The OCTOpod: Conversations with SUSE's Office of the CTO

Creating an Open Source Startup with Sheng Liang

August 05, 2021 Alan Clark - Office of the CTO, SUSE Episode 3
The OCTOpod: Conversations with SUSE's Office of the CTO
Creating an Open Source Startup with Sheng Liang
Show Notes Transcript

There are many potential benefits to adopting the open-source approach to building a start-up and here to discuss some of these and his experience in the field is SUSE's President of Engineering and Innovation, Sheng Liang. In our conversation, we get to share in some of Sheng's wisdom and business know-how, talking about pragmatic solutions to common issues from the world of open-source and start-ups more generally. Sheng explains why he thinks open-source is a smart path for businesses but is not the answer to every question. He gets into why the agility of open-source can also be of benefit to companies in their early phases, and the lessons that he learned during his work at, prior to joining Rancher Labs and SUSE. Sheng makes the point that without developing a relationship with your users and creating something that people actually care about, your ideas will never have an impact, and believes that being able to pivot when a great idea does not connect should always be an option. So for all this and more from our inspiring guest, be sure to listen in!

Key Points From This Episode:

  • The beginnings of Rancher Labs and Sheng's experience of starting an open-source company. 
  • How the open-source philosophy figured into the initial plans for Rancher. 
  • Differentiating an open-source approach from other business models.
  • Sheng's perspective on when going the open-source route is smart for business. 
  • Addressing the problem of the number of non-paying users in the open-source model. 
  • The two typical communities that open-source markets to: developers and users. 
  • Balancing innovation and stability at a new start-up; Sheng's best advice to this end. 
  • Sheng's approach to keeping the start-up spirit alive within Rancher now that it is a part of SUSE.
  • The most important considerations when deciding about whether to follow an open-source route. 
  • Advice from Sheng about software startups and the central question of creating engagement.
  • Weighing the feedback your company receives and what to take most notice of.

Links Mentioned in Today’s Episode:

Alan Clark

Sheng Liang on LinkedIn


Chris Aniszczyk

Linux Foundation


Red Hat








[0:00:03.5] Welcome to the OCTOPod, hosted by Alan Clark and brought to you by the SUSE & Rancher Community. Alan has spent his software career, focused on Open-Source advocacy and emerging tech. In The OCTOPod Season 1, Alan talks with experts across technology about trends and challenges in open source. From building communities to diversity and inclusion. To keeping the "open," in "open source."




[0:00:36.7] AC: Hi everyone, welcome to the OCTOPod. Today, I’m joined by Sheng Liang. He is the President of Engineering and Innovation at SUSE. He leads SUSE’s global team of engineers and is responsible for SUSE's rapid growth and expanding portfolio from the enterprise Linux operating system to the Rancher Container Management Platform.


Sheng was Co-Founder and CEO of Rancher Labs and CTO of the Cloud Platforms Group at Citrix Systems. Sheng, it’s great to have you here with us today.


[0:01:09.1] SL: Thank you Alan, it’s great to be here.


[0:01:10.9] AC: Sheng, this season, we’ve been talking a lot about open source and we really wanted to talk to you today about starting an open source company. How did Rancher Labs actually come about?


[0:01:23.1] SL: Yeah, thanks for the question and Rancher Labs was actually the second open source company I started back in 2008, I started a company called and there, we created an open source infrastructure as a service software platform. That was very successful and later on, we became part of the bigger, Open Stack effort, we were then acquired by Citrix throughout that process, we realized just working on open source infrastructure wasn’t solving our customer’s needs.


Most of our customers, they have applications to deploy, more so than infrastructure management, they have a devops management challenge and that’s where we got interested in Docker and later on the Kubernetes communities started going, so we became part of that and that’s really how Rancher came about.


[0:02:24.6] AC: That’s interesting. What you thought about Rancher and start putting it together, were you thinking that you wanted to make it a company based on open source? Was that one of the guiding beginnings or principles there from the beginning or was it something that grew out of the DNA of the idea?


[0:02:44.7] SL: Not necessarily. You know, we never start a business – I never started a business thinking it was going to be an open source business because starting open source businesses are very, very challenging. I would still say, it’s a little bit more challenging to make a viable open source business than non-open source business. I think the payoff and the impact could be very big.


