzilla at http://issues.basho.com. Feel free to submit bug reports there.  For patches, send a pull request on BitBucket.

2. Is there an HTTP interface for a bulk commit? CouchDB has this and it makes adding a large number of documents for an initial migration very simple. Riak seems to be "fast enough" for reads, but it seems to be quite slow for writes. I would expect much of this is from HTTP overhead on doc per request stores. (Apologies if this is in the documentation and I missed it ;).


There's no bulk-commit, but the HTTP interface supports Keep-Alive, so you should be able to submit multiple requests over the same HTTP connection.

3. Why are features that come in other projects as standard (web monitoring, site-to-site replication) bundled in an EnterpriseDB edition? This is a huge risk when considering Riak for adoption, especially when its not clear what future features might not make it into the open source version, especially if a larger company purchases Basho. For instance, if there is a great community patch for a monitoring interface or site-to-site replication or any other feature that you want to monetize, it seems doubtful that you would consider accepting it. I understand that you need to make your money, but I have to say, this model raises concerns, even if your intentions are benevolent. That said, hypothetically, I would be more interested in support, if it could be considered without having to weigh it against arbitrary bundling decisions.


You can start your development with the open-source version of Riak, then when you're ready to  add those features, talk to Basho about upgrading to EDS.  I believe you'll find Riak Open-Source to be very capable even without those features, and that when you're ready for the features, you'll also want the direct relationship with the Basho team that EDS provides.  In addition to the feature bullet-point stuff, you get input into future features, prerelease code and world-class support.

The dual-licensing model that Riak uses is not new, especially in the database world -- Sleepycat did it years ago.  The devs are also committed to releasing functional open-source products, not crippled ones.  Riak Open-Source is by no means crippled, EDS just has some extra sauce.

Sean Cribbs <[hidden email]>
Developer Advocate
Basho Technologies, Inc.


_______________________________________________
riak-users mailing list
[hidden email]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
Reply | Threaded
Open this post in threaded view
|

Re: Riak Bug Tracker

Stephen Day-2
Thanks for the quick response.

On Fri, Mar 12, 2010 at 1:56 PM, Sean Cribbs <[hidden email]> wrote:
We have a bugzilla at http://issues.basho.com. Feel free to submit bug reports there.  For patches, send a pull request on BitBucket.
Thanks. This was hard to find.

There's no bulk-commit, but the HTTP interface supports Keep-Alive, so you should be able to submit multiple requests over the same HTTP connection.
I will try this.
 

3. Why are features that come in other projects as standard (web monitoring, site-to-site replication) bundled in an EnterpriseDB edition? This is a huge risk when considering Riak for adoption, especially when its not clear what future features might not make it into the open source version, especially if a larger company purchases Basho. For instance, if there is a great community patch for a monitoring interface or site-to-site replication or any other feature that you want to monetize, it seems doubtful that you would consider accepting it. I understand that you need to make your money, but I have to say, this model raises concerns, even if your intentions are benevolent. That said, hypothetically, I would be more interested in support, if it could be considered without having to weigh it against arbitrary bundling decisions.


You can start your development with the open-source version of Riak, then when you're ready to  add those features, talk to Basho about upgrading to EDS.  I believe you'll find Riak Open-Source to be very capable even without those features, and that when you're ready for the features, you'll also want the direct relationship with the Basho team that EDS provides.  In addition to the feature bullet-point stuff, you get input into future features, prerelease code and world-class support.
Agreed, but other than support, this all comes with open source projects. I can get prerelease code from any of them; I can send an email to their list saying, "I want this project to Foo", and it gets taken under advisement. It just seems like an arbitrary line was drawn in the sand for what is called "enterprise".


The dual-licensing model that Riak uses is not new, especially in the database world -- Sleepycat did it years ago.  The devs are also committed to releasing functional open-source products, not crippled ones. 
It's interesting that you mention Sleepycat as a point in favor of a dual-license model, because I was considering mentioning them and MySQL as examples that this can can be risky for users. While these were successful businesses, both companies were purchased by Oracle and now their flagship projects sit in stagnation (debatable, I understand, but neither project is where PostgreSQL is now). Both of these companies were committed and did release very functional open source options. Now, decisions about these products are made solely based on the bottom line.
 
Riak Open-Source is by no means crippled, EDS just has some extra sauce.
This is clear. I wouldn't be writing you if I thought it was crippled. The concerns come from the future of the extra sauce, especially if the sauce spoils the meat. I guess the more accurate question would be, how are you going to ensure that the open source version doesn't become crippled when you start seeing dollar signs? What is the feature roadmap for Open-Source and EnterpriseDB versions?


Sean Cribbs <[hidden email]>
Developer Advocate
Basho Technologies, Inc.



_______________________________________________
riak-users mailing list
[hidden email]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
Reply | Threaded
Open this post in threaded view
|

Re: Riak Bug Tracker

Antony Falco
Stephen --

I think we understand your concerns and really appreciate your involvement with open source Riak.   

