On 24 February 2021, the National Stock Exchange, India's largest exchange, had a problem that had nothing to do with matching buyers and sellers. NSE had multiple telecom links from two different providers. That morning, NSE said, both providers reported instability on all their links. The trading system itself kept working. But the online risk management system, run by NSE's clearing arm NSE Clearing Ltd (NCL), was affected, even though it too was set up in a "high availability" mode. A risk system like this checks trades in real time, so that nobody takes on positions they can't cover. NSE said that without it, the market "could not continue normally".
So at 11:40 am, NSE halted trading in all segments. The rival exchange BSE kept running normally. NSE reopened at 3:45 pm, and the trading day was extended to 5 pm at NSE, BSE and a third exchange. All told, NSE was shut for about four hours.
The regulator, SEBI, asked NSE for a detailed root-cause analysis, and for an explanation of why trading had not moved to its disaster recovery site, a second setup meant for exactly this. SEBI's show-cause notice that August alleged that NSE and NCL had not ensured their systems were ready to move to that site, and had not recovered within their four-hour recovery time objective, the target for how quickly operations must be restored.
In June 2023 NSE and NCL settled the case by paying about ₹72.6 crore, without admitting or denying the findings.
- 2 telecom providers, and all links unstable
- Trading halted 11:40 am, resumed 3:45 pm: about four hours
- A four-hour recovery time objective, which SEBI said was not met
- About ₹72.6 crore settlement: NSE about ₹49.8 crore, NCL about ₹22.9 crore
Why it matters: Redundancy only helps if the backups don't fail together, and a disaster recovery site only helps if you have tested moving to it. Any system the main path can't run without needs the same protection as the main path.