A system glitch that led to the worst-ever all-day outage on the Tokyo bourse late last week was due to a setting error, the bourse said Monday and which put together a panel of outside directors to probe the cause of the malfunction.

"The system was supposed to automatically switch in case of a hardware breakdown...but it did not work when there was a memory malfunction," Yasuhiko Tamura, director of trading systems in charge of IT development at the Tokyo Stock Exchange, said in a press conference.

Photo taken Oct. 2, 2020, shows a financial data board at the Tokyo Stock Exchange. The bourse reopened normally the same day after trading in all shares was halted for the whole session the previous day due to a system glitch. (Kyodo)

After the glitch, the world's third-largest bourse following the New York Stock Exchange and the Nasdaq Stock Market in the United States revised the setting for automatic switching and confirmed it is functioning.

Such a setting is based on the entire system, and not that of memory alone, Tamura said, adding the exact cause of why the system failed to automatically switch is still being investigated.

The TSE also said Monday the panel of external directors will consider measures to ensure a similar incident does not occur again. The panel aims to compile a report as soon as possible, the TSE said.

Prime Minister Yoshihide Suga described the full-day suspension on Oct. 1, which was the first all-day halt since 1999 when the TSE's trading system was fully computerized, as "extremely regrettable."

"It's extremely important to develop an appropriate infrastructure," Suga said, as his government pushes to boost Tokyo's standing as a global financial hub and vows to make digitalization a top policy priority.

The government is aiming to accelerate such efforts at a time when the status of Hong Kong as a financial hub in Asia has been hurt by China's tightening of its grip on the semiautonomous territory.

The TSE's "arrowhead" trading system, developed by Fujitsu Ltd., was introduced in 2010 and updated in November last year.

In that month, the bourse and Fujitsu tested the system and confirmed the automatic backup system works in case of a hardware malfunction. But they did not test it by actual memory breakage, Tamura said.

After the updated system started running, the settings for the backup system were not changed and similar tests were not been conducted, Tamura said.

Fujitsu President Takahito Tokita apologized Monday for the system glitch as the developer vowed to fully investigate the cause.

The bourse said that at this point, it does not plan to seek compensation from Fujitsu.

It was not the first time the TSE has experienced a trading suspension while using Fujitsu's system.

A programming error led to suspension in the trading of all stocks for about three hours in 2005. In 2012, a server in the trading system failed but a backup was not switched on, prompting more than 200 issues to temporarily stop trading.

The bourse said last Thursday it detected a hardware defect at 7:04 a.m. and announced it would suspend all trading at 8:39 a.m. ahead of the market's opening. At 11:45 a.m., soon after failing to resume operations during the morning session, the TSE said it would halt operations for the entire day.

The bourse said it decided not to resume trading within that day as it could cause further disruption to market participants. The resumption required an entire system reboot that would affect orders already placed, the TSE said. The bourse resumed normal trading the following day.


Related coverage:

Fujitsu chief apologizes for Tokyo bourse glitch, vows probe

FOCUS: Glitch-caused trading halt threatens market trust in Tokyo bourse

Trading returns to normal at Tokyo bourse after worst glitch