Just to clarify, Riak EnterpriseDS is a blend of some features and advanced support with SLAs, not just a bunch of valuable features "arbitrarily" held back.  While this is still technically an open-core licensing model (OCLM) it differs from most OCLM offerings in that it allows us to push many more features _and_ support hours into the open source community

 It also underscores an oversimplification some make that holding back features is categorically BAD but holding back support is categorically GOOD.  Both are business models defined by what value they hold back and for which they charge users of software.  (I am intentionally ignoring NoSQL-as-a-service companies here because the differences are not germaine to the point you raised.)

In my experience,  support-only companies would never be as aggressive about answering questions on the riak-users list  as we are -- note how fast Rusty and Sean responded today to questions (on a Sunday no less).    Look at our response times in general to questions on Riak-users -- mine aside, of course.  Our calculated decision to productize a blend of some features and some support allows us to offer more of both to the community for free.  More companies should do this.

There is an economic component to our choice as well.  Support-only companies scale much differently than our approach with OCLM and require a LOT more capital, usually from those stalwarts of benevolence, venture capitalists.    

Support-only models make money not off subscription-based licenses but off of support.   Support staff who focus solely on customers soon outnumber developers,.  Support companies give away their software for free then look for ways to "monetize the community."  We are under no such pressure.   We don't market to our open source users.   We don't have to.   We are not under pressure from outside investors because of our choices.

At Basho we are 19 partners (we believe the idea of "founder" in open source software companies is silly and corrosive to the community ),  of which 15 of us are developers.  We are self-funded mainly, though we have had critical help from some angels interested in open source software and scaling infrastructure.  The other projects in NoSQL are either venture-funded or subsumed under large public corporations or licensed under AGPL (or all three!).  Does anyone feel safer with a company 51+% owned by venture or who answers to Wall Street?

Contrary to your fears about us setting some sort of enterprise trap "even if [our] intentions were benevolent,"  Basho is getting more open with every release.  

We started building Riak two years ago and open sourced it in August.  Since then, the open source code base has grown at many times the rate of the enterpriseDS code base.  We chose the apache license for its open nature.  We have plenty of companies (like EA) who use Riak open source.  In addition, we run the EnterpriseDS for Startups program -- startups pay us whatever they can for features  _and_ Advanced Support.  

Companies that disingenuously enter into open-source licensing arrangements and engage a community of users will fail.   The users will figure that out soon enough.   I respect your concerns and believe that part of participation in a community is not just respecting opposing ideas but encouraging them.   I ask only that you show us the reciprocal respect of commenting on our actual actions, not asking us to defend ourselves against some conjectural future evil doings. 

Please keep contributing ideas and questions and otherwise actively engaging with Riak.   It is perhaps a quaint notion nowadays but we believe dissenting views strengthen communities.  

Tony Falco
Basho Technologies
m:  617-417-4482

On Mar 12, 2010, at 3:24 PM, Stephen Day wrote:

Thanks for the quick response.

On Fri, Mar 12, 2010 at 1:56 PM, Sean Cribbs <[hidden email]> wrote:
We have a bugzilla at http://issues.basho.com. Feel free to submit bug reports there.  For patches, send a pull request on BitBucket.
Thanks. This was hard to find.

There's no bulk-commit, but the HTTP interface supports Keep-Alive, so you should be able to submit multiple requests over the same HTTP connection.
I will try this.
 

3. Why are features that come in other projects as standard (web monitoring, site-to-site replication) bundled in an EnterpriseDB edition? This is a huge risk when considering Riak for adoption, especially when its not clear what future features might not make it into the open source version, especially if a larger company purchases Basho. For instance, if there is a great community patch for a monitoring interface or site-to-site replication or any other feature that you want to monetize, it seems doubtful that you would consider accepting it. I understand that you need to make your money, but I have to say, this model raises concerns, even if your intentions are benevolent. That said, hypothetically, I would be more interested in support, if it could be considered without having to weigh it against arbitrary bundling decisions.


You can start your development with the open-source version of Riak, then when you're ready to  add those features, talk to Basho about upgrading to EDS.  I believe you'll find Riak Open-Source to be very capable even without those features, and that when you're ready for the features, you'll also want the direct relationship with the Basho team that EDS provides.  In addition to the feature bullet-point stuff, you get input into future features, prerelease code and world-class support.
Agreed, but other than support, this all comes with open source projects. I can get prerelease code from any of them; I can send an email to their list saying, "I want this project to Foo", and it gets taken under advisement. It just seems like an arbitrary line was drawn in the sand for what is called "enterprise".


The dual-licensing model that Riak uses is not new, especially in the database world -- Sleepycat did it years ago.  The devs are also committed to releasing functional open-source products, not crippled ones. 
It's interesting that you mention Sleepycat as a point in favor of a dual-license model, because I was considering mentioning them and MySQL as examples that this can can be risky for users. While these were successful businesses, both companies were purchased by Oracle and now their flagship projects sit in stagnation (debatable, I understand, but neither project is where PostgreSQL is now). Both of these companies were committed and did release very functional open source options. Now, decisions about these products are made solely based on the bottom line.
 
