Steve Mushero moved to China five years ago from Silicon Valley to consult for Chinese companies. He was hired on as chief technology officer by popular online video portal Tudou, but left to start ChinaNetCloud, a company that provides hosted cloud computing services and support to domestic and foreign-invested companies, including Nokia (NOK.NYSE, NOK1V.OMX), and popular group buying site Jigocity. He explained to CHINA ECONOMIC REVIEW what cloud computing actually is, why a good server administrator is hard to find, and why policy support for cloud computing is (mostly) unnecessary.
Q: What inspired you to found ChinaNetCloud?
A: One of the things I noticed while I was at Tudou was, we were famous and had money, but it was very hard to find operations people – people who could manage servers. We didn’t do any cloud computing at Tudou, but when I left there, I started this company – China’s first cloud computing company – which helps solve server operations problems for companies that can’t find operational people. We run the infrastructure for Chinese internet and gaming companies. We also build private clouds for companies. But our main business is actually running 24-7 operations; monitoring, management, backups, security and performance.
Q: Doesn’t China produce a lot of IT graduates? How can it be hard to find operations people?
A: Easy: All the schools turn out programmers, but almost nobody is turning out operations people. Operations is more of a trade thing than engineering and computer science, so you have to go to special schools. We interview all the graduates of the Linux schools in Shanghai – and there’s not many of them, maybe 50, compared with thousands of computer science graduates. Things are getting more complex, but half of our customers still don’t do backups because they don’t know how. We have the world’s largest internet community here and there is a serious mismatch between that and the condition of the infrastructure. The system is getting harder and harder to run because now we have e-commerce on the web. Five years ago you mostly had videos and blogs. If you lost a blog post, no big deal – but what if you lose a customer order?
Q: What about the network hardware? You’ve gone on the record saying the network is fragmented. Any improvement?
A: The backbone is developing – but the fragmentation I was referring to is not an issue of backbone, but the fact that northern China has one company running the network, China Unicom (CHU.NYSE, 600050.SH, 0762.HK), and southern China has another, China Telecom (CHA.NYSE, 0728.HK). Even within that, China Telecom is not actually one company; it’s one separate company per province or city. It’s not like they don’t want to interact, but there are territorial issues. My impression is that it is getting better; it’s now only a medium-serious problem.
Q: How’s business these days?
A: We’re doing very well. We recently attracted some investment from 500 Startups, which is run by Dave McClure. We have no competitors; we like that. But we’d like to have more players in the cloud. The more servers, the better for us. Doing what we do with these outsourced operations and running things for different people is actually very difficult. We used to be more concerned about competitors but we are less concerned all the time.
Q: Beijing is building a cloud computing center and lots of companies say they are investing it. What is cloud computing, exactly?
A: It’s very confusing. Different people mean different things when they talk about it. Not only are there different concepts, but every company wants to claim to be doing it, so whatever they are doing is cloud computing. That doesn’t help. Basically there are four different meanings. First, there is "software as a service," which is what most users think of when they think about cloud computing. That’s products like Salesforce.com or Google (GOOG.NASDAQ) Docs, where you have your software and documents in the cloud. The second type is called "platform as a service," which means you have parts of applications in the cloud. This is less common and you don’t see it in China much. For example, with Amazon (AMZN.NASDAQ) and Taobao, you can build an e-commerce site that uses their payment services. The third one is "infrastructure as a service," which is what we do. This is the most popular sort of cloud system.
Q : What exactly is that?
A: It’s a virtualized system. You are renting space in the cloud to do whatever you want to do. You don’t have to use your building or your facilities; you don’t have to buy anything. And if you don’t need the load anymore, it just goes away and you stop paying for it. Businesses like that.
Q: So what’s the fourth kind of cloud computing?
A: That refers to grid computing, used for distributed computing that requires a lot of processors.
Q: What’s your take on all these announcements by companies investing in cloud computing in China?
A: It’s not clear. Most of the investment is in research, but if you are researching cloud computing, what are you doing? The infrastructure certainly still needs a lot of work. There are lots of things to do to make cloud services faster and more widely available. But China is still in its infancy in terms of using software for driving growth. Most companies don’t even use ERP [enterprise resource planning] systems. Given the size of the economy, the amount of software used and the benefits this software delivers, its impact is still small.
Q: This sort of technology isn’t that new in developed economies. Is all this policy noise about cloud computing another case of Beijing using domestic companies to reinvent the wheel and block foreign entry?
A: No, the cloud isn’t that new. It’s not like we’re going to the moon and everything needs to be made from scratch. The main question is, will China be a competitive market, will there be foreign companies involved in developing cloud computing here? Yes, there will. Will it be regulated somehow with Chinese standards and so on? That I doubt. If it’s closed, people can’t use it anywhere else and it tends to fall behind. But the system is already moving very quickly, and standards take a long time to develop. I don’t see things being closed off. There aren’t standards for clouds in the rest of the world so it’s not clear how you would create new ones here. It’s not like wireless protocols where people have to agree on a standard. The cloud is very amorphous: How do you know if you are in a cloud or not?
Q: So what does the government need to do here?
A: I wouldn’t do anything at all. I’d let the market take care of it. Although cloud research centers do add a lot of value, so maybe some basic research support. That’s all.