Single Gateway Failure
System failure can occur at various points within the trading environment. These points include the HKEx Gateway and the Exchange host. This chapter’s discussion is restricted to total software and/or hardware failure and available options for disaster recovery.
When the customer has only one HKEx Gateway, and it fails, all traders’ orders behave as configured with the Exchange through the Order API.
When setting up their membership with the Exchange, clients pre-configure order behavior during a Gateway failure. When a single HKEx Gateway fails during trading hours, the Exchange either:
- Places all orders in an inactive state and deletes them at the end of the trading day
- Keeps orders working in the market
Clients pre-configure this behavior with the Exchange. This configuration occurs independent of the HKEx Gateway. In both cases, traders stop receiving data (prices, fills, order data) from the HKEx Gateway.
Gateway Software Recovery
After a failure, TTChron restarts any components scheduled to be running. After a successful restart of the Gateway, the following activities occur:
- Fills: From the Exchange, the HKEx Gateway requests and downloads all successful fills. The gateway does not download any fill information that has already been received from the Exchange.
- Orders: Upon reconnect, traders’ can resume trading. The HKEx Gateway downloads
all order information from the Exchange. The Order Book will be correctly
populated: all working orders are displayed as “working;” all inactive
orders are displayed as “on hold.”
The Order Server synthetically keeps all inactive orders “on hold” upon disconnect.
Gateway Hardware Recovery
HKEx Gateways that suffer hardware failure are unable to go back online until the situation is rectified via replacement of the affected components.
When failed hardware is replaced or corrected, and the system is restarted, recovery processes are identical to those listed in Gateway Software Recovery.