Riak Open-Source is by no means crippled, EDS just has some extra sauce.
This is clear. I wouldn't be writing you if I thought it was crippled. The concerns come from the future of the extra sauce, especially if the sauce spoils the meat. I guess the more accurate question would be, how are you going to ensure that the open source version doesn't become crippled when you start seeing dollar signs? What is the feature roadmap for Open-Source and EnterpriseDB versions?


Sean Cribbs <[hidden email]>
Developer Advocate
Basho Technologies, Inc.


_______________________________________________
riak-users mailing list
[hidden email]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com


_______________________________________________
riak-users mailing list
[hidden email]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
Reply | Threaded
Open this post in threaded view
|

Re: Riak Bug Tracker

Alexander Sicular
Thanks for the clarification. I, for one, appreciate the transparency. A different approach that hopefully works out for Basho. After all, developers gotta eat. -Alexander

On Mar 14, 2010, at 8:31 PM, Antony Falco wrote:

> Stephen --
>
> I think we understand your concerns and really appreciate your involvement with open source Riak.  
>
> Just to clarify, Riak EnterpriseDS is a blend of some features and advanced support with SLAs, not just a bunch of valuable features "arbitrarily" held back.  While this is still technically an open-core licensing model (OCLM) it differs from most OCLM offerings in that it allows us to push many more features _and_ support hours into the open source community
>
>  It also underscores an oversimplification some make that holding back features is categorically BAD but holding back support is categorically GOOD.  Both are business models defined by what value they hold back and for which they charge users of software.  (I am intentionally ignoring NoSQL-as-a-service companies here because the differences are not germaine to the point you raised.)
>
> In my experience,  support-only companies would never be as aggressive about answering questions on the riak-users list  as we are -- note how fast Rusty and Sean responded today to questions (on a Sunday no less).    Look at our response times in general to questions on Riak-users -- mine aside, of course.  Our calculated decision to productize a blend of some features and some support allows us to offer more of both to the community for free.  More companies should do this.
>
> There is an economic component to our choice as well.  Support-only companies scale much differently than our approach with OCLM and require a LOT more capital, usually from those stalwarts of benevolence, venture capitalists.    
>
> Support-only models make money not off subscription-based licenses but off of support.   Support staff who focus solely on customers soon outnumber developers,.  Support companies give away their software for free then look for ways to "monetize the community."  We are under no such pressure.   We don't market to our open source users.   We don't have to.   We are not under pressure from outside investors because of our choices.
>
> At Basho we are 19 partners (we believe the idea of "founder" in open source software companies is silly and corrosive to the community ),  of which 15 of us are developers.  We are self-funded mainly, though we have had critical help from some angels interested in open source software and scaling infrastructure.  The other projects in NoSQL are either venture-funded or subsumed under large public corporations or licensed under AGPL (or all three!).  Does anyone feel safer with a company 51+% owned by venture or who answers to Wall Street?
>
> Contrary to your fears about us setting some sort of enterprise trap "even if [our] intentions were benevolent,"  Basho is getting more open with every release.  
>
> We started building Riak two years ago and open sourced it in August.  Since then, the open source code base has grown at many times the rate of the enterpriseDS code base.  We chose the apache license for its open nature.  We have plenty of companies (like EA) who use Riak open source.  In addition, we run the EnterpriseDS for Startups program -- startups pay us whatever they can for features  _and_ Advanced Support.  
>
> Companies that disingenuously enter into open-source licensing arrangements and engage a community of users will fail.   The users will figure that out soon enough.   I respect your concerns and believe that part of participation in a community is not just respecting opposing ideas but encouraging them.   I ask only that you show us the reciprocal respect of commenting on our actual actions, not asking us to defend ourselves against some conjectural future evil doings.
>
> Please keep contributing ideas and questions and otherwise actively engaging with Riak.   It is perhaps a quaint notion nowadays but we believe dissenting views strengthen communities.  
>
> Tony Falco
> Basho Technologies
> m:  617-417-4482
>
> On Mar 12, 2010, at 3:24 PM, Stephen Day wrote:
>
>> Thanks for the quick response.
>>
>> On Fri, Mar 12, 2010 at 1:56 PM, Sean Cribbs <[hidden email]> wrote:
>> We have a bugzilla at http://issues.basho.com. Feel free to submit bug reports there.  For patches, send a pull request on BitBucket.
>> Thanks. This was hard to find.
>>
>> There's no bulk-commit, but the HTTP interface supports Keep-Alive, so you should be able to submit multiple requests over the same HTTP connection.
>> I will try this.
>>  
>>
>>> 3. Why are features that come in other projects as standard (web monitoring, site-to-site replication) bundled in an EnterpriseDB edition? This is a huge risk when considering Riak for adoption, especially when its not clear what future features might not make it into the open source version, especially if a larger company purchases Basho. For instance, if there is a great community patch for a monitoring interface or site-to-site replication or any other feature that you want to monetize, it seems doubtful that you would consider accepting it. I understand that you need to make your money, but I have to say, this model raises concerns, even if your intentions are benevolent. That said, hypothetically, I would be more interested in support, if it could be considered without having to weigh it against arbitrary bundling decisions.
>>>
>>
>> You can start your development with the open-source version of Riak, then when you're ready to  add those features, talk to Basho about upgrading to EDS.  I believe you'll find Riak Open-Source to be very capable even without those features, and that when you're ready for the features, you'll also want the direct relationship with the Basho team that EDS provides.  In addition to the feature bullet-point stuff, you get input into future features, prerelease code and world-class support.
>> Agreed, but other than support, this all comes with open source projects. I can get prerelease code from any of them; I can send an email to their list saying, "I want this project to Foo", and it gets taken under advisement. It just seems like an arbitrary line was drawn in the sand for what is called "enterprise".
>>
>>
>> The dual-licensing model that Riak uses is not new, especially in the database world -- Sleepycat did it years ago.  The devs are also committed to releasing functional open-source products, not crippled ones.  
>> It's interesting that you mention Sleepycat as a point in favor of a dual-license model, because I was considering mentioning them and MySQL as examples that this can can be risky for users. While these were successful businesses, both companies were purchased by Oracle and now their flagship projects sit in stagnation (debatable, I understand, but neither project is where PostgreSQL is now). Both of these companies were committed and did release very functional open source options. Now, decisions about these products are made solely based on the bottom line.
>>  
>> Riak Open-Source is by no means crippled, EDS just has some extra sauce.
>> This is clear. I wouldn't be writing you if I thought it was crippled. The concerns come from the future of the extra sauce, especially if the sauce spoils the meat. I guess the more accurate question would be, how are you going to ensure that the open source version doesn't become crippled when you start seeing dollar signs? What is the feature roadmap for Open-Source and EnterpriseDB versions?
>>
>>
>> Sean Cribbs <[hidden email]>
>> Developer Advocate
>> Basho Technologies, Inc.
>> http://basho.com/
>>
>>
>> _______________________________________________
>> riak-users mailing list
>> [hidden email]
>> http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
>
> _______________________________________________
> riak-users mailing list
> [hidden email]
> http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com


