Interview with China Cloud Veteran Steve Mushero
On building China's first cloud and internet managed services company, and the rise of the cloud in China.
I recently had the opportunity to speak to Steve Mushero, whose professional career is closely intertwined with the development of cloud computing in China. In 2008, he co-founded ChinaNetCloud, which he describes as China’s first cloud computing and internet managed services company. After maintaining a leading role at the company for over a decade, he recently returned to Silicon Valley.
If you’re interested in the history of Chinese cloud and tech, particularly the early days, this is a must-read. In our conversation, we discussed the growth of China’s cloud industry since the 2000s. This also included the circumstances that led to the founding of ChinaNetCloud, as well as what it was like to work with leading cloud providers like AWS and Alibaba Cloud.
The following is a transcript of our conversation, which has been edited for length and clarity.
Getting into the Chinese tech industry
Zac Haluza: Could you briefly sum up your background regarding China, IT, and the cloud?
Steve Mushero: My background is in tech. Originally in heavy manufacturing, industrial automation, robotics, and all that sort of stuff. That eventually morphed into, software, systems, ERP, finance systems, and so on. I was in Seattle and Silicon Valley, all over the place.
I eventually ended up in China because I wanted to be where the action was and see what was going on. I moved from Silicon Valley to Shanghai starting in 2005. I was there full-time by 2007. Even at the time, there was a huge connection between Shanghai and the Bay Area, so it was easy to meet people and get connected.
I ended up at Tudou, the big Chinese internet video company, which is actually older than and was bigger than YouTube at the time. And much closer to what Netflix is today, in some ways. I was the only foreigner. I was already right at the core of Chinese internet and video, because China had just put DSL in most places.
Around 2005-2007, people were just getting a) online in general, b) online with broadband connections, which were fairly ubiquitous, for twenty bucks a month. People had computers; people were bored. Suddenly video became hugely hot, like streaming is in the States now, but that was already in China in, let’s say, 2007.
People in the US were still watching network TV. You’d watch Sex in the City on Tuesday nights at 7 PM. People in China were already picking any show they wanted at any time.
I was there at Tudou for a while, helping them with their online systems, CDNs, and such. Then it became obvious that the operations and managing of all this infrastructure — servers, databases network, security, performance — all that stuff was a huge issue in China. There really wasn't a lot of skilled people to do it because the industry grew out of nothing.
We were already the world's biggest internet, with 400 or 500 million users by then. Bigger than the US, bigger than Europe, yet it was only about five years old. The skilled people available to do things — nobody had five years’ experience. And so that was an obvious issue. Even for us, this extremely well-funded and famous company, we couldn't find people to do stuff. So imagine what all the other guys were doing.
My business partner and I struck out on our own and started a managed services company, which was ChinaNetCloud, to take care of those issues and provide management and operation services for internet companies, startups, even games. The iPhone had just come out. Mobile, apps, infrastructure, and e-commerce were becoming big. And of course, we picked the three sexiest words: China, net, and cloud. We put them all together with an eye toward the cloud. But cloud was not there yet in China when we started in 2008. In fact, we were the first people to offer public cloud, as far as I know, in China at the end of 2008.
The early days of cloud
Steve Mushero: If we want to focus on the early history, because it’s not really well-documented — when we got started at the end of 2008, everything was in data centers and people did corporate data centers hardly at all. So it was all shared data centers, co-location. We started off managing that stuff. This is physical servers: Dell servers (Dell dominated China) in racks, in data centers. Server by server, rack by rack. We started managing all that stuff from day one.
Of course, cloud had started outside of the US. VMware was present in China at the time, but that was very much on-premise or in data centers. You might call it private cloud today, but it was very much virtualization. We actually used VMware for some things at Tudou. But there was no public cloud whatsoever, as far as I know.
So our business was management, right? And so the problem we had was we wanted to charge clients $100 or $200 to manage a server, and 10 servers would cost 10 times that. But startups would go and spend 5,000 per server to buy them through, for instance, Dell. Well, if you spent $50,000 buying 10 servers, you often didn't have the extra $1000 a month we wanted you to pay us to manage that, because all your cash had been spent on the servers. We wanted to start offering public cloud and virtualization — not because it's a great business, because even today, most public clouds lose money — but because we wanted the management.
We would sell you small slices, basically VMs. We would charge you $20 a month for that and $150 to manage it. And it would cost us nothing. That was our whole driving force. We never wanted to be a big cloud provider, but we wanted those management fees, and we wanted you to save money and spend most of it with us and not Dell.
We started that right around the end of 2008, early 2009, for various customers. Some customers wouldn't want that. We’d buy servers, use their servers, and start a cloud. There was nobody in the market at all. At least a year or two later, there was CloudEx, a spinout of 21Vianet, which was one of the main data center providers and probably the largest now.1 Sort of a copy of AWS, which was just getting going.2 That was going to be the first Chinese file provider [before it collapsed in 2011]. We never managed anything on it. We got very close.
