Korea and Kakao: '503 service temporarily unavailable'

By Jason Lim

My elderly parents moved into an apartment condo in New Jersey a few years ago when they found that they couldn't upkeep a single-family house any longer due to increasing frailty. Before they moved in, they had some renovations done to their kitchen and bathrooms to make them modern, clean, and stylish.

Recently, however, they started smelling a foul odor coming from their bathrooms. Despite multiple tries, they couldn't diagnose the source of the problem and had to tolerate the sewage-smelling odor until the latest batch of plumbers discovered that the former contractor hadn't connected the toilet pipes correctly and that human waste had literally been piling up. In short, while the outside of the apartment looked spic and span, the underlying infrastructure was shoddy.

As I write this, Kakao services are still down due to a fire in the SK data center that many large IT companies in Korea use to house their data. While the fire was extinguished fairly quickly, it still takes time to bring the power back up.

After a few hours, the co-heads of Kakao posted a message that made an obligatory apology and stated (my translation), "…We are doing everything we can to normalize our services by leveraging another data center. Kakao backs up our data distributed across multiple data centers domestically and has a redundant system to address issues arising from externalities. Immediately after the current fire, Kakao became aware of the relevant facts and right away began our failover procedures."

Technical translation becomes tricky since Korean use of English terminologies often varies in actual meaning. For example, when Koreans use the word, "system," they might actually mean a set of processes to get things done and not an actual system per se. I also assumed that, by "redundant system," they meant an automatic failover to a backup data center in real time.

Or did they? The statement goes on to say, "However, since it's rare that the entirety of a data center is affected, it's taking longer than expected to apply the fixes." Basically, even with the most generous interpretation, Kakao is admitting that it never planned for a contingency in which its primary data center wholly loses power. It's also admitting that the "fixes" that they had planned aren't adequate to the contingency that they just experienced.

To note, data centers go down all the time. It's not a Black Swan event. It's actually very foreseeable. You would think a company with the size and reputation of Kakao would have the systems and processes in place for an immediate enterprise failover and that this data center outage would be relegated to be a minor footnote on some obscure academic journal, not front-page news in every Korean newspaper.

At the same time, even if it takes Kakao much longer to restore services than they had originally planned, it's not a huge deal. Unless, of course, you are Kakao, on which the entirety of the nation depends on to make things work.

So, a prolonged Kakao failure is not just about an app being down. It's not even similar to Facebook or Instagram going offline. As of 2021, of the 50 million Kakao users in the world, 47 million were in South Korea. Although Kakao started out with messaging app Kakaotalk, it's now so much more for so many people. Imagine WeChat falling in China, except that Kakao's relative dominance in Korea is bigger. As Juwon Park, AP's Asia entertainment editor tweeted, "Koreans rely on Kakao as an everyday messenger app, bank, cab hailer, shopping, bicycle renting, restaurant reservation service, etc. Everything stopped working today."

This means that a Kakao failure immediately becomes a critical infrastructure issue for Korea. It literally creates a single point of failure for every system, every service and every transaction. In fact, it creates a single point of failure for the vast majority of human interfacing and communications for the entirety of a nation. Sure, Korea is the most connected country in the world. But it's connected via a single point of failure.

And it failed. Consequently, Korea stopped.

I am not familiar with Korea's regulatory framework for IT companies. However, it's pretty obvious that the government must ensure that something which intimately and pervasively affects the lives of virtually every citizen of the country is resilient enough to keep going despite natural disasters, cyberattacks, fire or other physical and cyber threats.

The underlying plumbing, so to speak, the electric, gas lines, water and every system allow occupants to live healthily and happily in their own homes without having to smell the stench of their own waste. If these things don't work or are shoddily done, they condemn vulnerable populations to live without the basic infrastructure that they depend on. Worse, the weight of shiny new things that are built on top of a shoddy foundation will cause it eventually to crumble and expose the whole thing as a house of cards.

Jason Lim (jasonlim@msn.com) is a Washington, D.C.-based expert on innovation, leadership and organizational culture.