By going with open source you’re giving up one of the very obvious means of monetization very early on, in exchange for maybe awareness, maybe a broad options but that’s in general, very difficult to pull off. As with any startup, you got to start with the market need, the product market fit and then decide specifically, based on what kind of products you want to create, what will be the best way to distribute it and to market it. Open source would be one way to do it. It so happens that it worked for me in the last couple of times.


[0:03:54.0] AC: That’s very interesting and that actually ties into something that we talked about with Chris Aniszczyk last week. Chris, as you know is from the Linux Foundation and has been one of the leaders, very visible within CNCF. He said to us, open source, quoting from Chris, “Open source is not a business model, it’s a method to collaborate and develop software. How your business makes money is completely different.”


Can you comment a little bit in Chris’ notion here about how the business model is different and separate from the community?


[0:04:33.5] SL: Yeah. I mean, generally, open source communities, a lot of times consisting of developers and IT practitioners, they’re not first and foremost looking for something to buy. They may have a problem, I think they’re a barrier to adopting some new and interesting technologies pretty low. Open source makes it far more interesting, far more attractive usually for target user base to adopt the technology. 


A lot of times, these users are not necessarily the ones whoa re going to buy software or purchase software and if you sell to consumers, generally, your users are buyers but if you sell to a business or an enterprise, the users are not necessarily making the decision to purchase software or they may not individually have the budget to make a purchase.


That is why you need to make that association from adoption and use to eventually solving a business problem and that these users that would help you to get the budget released from their organization to pay for a commercial subscription like SUSE and Rancher, companies like those, what we’re doing today. 


[0:05:56.8] AC: That makes sense. Let’s change the direction here just a little bit and maybe try to connect some of the dots here on the business acumen in building successful business around open source. There’s lots of examples where businesses were unable to monetize the potential value, right? They’ve used and leveraged open source with the notion or the idea that it would bring them value, right? 


Like you said, awareness or help them to innovate quickly or whatever. But there’s been several cases, some of them very famous where competitors actually achieve the benefits to their detriment and the question, what I’m looking for here, is can you provide some guidelines or examples of when open source is smart business?


[0:06:42.8] SL: Yeah. I would say, generally, open source is smart business. I mean, the question of, if you had a piece of open source technology that it has achieved mainstream adoption. It is true, you may not still be able to build a successful business on top of it but this is very similar to asking, say, if you had a website that a lot of visitors, that still doesn’t mean you can be able to good web business.


I would say, in most cases, you can. In a lot of cases, folks have been spectacularly successful. That’s why as long as we’re talking startups, I mean, we have to talk about investors and that whole ecosystem. Generally, it’s how there’s truth or even an axiom that it is safe to invest in technology with adoption, open source technology, with adoption or with the promise of widespread adoption.


Then, the second step of figuring out a business model on top of it is relatively low risk. It’s not without risk but it’s relatively low risk, there’s just so many ways of creating a viable business out of open source technologies with wide adoption.


[0:08:00.3] AC: Yeah, but as you say that, totally get the adoption, it’s the second step there that received some criticism, right? One of the business investment criticisms of open source is that a percentage of the users won’t pay, right? It must be interesting to pitch for business capital when faced with this or similar criticisms. How do you address those?


[0:08:22.3] SL: Yeah, actually, in the history of Rancher, we’ve always had this challenge. We always viewed it as more of an opportunity. Throughout the history of Rancher, we’ve had about 1% of our user-base that paid and we consider that our greatest strength as supposed to a weakness because that remaining 99% brought us widespread awareness and it represents extremely lucrative, untapped market that remained to be tapped and monetized as the technology hits mainstream.


At least in my career, I have personally, as a product person and as an engineer, I’ve spent far more of my brain cycles worrying about how to grow that 99% versus growing that 1% of the customers who pay. There’s just so many reasons for people who want to pay for the software once they use it for mission critical production use. They would need the technical support, they would need the ISV certifications, they would need the IHV or cloud certifications and they will need maybe some proprietary add-ons that we’d be able to offer on top of the 100%, for the open source software.


Like I said, generally, they’re about a handful of tricks that these open source companies can use to derive revenue and exactly which lever you pull, which trick you go for, again, depends on the product and the market but these have been proven and I think at this point, my experience is it’s not been nearly as difficult to convince the investors or customers that you do have a viable business.


