Last year, Alibaba turned what Business Insider’s Jeremy Burke called an “obscure, made-up Chinese holiday” into a shopping event that surpassed even Black Friday. On China’s Nov. 11 “Singles’ Day”, the company raked in $14.3B! Expectations are high for 2016, and Alibaba and other retailers are no doubt gearing up for the marketing blitz of the century. But sometimes success comes with its own set of challenges, and in e-commerce it is often the case that it’s easier to get masses of people to your site than it is to complete their shopping cart checkouts once they arrive. E-commerce sites frequently experience performance problems when tasked with concurrently transacting the astronomical number of purchases experienced during events like Singles’ Day.
If you build it, they will come, but will they be able to check out?
Amazon encountered these performance problems with its most recent Prime Day. While the second annual event exceeded the original by more than 60 percent worldwide, it was also marred by technical problems. As Jason Ankeny reported, “Many Amazon Prime subscribers hoping to score Prime Day deals and discounts instead took to Twitter to complain about “add to cart fail” messages generated when they attempted to purchase sale-priced items. Some added the hashtag “#PrimeDayFail” to their tweets.” Amazon made a killing, to be sure, but it leaves one to wonder how much more dough would they have raked in if everything had worked like it was supposed to?
Will Alibaba do it better?
So, the world will watch with great interest to see if retailers serving the Chinese market learn from Amazon’s technical missteps and deliver a seamless customer experience that fully capitalizes on all of the traffic and interest they’re likely to get this year. Can they do it better?
No doubt they’re feverishly preparing for an unprecedented onslaught of traffic. However, much of what was experienced in the U.S. may not necessarily stem from lack of preparation, but instead misplaced reliance on an underlying database technology that simply isn’t built to scale to the Prime Day/Singles Day levels of traffic.
Watching helplessly as profits get sucked into a MySQL vortex
MySQL databases, which power the majority of online retail sites, are designed to run on a “single box”, and scaling them out in cloud computing fashion is quite a complicated and risky feat of software manipulation. Put simply, it doesn’t matter how fancy your car is if the engine doesn’t have the horsepower to move it up the hill. MySQL was designed for another generation of applications, and no matter how skilled your team is at pushing it beyond its natural limits, those who rely on it are bound to experience performance problems at exactly the time when they stand to benefit most from heavy consumer traffic.
AWS of course has massive resources– if anyone can scale a relational database, it ought to be them. But the fact that even they had trouble illustrates the fundamental issue that online retailers face in regards to MySQL scalability.
Alibaba’s best shot for ensuring a seamless customer experience
Scale-out SQL, the most economical and currently viable approach to these performance problems, is no pipe dream and is in fact used by many of the world’s top e-commerce companies, such as Japan’s Rakuten. With the ability to scale indefinitely to performance levels by merely adding server nodes (in true cloud-computing fashion), combined with the ACID compliance of a relational database, these companies are able to handle massive swings in traffic without a hiccup. Combined with in-memory capabilities, modern scale-out SQL technology affords enough speed to outpace any traffic demands that are likely to be experienced.
What’s under the hood of Alibaba? We can only guess, but the site performance as millions of users attempt to check out simultaneously will be a top indicator of whether or not they’re using scale-out SQL, or old-school single-box MySQL. When a SQL database is architected to scale out (and not just up), it ought to be able to handle not just massive traffic, but massive concurrent transactions– and concurrency is key when you’re talking about what may turn out to be the world’s biggest online shopping day. After all, millions of shoppers aren’t going to “wait in line” to check out. They’ll expect to make their purchases immediately; and if they don’t they’ll probably abandon their cart.