_______________________________________________
riak-users mailing list
[hidden email]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
Reply | Threaded
Open this post in threaded view
|

Re: Riak Bug Tracker

Stephen Day-2
In reply to this post by Antony Falco
Tony:

Thank you for the detailed response.

As a company, I believe that Basho is a shining example of a dual-license open-source business model. Your support for the open-source community is definitely not in question. You obviously recognize the potential that community involvement has to improve the end product.

I'm sorry if I came off a little bit trollish and relegated my questions to conjecture rather than specifics, but this is the question I am interested in (though it may have come off as a rhetorical device):

Is there a feature roadmap for the open source vs enterprise versions? If so, what is the road map for the open source version?

As far as "future evils" go (I honestly feel bad that you read it that way ;) ), I am not worried in the least bit about riak or Basho. Again, thank you for your time and keep up the good work.

steve

On Sun, Mar 14, 2010 at 5:31 PM, Antony Falco <[hidden email]> wrote:
Stephen --

I think we understand your concerns and really appreciate your involvement with open source Riak.   

Just to clarify, Riak EnterpriseDS is a blend of some features and advanced support with SLAs, not just a bunch of valuable features "arbitrarily" held back.  While this is still technically an open-core licensing model (OCLM) it differs from most OCLM offerings in that it allows us to push many more features _and_ support hours into the open source community

 It also underscores an oversimplification some make that holding back features is categorically BAD but holding back support is categorically GOOD.  Both are business models defined by what value they hold back and for which they charge users of software.  (I am intentionally ignoring NoSQL-as-a-service companies here because the differences are not germaine to the point you raised.)

In my experience,  support-only companies would never be as aggressive about answering questions on the riak-users list  as we are -- note how fast Rusty and Sean responded today to questions (on a Sunday no less).    Look at our response times in general to questions on Riak-users -- mine aside, of course.  Our calculated decision to productize a blend of some features and some support allows us to offer more of both to the community for free.  More companies should do this.

There is an economic component to our choice as well.  Support-only companies scale much differently than our approach with OCLM and require a LOT more capital, usually from those stalwarts of benevolence, venture capitalists.    

Support-only models make money not off subscription-based licenses but off of support.   Support staff who focus solely on customers soon outnumber developers,.  Support companies give away their software for free then look for ways to "monetize the community."  We are under no such pressure.   We don't market to our open source users.   We don't have to.   We are not under pressure from outside investors because of our choices.

