Community and Cloud.com: Interview with Mark Hinkle

Musings on community from Cloud.com's VP of Community

Mark Hinkle
A few weeks ago, Cloud.com snapped up Mark Hinkle from Zenoss. Hinkle has been doing community development for Zenoss for some time, so I was curious to see what could lure him away and what Cloud.com's community strategy might be.

Tell us a bit about Cloud.com and where it's at now with regards to community — how many contributors do you have? Is community contribution a significant part of cloud.com's development at this point?

Hinkle: Our community is fairly small since launching in May of this year but we do have a growing number of users and advocates (like the guys from Greenqloud in Iceland) but we are still in our infancy. We also have some projects going on where we have strong partnerships with RightScale and CloudScaling who are helping users be successful with our software but much of that has not been institutionalized beyond those organizations. I think because we are on an aggressive development schedule our two most significant forms of contribution are going to be in forum participation, product feedback and bug submissions especially during the beta periods. Over time as more users get their cloud infrastructure running I am sure there will be many other more tangible and substantial types of contribution like patches and code contributions. As you look at most projects less than 1% of users offer any kind of code contribution so for now the priority is to make Cloudstack attractive for widespread adoption.

What prompted the move to Cloud.com?

Hinkle: There are really three reasons I chose to go to Cloud.com. First, I prefer to go where the action is. In 1995, I worked in consumer Internet; In 2000 I went to work in the application server provider (ASP) space, and soon after virtualization. The one reoccurring theme is that all the companies were heavy users or developers of open source software, which will continue to be a driving factor in my decisions.

Given my penchant for the bleeding edge of technology I can't think of a more interesting, fast-paced technology segment than cloud computing. The unique angle in open source about cloud computing is that in other software segments like ERP, database and operating system you could easily point to a proprietary technology and say just like "____ but open source." Cloud computing as an industry is still developing and even for start-up it's still possible to be a leader rather than a follower, something I find very attractive."

Second, I was very impressed by the organizations desire to build a strong open source project. It's in the in the DNA of the company at every level from the board to execs to development team. A lot of open source companies say that but fail on their commitment to developing a community project, I am confident that's not the case here.

Finally, I really think our technology is exceptional and has some advanced features that are extraordinary in comparison to most cloud computing software (especially support for all hypervisors in the OSS version). This is really unique, compared to the usual case in open source software where you are playing catch-up with established proprietary software and the software is often labeled "good enough for open source" but not the most complete. With Cloudstack, I think our cloud computing software could legitimately be the best and be open source. I think that's a rare opportunity.

Looking back on your tenure with Zenoss — what did you learn about community there that you didn't know when you joined the company? How will you apply that at Cloud.com?

Hinkle: The most important thing I learned is that you can't cause a community to do what you want, rather you need to be there to enable those community members to do what is useful to them. Often contributions come in ways well outside the realm of core code development (recommendations, ideas, introductions, etc.) and you need to make sure to recognize their contributions and help institutionalize them.

Also I learned that Bill Joy's Law is still well in force ("No matter who you are, most of the smartest people work for someone else"). The takeaway is simply that in an open source development model you are in a unique position to capitalize on the expertise of these smart people. At Cloud.com I am very impressed by depth-of-knowledge our internal resources possess but I look forward at those ideas that I expect to come from outside the company.

What are you hoping to accomplish in the first year with the company?

Hinkle First and foremost my goal is to build a community of inclusion (I borrowed that term from Opscode CTO Adam Jacob), which means that we take all comers and don't try to only attract a certain type of user but rather all users from tinkerers to enterprise users. Then we'll focus on making those community members successful with CloudStack. I really like the Red Hat's philosophies spelled out in their book, The Open Source Way and hope to learn from their success.

Beyond that I think it's important to make strong community partnerships with complimentary projects and open source groups and to drive collaboration. I think by being a cloud computing platform we have a lot of opportunities to work with a wide-variety of technologies including management projects like Puppet, Nagios, Chef, Zenoss, etc. as well as virtualization projects like Xen and KVM.