Then there was a company called HiChina (万网). HiChina was predominantly a DNS provider, like network solutions. They were the main registrar at the time, I think. So if you wanted games.cn, games.com, or whatever, they were the registrar. They also provided email services — a whole bunch of these sort of mixed-up consumerish (but really small business) type of internet services. They started their own cloud right around the same time, 2010-ish.
This is all leading to Alibaba. They were offering cloud before Alibaba was.3 We actually just started using them and partnered with them and were ready to use their services when Alibaba bought them.4
Alibaba was already offering cloud services by that time. But they were still super small. But then Ali bought HiChina, merged those together, and eventually merged a cloud together. By now, we had been through our own cloud, the defunct cloud [with CloudEx] and HiChina, which essentially died also by being bought in. Ali was actually the number-four player, and the only one that survived. All in about 24 months.
Then we started working with Ali, but it was very challenging. They wanted us to help them with features, reviewing, management, and operations. And then the management — and particularly market sales and marketing — of cloud switched back and forth in a never-ending war that lasted two or three years between Ali and HiChina management. And at one point, our counterparty on the Ali side changed every month because it was just in disarray. Sometimes the prior person came back and then they fired all the salespeople.
Ali was run by tech folks, and they didn’t really get sales at all, because they’re Alibaba — they don’t have to sell; they’re famous. So the idea of having a sales team, marketing, and partnerships was new to them. It was management turmoil.
As Alibaba was preparing for their IPO in the same period,5 there were issues running Ali, cause they also had a Google fight on Android. They were building their own operating system and their own phone. And that was mixed into all of this mess too.
At the end of all of that, Ali was the only cloud provider operating [before AWS entered the China market]. That's how it went from nothing to Ali.
China’s cloud matures
Zac Haluza: So after that period, things started to even out a bit and get a bit more consistent, in terms of the partners you were working with.
Steve Mushero: Yeah. So Ali was and is the big domestic player. Tencent started introducing their cloud probably a little bit later, but they were never a significant factor.6 The only other domestic large player was Huawei, which today is doing more enterprise cloud. But they weren't even in the picture that early on. It wasn't until 2017-2019 at least that anyone even wanted to use them.
Those are the three big domestic: Aliyun, Tencent, and Huawei. Huawei today is predominantly private cloud and enterprise cloud.
Then there were three well-funded Chinese startups: Xingcloud, UCloud, and Archnet. Xingcloud went nowhere eventually. I remember the third one was based somewhere like Wuhan, and they took a different approach. They bought up a bunch of things, and Intel invested. So they were floating around by 2015-2016. A lot of money got put into those guys, somewhat foolishly. They kind of never went anywhere because obviously, the big ones are going to dominate.
There were three big Chinese — Ali and so on — three small Chinese, and then there were sort of two or three international: AWS, Azure. And to some extent, IBM. They tried a little bit.
AWS started sniffing around the market around 2011-2012, maybe. We started talking to them and offering advice on how to do stuff in China and so on. They eventually, after a long time, got going [when they launched their Beijing region in 2014]. We were one of their launch partners, the managed service provider for them. We were the first managed service provider. I was the first Amazon-certified person in China around that time.
But they were held artificially small by all the regulatory issues around things, as a foreign cloud. And so they grew very slowly early on. They had not only the VIE but they had to have a partner, and they only had one data center, and they had to invest a lot in China. So they had really limited functionality. It was hard to do a lot for many, many years. Until later, when they became legal and so on, they really didn't invest in features. They sort of stood up a core AWS with, you know, S3 and a few things, but, but none of the more advanced features that were developing rapidly outside of China.
But we worked with AWS right in the beginning. We already had experience from overseas, and they had a lot of good people, and they spent a ton of money and they hired a lot of people. AWS eventually got going.
AWS was completely divorced from the rest of the world. It had its own cybersecurity and users and all that. It had a region, but it wasn’t a linked-in region.
Azure came in, I want to say, a little before AWS. They did a partnership, a much more hands-off regulatory legal partnership with 21Vianet also, who was supposed to essentially run the whole thing, just using Azure technology. Which isn't fully what happened, but was kind of the goal. Microsoft did that to get Azure running. Microsoft also really wanted to get Office 360 running.
Because that would let them for the first time really capture revenue from Microsoft Office that was not usually captured before this, because people would pirate it. Everybody had pirated copies of Office, and nobody paid anything. Obviously, as soon as you go online, that picture changed. That was a big strategic priority for Microsoft.
Azure never did well in China. They begged us to work with them for years and years. We said that we’d love to, bring us some customers… But they were never really able to do that. They were and are around. I think Office was way more successful.
That's how AWS and Azure ended up there. Google Cloud of course never happened, because by then Google had been blocked. So that was its own separate set of messes.
That's sort of the genesis. It started from nowhere, and we started things, I guess. There were some hit-and-miss players, and then it morphed into where it is today, which is really Ali, to some extent Tencent, Huawei, and then AWS and Azure. Those five are the main players for most things today, and that was already the case by 2015.
Policy wrangling
Zac Haluza: You mentioned earlier how AWS came up against a lot of regulatory issues, general policies, and just a lot of obstacles in front of them. I was curious if you faced a lot of obstacles from Chinese policies in your experience with ChinaNetCloud.