At Basho we are 19 partners (we believe the idea of "founder" in open source software companies is silly and corrosive to the community ),  of which 15 of us are developers.  We are self-funded mainly, though we have had critical help from some angels interested in open source software and scaling infrastructure.  The other projects in NoSQL are either venture-funded or subsumed under large public corporations or licensed under AGPL (or all three!).  Does anyone feel safer with a company 51+% owned by venture or who answers to Wall Street?

Contrary to your fears about us setting some sort of enterprise trap "even if [our] intentions were benevolent,"  Basho is getting more open with every release.  

We started building Riak two years ago and open sourced it in August.  Since then, the open source code base has grown at many times the rate of the enterpriseDS code base.  We chose the apache license for its open nature.  We have plenty of companies (like EA) who use Riak open source.  In addition, we run the EnterpriseDS for Startups program -- startups pay us whatever they can for features  _and_ Advanced Support.  

Companies that disingenuously enter into open-source licensing arrangements and engage a community of users will fail.   The users will figure that out soon enough.   I respect your concerns and believe that part of participation in a community is not just respecting opposing ideas but encouraging them.   I ask only that you show us the reciprocal respect of commenting on our actual actions, not asking us to defend ourselves against some conjectural future evil doings. 

Please keep contributing ideas and questions and otherwise actively engaging with Riak.   It is perhaps a quaint notion nowadays but we believe dissenting views strengthen communities.  

Tony Falco
Basho Technologies
m:  617-417-4482

On Mar 12, 2010, at 3:24 PM, Stephen Day wrote:

Thanks for the quick response.

On Fri, Mar 12, 2010 at 1:56 PM, Sean Cribbs <[hidden email]> wrote:
We have a bugzilla at http://issues.basho.com. Feel free to submit bug reports there.  For patches, send a pull request on BitBucket.
Thanks. This was hard to find.

There's no bulk-commit, but the HTTP interface supports Keep-Alive, so you should be able to submit multiple requests over the same HTTP connection.
I will try this.
 

3. Why are features that come in other projects as standard (web monitoring, site-to-site replication) bundled in an EnterpriseDB edition? This is a huge risk when considering Riak for adoption, especially when its not clear what future features might not make it into the open source version, especially if a larger company purchases Basho. For instance, if there is a great community patch for a monitoring interface or site-to-site replication or any other feature that you want to monetize, it seems doubtful that you would consider accepting it. I understand that you need to make your money, but I have to say, this model raises concerns, even if your intentions are benevolent. That said, hypothetically, I would be more interested in support, if it could be considered without having to weigh it against arbitrary bundling decisions.


You can start your development with the open-source version of Riak, then when you're ready to  add those features, talk to Basho about upgrading to EDS.  I believe you'll find Riak Open-Source to be very capable even without those features, and that when you're ready for the features, you'll also want the direct relationship with the Basho team that EDS provides.  In addition to the feature bullet-point stuff, you get input into future features, prerelease code and world-class support.
Agreed, but other than support, this all comes with open source projects. I can get prerelease code from any of them; I can send an email to their list saying, "I want this project to Foo", and it gets taken under advisement. It just seems like an arbitrary line was drawn in the sand for what is called "enterprise".


The dual-licensing model that Riak uses is not new, especially in the database world -- Sleepycat did it years ago.  The devs are also committed to releasing functional open-source products, not crippled ones. 
It's interesting that you mention Sleepycat as a point in favor of a dual-license model, because I was considering mentioning them and MySQL as examples that this can can be risky for users. While these were successful businesses, both companies were purchased by Oracle and now their flagship projects sit in stagnation (debatable, I understand, but neither project is where PostgreSQL is now). Both of these companies were committed and did release very functional open source options. Now, decisions about these products are made solely based on the bottom line.
 
Riak Open-Source is by no means crippled, EDS just has some extra sauce.
This is clear. I wouldn't be writing you if I thought it was crippled. The concerns come from the future of the extra sauce, especially if the sauce spoils the meat. I guess the more accurate question would be, how are you going to ensure that the open source version doesn't become crippled when you start seeing dollar signs? What is the feature roadmap for Open-Source and EnterpriseDB versions?


Sean Cribbs <[hidden email]>
Developer Advocate
Basho Technologies, Inc.


_______________________________________________
riak-users mailing list
[hidden email]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com



_______________________________________________
riak-users mailing list
[hidden email]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
Reply | Threaded
Open this post in threaded view
|

Re: Riak Bug Tracker

Antony Falco
Stephen --

Thanks for your considered and considerate response.  And thanks for taking on Issues 75 and 76.

I did not find your comments troll-ish.  Your comments (even the "evil" ones) were informed by the way the debate about open source typically shapes up -- often with a lot more acrimony.   However, as this exchange shows, the debate is turning into a practical discussion as open-source business models mature.  I have to say my own response was reductive -- in the last several months I have met some very impressive and fair-minded people working in venture who get both infrastructure and the proper value of community and open source.  The real point is that some companies and investors view communities as markets from which to derive profit.

