During moments of extreme activity like 26 November 2020, we receive and process thousands of updates per second, predominantly generated by our (option) market makers, the number of which is growing, resulting in high quality and competitive options orderbooks. Furthermore due to expanded expiries tradable and the underlying’s moves, we have more option strikes vs ever before.
The growing platform activity, reduction in order latencies, market volatility and matched volumes in combination with the new and much faster events system (market data system improvements implemented 10 November 2020) has unfortunately lead to Mnesia issues, our Erlang database, causing unscheduled downtime twice. Especially during moments of high volatility, the platform needs to be robust and reliable, so these issues were the utmost priority for us and we have made long term changes to ensure this won’t happen again.
Since the incidents we have implemented 4 hotfixes and also proposed two optimisation changes in the Erlang code itself to the Erlang Team. We are convinced we have now mitigated the causes of both issues. We sincerely apologize for the inconvenience this has caused.