[0:10:20.1] AC: Do you think just to follow on to that, I would imagine that as you said, it hasn’t been as hard to convince, albeit that has changed dramatically over the last say five, 10, 15 years, do you think?


[0:10:33.8] SL: That’s indeed the case, when I started my first company,, we initially – I don’t think we did open source and we quickly did do open source, then we switched to Apache open source and throughout this process, back then, in 2009, 2010, that timeframe, there were definitely still a lot of people complaining.


Open source is a very difficult business model, they’re really only two companies that’s been successful at that time and that was Red Hat and SUSE, but outside of that, you don’t’ really see anyone else. Look at the last 10 years, the last decade or so, you know, there’s just been a constant march of extremely successful open source businesses. You know of course, Red Hat and SUSE continues to grow but then on top of that you’ll have Elastic, MongoDB, Confluence, you have HashiCorp or Rancher, it just goes on and on right? At this point it is no longer in question that you can build an open source business. 


[0:11:41.3] AC: Very good, so let’s change directions again just a tiny bit and we touched on this at the very beginning, about the community and what you would be looking — so, first, the question is when you’re developing the business and you want to leverage open source as part of that business model, what are the benefits that you’re after? What are you looking for from the community? 


[0:12:04.8] SL: Yeah, I mean there are typically two things when it comes to the community and two types of community. You have a developer community, people who can potentially contribute code or bug fixes or things like that directly into your code base and the other would be your user community and I always believed user community is what’s truly valuable for open source. Developer community interestingly is much, much harder to achieve also. 


It is a lot easier for open source projects to grow its user community than developer community because by definition, the entry to barrier to get someone to contribute to the software is just a lot harder and only a handful, literally a very, very small number of open source projects achieve the level of having a vibrant multi-vendor developer community and those will be projects like Kubernetes and Linux and you know they are at that level but the majority of the successful open source projects, they’re successful because they have a huge user community. 


Projects like Rancher for example, extremely successful, right? Hundreds of millions of downloads and tens and thousands of live deployments running at any given time. That level of scale is what makes Rancher a viable open source project and that’s the foundation that we could be validate very successful and very fast growing enterprise business.


[0:13:41.7] AC: Which leads to, maybe this is too big of a jump here but at least to an interesting question, right? Because user and developers have sometimes different goals and different perspectives and one of those is around stability versus innovation. Should a startup focus on stability or fast innovation from an open source community? 


[0:14:03.5] SL: I would say it’s a lot easier for a startup to focus on users and focus doesn’t mean it’s easy to get a lot of users. For an open source project to get users is just about as difficult as for a say, a musician to get audiences and we all know how hard that could be. We definitely are in competition for people’s limited amount of time and limited amount of attention and that is the competition for users, competition for adoption is fundamentally going to make the project higher quality, easier to use and naturally be able to solve a real problem. 


If you don’t focus on that then it really doesn’t matter. You know, it doesn’t matter these days to launch a project and say, “You know, I’ve got the best technical innovation. I’ve got the fastest performance” but this is sort of like saying, “I compose the piece of music. It’s got perfect compliance with every single music theory in the world, right? But why isn’t anyone listening to it?" That is the issue. 


[0:15:21.8] AC: Okay, that makes sense. As we know, Rancher was recently acquired by SUSE and since SUSE is also an open source company, you know many of the aspects of the business and the people and the cultures with SUSE & Rancher of course align very well, right? Rancher is well known for it’s startup spirit, so how are you keeping that spirit alive as part of SUSE? 


[0:15:49.8] SL: Yeah, I mean today, Rancher team and the SUSE team they more or less function as independent engineering teams inside SUSE and that’s both good and bad. It’s good because that way, both teams are given a fair amount of continuity and they can progress at their own pace but it is also bad maybe they have not been able to cross-pollinate as much as they could have been and it’s out of necessity too because Linux business and our container business are very different and the pace of innovation is different. 


The stage of market development is different. The customer and user expectation is also very, very different. In the Linux world, there is more of an established practice to work with upstream code, whereas to push all the innovation upstream. Whereas when Rancher was born and especially, you know, that was sort of the more recent way of doing these open source projects, it tend to be more focused on adding value, as a separate open source project like Rancher driven by a vendor to overcome a bunch of upstream open source projects. But maybe vendors aren’t necessarily just taking the upstream and productize it as is. There are some fairly subtle differences but they are important differences and we got to manage those two efforts differently. 