To answer your question:  Yes there is a roadmap of open source and enterprise features.  I think the best way to present it is as a discussion.  Do you want to do a podcast with us?

Take care,

Tony Falco
Basho Technologies
m:  617-417-4482

On Mar 15, 2010, at 6:56 PM, Stephen Day wrote:

Tony:

Thank you for the detailed response.

As a company, I believe that Basho is a shining example of a dual-license open-source business model. Your support for the open-source community is definitely not in question. You obviously recognize the potential that community involvement has to improve the end product.

I'm sorry if I came off a little bit trollish and relegated my questions to conjecture rather than specifics, but this is the question I am interested in (though it may have come off as a rhetorical device):

Is there a feature roadmap for the open source vs enterprise versions? If so, what is the road map for the open source version?

As far as "future evils" go (I honestly feel bad that you read it that way ;) ), I am not worried in the least bit about riak or Basho. Again, thank you for your time and keep up the good work.

steve

On Sun, Mar 14, 2010 at 5:31 PM, Antony Falco <[hidden email]> wrote:
Stephen --

I think we understand your concerns and really appreciate your involvement with open source Riak.   

Just to clarify, Riak EnterpriseDS is a blend of some features and advanced support with SLAs, not just a bunch of valuable features "arbitrarily" held back.  While this is still technically an open-core licensing model (OCLM) it differs from most OCLM offerings in that it allows us to push many more features _and_ support hours into the open source community

 It also underscores an oversimplification some make that holding back features is categorically BAD but holding back support is categorically GOOD.  Both are business models defined by what value they hold back and for which they charge users of software.  (I am intentionally ignoring NoSQL-as-a-service companies here because the differences are not germaine to the point you raised.)

In my experience,  support-only companies would never be as aggressive about answering questions on the riak-users list  as we are -- note how fast Rusty and Sean responded today to questions (on a Sunday no less).    Look at our response times in general to questions on Riak-users -- mine aside, of course.  Our calculated decision to productize a blend of some features and some support allows us to offer more of both to the community for free.  More companies should do this.

There is an economic component to our choice as well.  Support-only companies scale much differently than our approach with OCLM and require a LOT more capital, usually from those stalwarts of benevolence, venture capitalists.    

Support-only models make money not off subscription-based licenses but off of support.   Support staff who focus solely on customers soon outnumber developers,.  Support companies give away their software for free then look for ways to "monetize the community."  We are under no such pressure.   We don't market to our open source users.   We don't have to.   We are not under pressure from outside investors because of our choices.

At Basho we are 19 partners (we believe the idea of "founder" in open source software companies is silly and corrosive to the community ),  of which 15 of us are developers.  We are self-funded mainly, though we have had critical help from some angels interested in open source software and scaling infrastructure.  The other projects in NoSQL are either venture-funded or subsumed under large public corporations or licensed under AGPL (or all three!).  Does anyone feel safer with a company 51+% owned by venture or who answers to Wall Street?

Contrary to your fears about us setting some sort of enterprise trap "even if [our] intentions were benevolent,"  Basho is getting more open with every release.  

We started building Riak two years ago and open sourced it in August.  Since then, the open source code base has grown at many times the rate of the enterpriseDS code base.  We chose the apache license for its open nature.  We have plenty of companies (like EA) who use Riak open source.  In addition, we run the EnterpriseDS for Startups program -- startups pay us whatever they can for features  _and_ Advanced Support.  

Companies that disingenuously enter into open-source licensing arrangements and engage a community of users will fail.   The users will figure that out soon enough.   I respect your concerns and believe that part of participation in a community is not just respecting opposing ideas but encouraging them.   I ask only that you show us the reciprocal respect of commenting on our actual actions, not asking us to defend ourselves against some conjectural future evil doings. 

Please keep contributing ideas and questions and otherwise actively engaging with Riak.   It is perhaps a quaint notion nowadays but we believe dissenting views strengthen communities.  

Tony Falco
Basho Technologies
m:  617-417-4482

On Mar 12, 2010, at 3:24 PM, Stephen Day wrote:

Thanks for the quick response.

On Fri, Mar 12, 2010 at 1:56 PM, Sean Cribbs <[hidden email]> wrote:
We have a bugzilla at http://issues.basho.com. Feel free to submit bug reports there.  For patches, send a pull request on BitBucket.
Thanks. This was hard to find.

There's no bulk-commit, but the HTTP interface supports Keep-Alive, so you should be able to submit multiple requests over the same HTTP connection.
I will try this.
 

3. Why are features that come in other projects as standard (web monitoring, site-to-site replication) bundled in an EnterpriseDB edition? This is a huge risk when considering Riak for adoption, especially when its not clear what future features might not make it into the open source version, especially if a larger company purchases Basho. For instance, if there is a great community patch for a monitoring interface or site-to-site replication or any other feature that you want to monetize, it seems doubtful that you would consider accepting it. I understand that you need to make your money, but I have to say, this model raises concerns, even if your intentions are benevolent. That said, hypothetically, I would be more interested in support, if it could be considered without having to weigh it against arbitrary bundling decisions.


