The “true cause” of the large-scale failure of the Zengin system revealed – Zengin Net and NTT Data announced – CNET Japan

2023-12-01 09:18:00

On December 1st, the National Bank Fund Settlement Network (Zengin-Net) and NTT Data revealed the true cause of the large-scale outage of the Zengin system that occurred between October 10th and 11th.

The Zengin System is a system that processes daily transfers and remittances in real time, and is used by almost all deposit-taking financial institutions in Japan. In October, 10 banks, including Bank of Mitsubishi UFJ and Resona Bank, experienced system failures that continued for two full days, including the inability to transfer funds to other banks.

The failure occurred immediately after the Zengin System’s relay computer was replaced with a new model, the RC23 series, and commercial operations began. This is because the “index table for processing interbank fees” in the RC23 series was corrupted, and an error occurred when referencing the table.

One relay computer was installed in Tokyo and one in Osaka for redundancy, but because both computers were switched to the new RC23 series at the same time, software failures occurred on both computers. It failed to fulfill its role as redundancy.

Note that the corrupted index table was to be expanded from the load file that holds the table’s initial settings when the relay computer was started. According to the announcement, there was a problem in the table creation process of the program that generates this load file. Specifically, there was not enough work space to temporarily secure in memory during expansion.

Why did the shortage of work space occur? According to the announcement, in the development of the RC23 series, the size of one of the four tables used when generating load files was expanded due to the OS version upgrade.

Note that the load file generation program was designed to expand four tables at once in a temporarily allocated area. However, in the manufacturing process of NTT Data’s development process, it was mistakenly assumed that each table would be developed individually, and the temporarily secured work area was not expanded.

In addition, a re-examination of the revised content by manufacturing experts within NTT Data failed to point out the need for expansion. As a result, a large-scale failure occurred.

1701439693
#true #largescale #failure #Zengin #system #revealed #Zengin #Net #NTT #Data #announced #CNET #Japan

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.