[0:17:18.7] AC: Okay, so it sounds like you’re managing to those differences more than what was Rancher and what was SUSE, right? 


[0:17:26.0] SL: That’s right, you just got to be pragmatic. You know, it will be wrong to say, “Now we’re the same company and we have to follow the same process.” Linux is obviously a bigger business and therefore, if Rancher just followed the Linux way or Rancher is obviously more agile if Linux followed the Rancher way, and I think if we took any one of these approaches it would have killed those businesses. We just got to be pragmatic. 


[0:17:50.2] AC: Interesting, so thank you. Before we run out of time here, I really wanted to ask you the question, if you were to go back and do it again or if you were going to do it again somewhere else sometime in the future, would your next startup be open source based? 


[0:18:07.7] SL: Probably but like I said that I would just always tell someone who is asking for a device, I mean if he has given me a very concrete idea on what he’s going to do or this is the sort of code we’ll be writing, then it will be a lot easier to say, “Well, it didn’t make sense to distribute it as open source software.” Then create the adoption, create the grounds well then be able to enterprise subscription business out of that but if the idea is not as concrete as that yet, then it’s harder to say, because there are just many, many different other kinds of models that are equally successful, if not more successful. 


The SaaS business model is extremely successful, right? That is very good if you are doing more application level stuff. There is really no reason you have to write your business application as open source software. I think my answer is it just depends on exactly what you want to do. 


[0:19:03.5] AC: Makes sense. Any final advice for those types of folks that have the idea of following your footsteps and creating a successful software startup? Any final words of advice? 


[0:19:17.8] SL: Yeah, definitely. I think it’s just important to ask yourself, will people care? Would they care enough to take a look? Will they care enough to contribute? Will they care enough to download and use? Will they care enough to join your crusade to make it a better piece of software? That’s very important. Because I know a lot of times we’ve got our own insight and we’ve got our sweat and tears into a piece of software and it is easy to get carried away. Whereas I think people’s taste and the real adoption, those things are hard to fake. 


Generally, if you put your piece of software out there, if it is not getting any traction after a few days or a month or whatever, then it is probably time to change direction and that’s why open source is – I mean, I disagree with Chris a little bit, it is even a great business model because in that sense, because it allows you to pivot much faster than a traditional business where you’d have to involve sales and marketing and those efforts to really take months to play out, right? 


[0:20:26.0] AC: Takes months to years, yeah. 


[0:20:28.7] SL: Yeah, sometimes years because you can’t have a false negative, you know? That is even worse. If only you kept that in, you know, if you had a little bit more perseverance, right? You would be successful, wouldn’t that be a shame? But open source generally, I mean you put your stuff out there literally within days you get a good read and they’re like – and within a week or two or a month or two, I mean you know, it is probably not going to change unless you make a change, okay? 


[0:20:57.5] AC: Makes sense. Gauging that feedback, how do you know that you are getting the right feedback, the right audience, right? 


[0:21:05.2] SL: I would just say, I mean there is no right or wrong audience. There is only the level of engagement that you can – those are our objective numbers. You know, how many bug reports have you seen? How many comments are you getting back, right? Are people loving it, hating it? I think even, do they give a damn? That’s the key, as long as people care, do they care? If all these people care, then there is something in there. If nobody cares, then try something else.


[0:21:34.9] AC: Try something else, perfect. 


Well, thank you Sheng. It’s been terrific spending a few minutes with you today. Thank you for your time and thank you for participating with us and sharing your wisdom with us. 


[0:21:47.0] SL: Thanks, Alan. 


[0:21:48.2] AC: Thanks everyone for listening to this podcast. Catch us again soon. 




[0:21:53.6] ANNOUNCER: Thank you for tuning in to this episode of The OCTOPod. Subscribe to the show wherever you listen to podcast, so you’ll never miss an episode. While you’re at it, if you find value in the show, we’d appreciate a rating on iTunes or if you simply tell a friend about the show, that’ll help us out too. You can connect with us at SUSE & Rancher Community at