Yesterday's trading halt at the Sydney Future's Exchange (SFE) has again called into question the state of disaster recovery and business continuity planning among local enterprises.
The SFE issued a public notice yesterday stating all SFE markets would be stopped at 12.45pm to address "an urgent hardware rectification" unrelated to trading volumes.
"Advanced notice is provided to all users in order that this unscheduled market pause is managed as effectively as possible," according to the SFE.
The expected trading downtime was between 45 and 75 minutes, but media reports indicate it was even longer.
When the ASX merged with the SFE mid last year, managing director and CEO Robert Elstone pledged to make sure the customers don't suffer any disruption to service levels or the quality of service generally.
IBRS senior analyst Kevin McIssac said he is "really surprised" at the downtime and although the ASX claims trading volume was not responsible, he believes it is possible a secondary system was unable to cope with the present high trading volumes.
"They may have a failover strategy, but with the high volumes yesterday that might not have been sufficient," McIsaac said, adding bringing the system back in an hour was good.
"Even if you do have a secondary site doing anything in less than 10 minutes is unusual. How could you have a hardware failure on such a system that is designed for high-availability?"
McIssac said he wouldn't draw a relationship between Australia's lack of preparedness for disaster recovery and stock markets which are designed to be highly available.
According to a HP media release dated January 18, 2001, the SFE purchased an array of Compaq AlphaServer DS20E systems running OpenVMS to host its new allocation, clearing and settlement application.
The systems are attached to a Fibre Channel SAN and run the Oracle database.
ASX public relations manager Kerrin Oshry would not disclose the type of computer systems or software running the SFE's trading platform.
"The only thing we have advised the market is that it was hardware related," Oshry said.
If the AlphaServers are still in use, a hardware failure is not a good sign for HP, which prides itself on having more than 100 of the world's stock exchanges rely on either AlphaServer or NonStop systems.
The SFE markets recommenced without a "pre-open" and traders had an opportunity to cancel orders prior to the resumption of trading.