Back to the archive

The encyclopedia · Engineering & Operations · Technical decision · 2022

Kakao ran South Korea's messaging on one data center — a fire took the country offline

A fire at SK C&C's Pangyo data center took KakaoTalk and Kakao Pay down for nearly a day. Kakao had no backup; Naver, in the same building, was back in hours.

Kakao · 2022

What happened

Kakao runs much of South Korea's daily life online. Its KakaoTalk messenger had 43 million monthly active users in a country of 51 million, and its Kakao Pay and Kakao T apps handled payments and taxi-hailing. Much of that ran through a single data center: an SK C&C facility in Pangyo, just south of Seoul, which it shared with rival Naver.

On the afternoon of 15 October 2022, a fire broke out at the SK C&C building. Power was cut for safety, and the fire took about eight hours to extinguish. KakaoTalk, Kakao Pay and Kakao T went down and stayed down for nearly a day, disrupting payments, online transactions and taxi-hailing across the country — affecting, as one report put it, nearly all sectors, even finance and transportation.

The fire was the trigger, but the outage was the architecture. Experts blamed Kakao for having no disaster-recovery plan or backup site for the country's biggest messenger; had it kept a hot site, they said, the disruption would have been resolved quickly. The proof was next door: Naver, in the same building, had an off-site backup and restored its services within hours. Kakao said it had not expected the power outage and that replicating so much data took time; it announced it was building its own data center in Ansan.

The contrast made the lesson unmistakable. Two companies in the same building, the same fire — and one was back in hours while the other was down for a day. The difference was a backup site one had invested in and the other had not. For a service a country depends on, redundancy is not an expense to defer; it is the price of being essential.

Why it happened

  • Kakao concentrated an essential national service in one data center with no disaster-recovery site, so a single failure took down messaging, payments and transport at once
  • The fire was the trigger, but the day-long outage was the result of having no backup; Naver, in the same building, had an off-site backup and recovered in hours
  • A service used by 43 million of 51 million people is infrastructure; an outage is a national disruption, not a technical hiccup, and the operator answers for it
  • Disaster recovery looks like an unnecessary cost until the day it is needed; Kakao's outage was that day, and building its own data center was the admission
What it costa country's services down for nearly a daycostly

The lesson

Redundancy is the price of being essential. Kakao ran a country's messaging on one data center with no backup; a fire cut power and it was down a day. Naver had a backup and recovered in hours.

Sources

spotted an error? The club wants to know.

Comments · 0

    Sign in to join the comments.

    More like this

    Somewhere, someone solved the problem this company failed at. 2nd Opinion →