Steve Mushero: We did not. With AWS it was endless roadblocks, but for us, no. We were small and not famous, I suppose, but we also structured things right from the beginning, having had experience in the market.
So even when we were offering public cloud, we did it in part through partners. For example, we had data center partners. We would buy the hardware usually, or we’d lease it, which was already hard enough. But the controls in China are around bandwidth. So we would buy servers, put them in a data center for you, and you would sign a contract with us for those and the management. That's just a service; that was not restricted. But when it came to the bandwidth — the actual connection, the wire to the internet — that you had to buy separately through a licensed proper third party, usually the data center. You had a legal contract between you, the customer, and the data center bandwidth provider. That is where the legal liability came in. So that's where you said you're not going upload porn or political stuff or anything illegal.
So we pre-organized all this stuff and had all these relationships. To you, it was just a few pieces of paper, but you actually were working with two or three vendors simultaneously, so that everyone was legally protected and it all worked.
Plus, we were completely above board on doing things right — legally and correctly in China. And you may not like that. You may not want to adhere to government regulations, but this is how it is. We published books about this, about censorship, licensing, registrations, all that sort of stuff.
That's also a way you avoid problems and scrutiny. Usually, the way that it would happen is it would go to the data center and bandwidth provider who would have a complaint, or a problem, or a police thing. Then it filtered down to us.
People would allow unmoderated forums for discussion of, say, your products. And you don’t want to do that because, for one thing, it ends up with all kinds of unmoderated stuff. We had one customer where someone uploaded, you know, Buy guns, here's a phone number. You might have observed guns are not that common in China, nor are they legal. So this gets reported to the police. The police go to the data center. The data center comes to us, and we go to the customer and tell them this has to be taken down. And it was a Friday afternoon or something. We pulled either the whole site or that part of the site offline.
This is for a famous multinational company, who will remain nameless. But you just don’t fool around with this stuff. It’s very simple. So they make a complaint to us on Monday. You know, “we're gonna sue you,” and all these things. We're like, look, this is really simple. There was illegal content on your site. We took it off. Sorry if you don't like that; feel free to sue us.
Because they didn't want to pay for moderation. They didn't have the budget for it. They didn't have someone who was gonna watch it. This is not complicated.
That's where issues would come up. We never ran into news or politics or other sorts of things you might think would be the problem. You're always trying to thread the right needle with licensures and all that kind of stuff. But broadly we never had problems because we paid attention.
Final reflections
Zac Haluza: I’m kind of curious, after all the time that you’ve spent in China, is there anything that particularly stands out, like a skill or learning? One in particular that’s more helpful to you now when dealing with clients and startups [while working in Silicon Valley]. The biggest takeaway.
Steve Mushero: Good question. Obviously, for any sort of expat living, managing, building in another country, especially one that's quite a bit different than your own, learning and understanding how to do that is really challenging, but really useful. I'm still a huge fan of people working and living overseas, whether it's China — maybe not so much anymore — or it's going to Vietnam or Cambodia or Africa, it doesn't matter where. I'm just a huge fan of that. Particularly people are young, to go and build and learn stuff. It just gives you so many more skills. Even if you hate it, it still gives you all kinds of skills. And of course, your job is not to hate it, but to find a way to make it work. It's gonna be frustrating no matter what you do.
For us, it wasn't only working with Chinese teams. We had 25 different countries in our workforce. People from all over the world, really. And getting all of them to mesh together and working through all of that is just a huge skill. That's not really China-specific. I mean, “China-specific” is more than just learning about China, right? And like a lot of places, being able to accept that some of the things you can change and some things you can't.
But that's generic stuff. I wouldn't say there's anything specific per se, other than to go do it.
Zac Haluza: Yeah. And that makes sense. Even just reflecting on my own experience, in how that's shaped how I approach things in my career or life. I guess it's less of several individual things that stand out and just more of this kind of cumulative snowball effect. And I guess that’s also really related to how you grow as an individual.
Steve Mushero: Yeah, exactly.
My sincere thanks to Steve for agreeing to chat about his background in Chinese cloud and more. You can get in touch with him on LinkedIn, and you can learn more about his experiences (including past interviews) on his personal website.
Some more details on CloudEx provided by Steve: They had 100 customers in the middle of 2009 and began facing issues by August 2011. They stopped accepting new customers in September 2011.
AWS’s operations in China began at the end of 2013, when it announced the opening of its China (Beijing) region. In 2017 it opened a second Chinese region in Ningxia.
Alibaba Cloud (阿里云, or Aliyun) was founded in 2009 but only started offering public services on a large scale in 2011, following public beta testing in 2010. (According to their Baidu Baike page.)
More details: HiChina was acquired by Alibaba in 2009 and later merged with Alibaba Cloud, ultimately becoming a brand of services offered under the Alibaba Cloud umbrella.
Alibaba’s record-breaking IPO on the NYSE took place in 2014.
According to the timeline on their Baidu Baike page, Tencent Cloud opened up to users in late 2013.
Thanks for the interview!
Great stuff! That's a keeper.