In a nutshell my goal is to make sure our software easy to use and the community attractive to participate in.

It looks like Cloud.com is pursuing a "open core" model with Cloud.com - some features held back in the enterprise edition. What's "missing" from the community edition?

Hinkle: You hit on my biggest open source pet peeve; I am not a fan of the term open core as I find it can be pejorative. I feel strongly that as a business that has an open source project you need to look at things this way. You must have an open source project that stands on it as own and provides a lot of value without the commercial extensions. If things are working right the commercial entity is something that appeals to some but not all users. I'm no Pollyanna though, as a business you need to sell something that people want, it doesn't matter if it's software or services in this business.

What I have observed though is that among businesses that have a significant open source component that they often try to coerce upgrades to commercial versions of their software. That's why I don't like the "open core" moniker. I think most open core companies end up creating a business out of upselling minute features that add very little value. Admittedly solving this problem has been difficult for the industry and there are very few companies that I think got it right. Red Hat's probably done the best job of it by sponsoring a thriving Fedora community and selling Red Hat Network and Satellite Server for Red Hat Enterprise Linux.

I am still trying to figure out the specifics but I believe the best business model adheres to the Tim O'Reilly principle, "Create more value than you capture." Which when done right, can yield both a thriving open source community and a profitable business.

With that being said the only thing Cloud.com currently plans to hold back at least for now from the open source cloud software, Cloudstack Community Edition is the billing and charge back capabilities. We still have other customer specific features that we don't distribute today but that's largely because they don't make sense for the majority of users. What I expect is that we will continue to broaden the open source Cloudstack but will target the best use cases for broad adoption.

Today a few of the features in our Cloudstack Service Provider Edition wouldn't make sense for the majority of users. As a business we are often called upon to add features to address very specific use cases but wouldn't make sense for many other users. What we are doing is are looking to hire a team to work on features that are requested by our community users so we can be responsive to their needs. As is usually the case there will be a large overlap between community and commercial editions.

Also we are a young company I think we still have a lot of things to figure out. As we gain users I think we'll learn from them what products and services they want to buy and address those requests. Right now we are in an enviable position we are outpacing our business plan for sales and we have staggeringly large number of commercial prospects. I hope to have the same problem with the community side of things shortly.

Copyright assignment and contributor agreements have been fairly contentious in 2010 - what's the situation for contributors to cloud.com's community? Do you have a copyright assignment requirement — and is any of the code being contributed by the community likely to turn up in a proprietary version?

Hinkle: We do use a contributor agreement that's almost identical to the Apache contributor agreement that grants Cloud.com both copyright assignment and patent license. Since Cloudstack Community Edition and Cloudstack commercial editions are sharing a large amount of code then it's probable that contributed code would be in both. Though we don't expect that we would take code in the commercial version and not include it the community edition.

Why not use the Affero GPL (AGPLv3) rather than GPLv3?

Hinkle: Two reasons, one GPL is the most common and well-understood license used in open source software and we feel people are simply more comfortable with it. As for AGPL we feel like this would alienate service providers who may want to extend our code and build a business on it. It's our opinion that AGPLv3 would only protect our interests and not our broader community.

What's the company's position on OpenStack — any chance of participating in that community?

Hinkle: We are very excited about OpenStack and publicly announced support for the project on the day it was announced. We have a public statement of why we think OpenStack is a good thing and our reasoning for participating. Now the harder part is figuring out how to contribute especially to the cloud compute part of the project. In addition, Microsoft has already enlisted us to develop support for Hyper-V for OpenStack and in October we submitted that code to the project and did so even before working on integrating Hyper-V with CloudStack.

Join the Network World communities on Facebook and LinkedIn to comment on topics that are top of mind.
Now read: Getting grounded in IoT