You can start your development with the open-source version of Riak, then when you're ready to  add those features, talk to Basho about upgrading to EDS.  I believe you'll find Riak Open-Source to be very capable even without those features, and that when you're ready for the features, you'll also want the direct relationship with the Basho team that EDS provides.  In addition to the feature bullet-point stuff, you get input into future features, prerelease code and world-class support.
Agreed, but other than support, this all comes with open source projects. I can get prerelease code from any of them; I can send an email to their list saying, "I want this project to Foo", and it gets taken under advisement. It just seems like an arbitrary line was drawn in the sand for what is called "enterprise".


The dual-licensing model that Riak uses is not new, especially in the database world -- Sleepycat did it years ago.  The devs are also committed to releasing functional open-source products, not crippled ones. 
It's interesting that you mention Sleepycat as a point in favor of a dual-license model, because I was considering mentioning them and MySQL as examples that this can can be risky for users. While these were successful businesses, both companies were purchased by Oracle and now their flagship projects sit in stagnation (debatable, I understand, but neither project is where PostgreSQL is now). Both of these companies were committed and did release very functional open source options. Now, decisions about these products are made solely based on the bottom line.
 
Riak Open-Source is by no means crippled, EDS just has some extra sauce.
This is clear. I wouldn't be writing you if I thought it was crippled. The concerns come from the future of the extra sauce, especially if the sauce spoils the meat. I guess the more accurate question would be, how are you going to ensure that the open source version doesn't become crippled when you start seeing dollar signs? What is the feature roadmap for Open-Source and EnterpriseDB versions?


Sean Cribbs <[hidden email]>
Developer Advocate
Basho Technologies, Inc.


_______________________________________________
riak-users mailing list
[hidden email]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com




_______________________________________________
riak-users mailing list
[hidden email]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
Reply | Threaded
Open this post in threaded view
|

Re: Riak Bug Tracker

Stephen Day-2
No worries all around. It looks like we're definitely bringing this discussion into a productive area.

As far as a podcast goes, I'm quite flattered by the request, and would obviously love to take part. Let me speak with my employer and we'll see what we can get scheduled.

steve

On Tue, Mar 16, 2010 at 6:21 AM, Antony Falco <[hidden email]> wrote:
Stephen --

Thanks for your considered and considerate response.  And thanks for taking on Issues 75 and 76.

I did not find your comments troll-ish.  Your comments (even the "evil" ones) were informed by the way the debate about open source typically shapes up -- often with a lot more acrimony.   However, as this exchange shows, the debate is turning into a practical discussion as open-source business models mature.  I have to say my own response was reductive -- in the last several months I have met some very impressive and fair-minded people working in venture who get both infrastructure and the proper value of community and open source.  The real point is that some companies and investors view communities as markets from which to derive profit.

To answer your question:  Yes there is a roadmap of open source and enterprise features.  I think the best way to present it is as a discussion.  Do you want to do a podcast with us?

Take care,

Tony Falco
Basho Technologies
m:  617-417-4482

On Mar 15, 2010, at 6:56 PM, Stephen Day wrote:

Tony:

Thank you for the detailed response.

As a company, I believe that Basho is a shining example of a dual-license open-source business model. Your support for the open-source community is definitely not in question. You obviously recognize the potential that community involvement has to improve the end product.

I'm sorry if I came off a little bit trollish and relegated my questions to conjecture rather than specifics, but this is the question I am interested in (though it may have come off as a rhetorical device):

Is there a feature roadmap for the open source vs enterprise versions? If so, what is the road map for the open source version?

As far as "future evils" go (I honestly feel bad that you read it that way ;) ), I am not worried in the least bit about riak or Basho. Again, thank you for your time and keep up the good work.

steve

On Sun, Mar 14, 2010 at 5:31 PM, Antony Falco <[hidden email]> wrote:
Stephen --

I think we understand your concerns and really appreciate your involvement with open source Riak.   

Just to clarify, Riak EnterpriseDS is a blend of some features and advanced support with SLAs, not just a bunch of valuable features "arbitrarily" held back.  While this is still technically an open-core licensing model (OCLM) it differs from most OCLM offerings in that it allows us to push many more features _and_ support hours into the open source community

 It also underscores an oversimplification some make that holding back features is categorically BAD but holding back support is categorically GOOD.  Both are business models defined by what value they hold back and for which they charge users of software.  (I am intentionally ignoring NoSQL-as-a-service companies here because the differences are not germaine to the point you raised.)

In my experience,  support-only companies would never be as aggressive about answering questions on the riak-users list  as we are -- note how fast Rusty and Sean responded today to questions (on a Sunday no less).    Look at our response times in general to questions on Riak-users -- mine aside, of course.  Our calculated decision to productize a blend of some features and some support allows us to offer more of both to the community for free.  More companies should do this.

