He was Alipay’s very first intern, and now the CTO of Ant. This is his story.
Recently, Ant Financial Chief Technology Officer Cheng Li (nickname Lusu) met with a group of interns at the company’s Hangzhou headquarters, where he spoke of his experience as Alipay’s very first trainee before it launched in 2004. These are excerpts of his comments.
I was Alipay’s very first intern back in 2004. Given the current requirements of the company, I probably would not have made it in today.
But back then, I was one of just four engineers put on a project to spin Alipay out into a third-party payments platform. Even though I was an intern, I was made chief architect of the project because of my previous experience with (e-commerce website) Taobao.com, and because of everyone’s trust in me.
Our plan was to go “live” in May, before other competitors could enter the market. We promised Teacher Ma (Alibaba founder Jack Ma) that we would meet that deadline, but in mid-April, some colleagues raised the concern that the system had serious problems.
My supervisor backed me up, and that made me feel really touched. But I was still uneasy. So I took another close look at the code and concluded that my colleagues were right to raise their concerns. As I reflected on the situation, I realised I had relied too much on unproven technology to use in so critical a system.
I decided to pulled an all-nighter to try and devise a new system from scratch. The next morning, we gathered to discuss our next steps. Instead of blaming me for my error, everyone voiced support for the new architecture. We were all determined that the project should succeed.
We successfully launched the new system in May, and that also marked the end of my internship. In July, I officially joined the company as an employee.
Alipay’s early development was meteoric and the system had to continually improve to keep up with surging transaction volumes. We coded day and night. Then in 2007, we decided to entirely rebuild the system we had created just two years earlier, because it was the only way to support Alipay’s future growth.
This total refactoring was one of the most famous projects in Alipay’s history. But we committed a big taboo by using the latest — hence unproven — technology. We wrote all of the middleware ourselves, which meant hundreds of millions of lines of data accumulated over the years had to be reformatted.
As the launch date approached, we were still not confident that the system was ready. In the end, a decision was made to launch at the stroke of midnight on January 22, come what may.
But when the appointed time came, we realised the system still couldn’t be launched because the accounts could not be balanced. This would be a disastrous error if the system went live. By the time we found the root cause of the problem, it was already almost 3 p.m. that afternoon.
There were two unforeseen scenarios when we finally launched the new system. Firstly, because user and merchant demand had been suppressed for a whole day amid the upgrade, the pent-up orders overwhelmed the system. Second, the system kept displaying transaction errors.
I remember continually rebooting and troubleshooting the system, and the whole time, there was a voice in my head repeating: “We’re doomed. We’re doomed.”
But the system gradually stabilised as it worked through the backlog and fulfilled the merchant orders. The errors turned out to be false alarms and procedural in nature, rather than systemic.
Looking back, I think we were really courageous. We made a lot of poor technical decisions on hindsight. But we pushed through no matter what, and because of that, we laid the foundations for Alipay’s development.
The year 2009 brought us another major turning point with the launch of November 11 Singles’ Day (Alibaba’s annual online shopping festival). By 2010, in its second year, transaction volumes surged multiple times from the start of the festival at the stroke of midnight.
No one dared rest that night, and by 8 a.m. in the morning, we knew it was going to be a rough day because our system was simply not designed to handle such volumes.
So we put all of our computing resources where they were needed for payments, and disabled non-essential functions to free up our servers.
After that challenge came another gigantic one: moving our technology architecture onto the cloud. This took almost three years of efforts, and by May 2013 we were fully on the cloud.
At that moment, I felt this team had a very clear idea of our direction for the future: to build our very own platform and also build our own core database.
No platform in the world has had to deal with the volumes that we do, and that is why we need to create our own technology solutions. And this technology is our dream, the dream that we will go “all in” to achieve.
From 2009 to 2019, after much perseverance, we finally developed our own backend infrastructure capable of supporting all of our businesses, and that of many other companies.
Ant Financial is what we are today, because of our trust in the young. We give them a very good stage, allowing each young person to grow and shape the future.
And also because of our courage. Today Alipay still faces a lot of challenges, but as long as we band together, as long as we have courage, I believe we can conquer these challenges.
I’m confident, because we have already created many miracles in this process. And I look forward to more courageous young people with dreams, to join us in shaping the future.