by clicking the arrows at the side of the page, or by using the toolbar.
by clicking anywhere on the page.
by dragging the page around when zoomed in.
by clicking anywhere on the page when zoomed in.
web sites or send emails by clicking on hyperlinks.
Email this page to a friend
Search this issue
Index - jump to page or section
Archive - view past issues
Connexus : Issue 38
Avoiding downtime DiSAStERS C orrupted files recently caused payroll and transaction delays for thousands of customers at NAB. Air-conditioning problems at Westpac dragged down merchant ATM, EFTPOS and online banking networks. Disk failure at ANZ resulted in internet banking outages. Routine maintenance at CBA forced the bank to run ATM machines in offline mode. Planning and preparing for when things go wrong is the way to avoid such downtime disasters, says Barry Seto, general manager of support at Ultradata. When Westpac’s air-conditioning failed, it should have had a backup system ready to go. “Is regular maintenance conducted? Is the unit past its usable life? If it fails without reason and without backup, is a high-priority service contract in place?” asks Seto. He says ANZ’s disk failure is a lesson that robust architecture – be it a server, data line, disk device or entire data centre – is not cheap, but it’s important. “ANZ’s disk failure seems to be a single point of failure that brought down the system. Architecture should be designed to minimise the number of potential points of failure and, if your primary link to the outside world fails, you should have a secondary link to fall back on. For example, you don’t want a single data link for all ATM connections.” Some activities at data centres are mundane and repetitive, but their technology teams need to stay focused, says Branko Panich, partner and lead technology adviser for the financial services industry at Deloitte Touche Tohmatsu. “In this environment, a small error can cause huge problems. A simple mistake can impact on a lot of customers,” he says. It’s also critical to have accurate information about what’s going on in the operations area. “System performance needs to be tested and monitored,” says Panich. “How does the system react under different network configurations? As system performance slows down, metrics are required to help diagnose the problem quickly and resolve the issue before it escalates into a full outage.” If an outage does occur, customers need to be informed quickly about what’s happening. “Typically, the focus is on getting the system up as quickly as possible,” says Panich, “but ignoring the customers means they are likely to get even more annoyed.” – Cynthia Karena is a freelance writer. With a spate of outages causing havoc for the big four, avoiding similar high-profile crises is top of the list for mutuals. By Cynthia Karena TECHNOLOGY 56 Connexus www.abacus.org.au “...everyone knows what to expect when things go wrong and it is clear what steps need to be taken in a worst-case scenario.” David Rowe, Hume Building Society