There is an economic component to our choice as well.  Support-only companies scale much differently than our approach with OCLM and require a LOT more capital, usually from those stalwarts of benevolence, venture capitalists.    

Support-only models make money not off subscription-based licenses but off of support.   Support staff who focus solely on customers soon outnumber developers,.  Support companies give away their software for free then look for ways to "monetize the community."  We are under no such pressure.   We don't market to our open source users.   We don't have to.   We are not under pressure from outside investors because of our choices.

At Basho we are 19 partners (we believe the idea of "founder" in open source software companies is silly and corrosive to the community ),  of which 15 of us are developers.  We are self-funded mainly, though we have had critical help from some angels interested in open source software and scaling infrastructure.  The other projects in NoSQL are either venture-funded or subsumed under large public corporations or licensed under AGPL (or all three!).  Does anyone feel safer with a company 51+% owned by venture or who answers to Wall Street?

Contrary to your fears about us setting some sort of enterprise trap "even if [our] intentions were benevolent,"  Basho is getting more open with every release.  

We started building Riak two years ago and open sourced it in August.  Since then, the open source code base has grown at many times the rate of the enterpriseDS code base.  We chose the apache license for its open nature.  We have plenty of companies (like EA) who use Riak open source.  In addition, we run the EnterpriseDS for Startups program -- startups pay us whatever they can for features  _and_ Advanced Support.  

Companies that disingenuously enter into open-source licensing arrangements and engage a community of users will fail.   The users will figure that out soon enough.   I respect your concerns and believe that part of participation in a community is not just respecting opposing ideas but encouraging them.   I ask only that you show us the reciprocal respect of commenting on our actual actions, not asking us to defend ourselves against some conjectural future evil doings. 

Please keep contributing ideas and questions and otherwise actively engaging with Riak.   It is perhaps a quaint notion nowadays but we believe dissenting views strengthen communities.  

Tony Falco
Basho Technologies
m:  617-417-4482

On Mar 12, 2010, at 3:24 PM, Stephen Day wrote:

Thanks for the quick response.

On Fri, Mar 12, 2010 at 1:56 PM, Sean Cribbs <[hidden email]> wrote:
We have a bugzilla at http://issues.basho.com. Feel free to submit bug reports there.  For patches, send a pull request on BitBucket.
Thanks. This was hard to find.

There's no bulk-commit, but the HTTP interface supports Keep-Alive, so you should be able to submit multiple requests over the same HTTP connection.
I will try this.
 

3. Why are features that come in other projects as standard (web monitoring, site-to-site replication) bundled in an EnterpriseDB edition? This is a huge risk when considering Riak for adoption, especially when its not clear what future features might not make it into the open source version, especially if a larger company purchases Basho. For instance, if there is a great community patch for a monitoring interface or site-to-site replication or any other feature that you want to monetize, it seems doubtful that you would consider accepting it. I understand that you need to make your money, but I have to say, this model raises concerns, even if your intentions are benevolent. That said, hypothetically, I would be more interested in support, if it could be considered without having to weigh it against arbitrary bundling decisions.


You can start your development with the open-source version of Riak, then when you're ready to  add those features, talk to Basho about upgrading to EDS.  I believe you'll find Riak Open-Source to be very capable even without those features, and that when you're ready for the features, you'll also want the direct relationship with the Basho team that EDS provides.  In addition to the feature bullet-point stuff, you get input into future features, prerelease code and world-class support.
Agreed, but other than support, this all comes with open source projects. I can get prerelease code from any of them; I can send an email to their list saying, "I want this project to Foo", and it gets taken under advisement. It just seems like an arbitrary line was drawn in the sand for what is called "enterprise".


The dual-licensing model that Riak uses is not new, especially in the database world -- Sleepycat did it years ago.  The devs are also committed to releasing functional open-source products, not crippled ones. 
It's interesting that you mention Sleepycat as a point in favor of a dual-license model, because I was considering mentioning them and MySQL as examples that this can can be risky for users. While these were successful businesses, both companies were purchased by Oracle and now their flagship projects sit in stagnation (debatable, I understand, but neither project is where PostgreSQL is now). Both of these companies were committed and did release very functional open source options. Now, decisions about these products are made solely based on the bottom line.
 
Riak Open-Source is by no means crippled, EDS just has some extra sauce.
This is clear. I wouldn't be writing you if I thought it was crippled. The concerns come from the future of the extra sauce, especially if the sauce spoils the meat. I guess the more accurate question would be, how are you going to ensure that the open source version doesn't become crippled when you start seeing dollar signs? What is the feature roadmap for Open-Source and EnterpriseDB versions?


Sean Cribbs <[hidden email]>
Developer Advocate
Basho Technologies, Inc.


_______________________________________________
riak-users mailing list
[hidden email]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com





_______________________________________________
riak-users mailing list
[hidden email]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com