Sidebar to Feature:
Paymark, which processes 80% of New Zealand’s EFTPOS payments, claims its reliability has helped drive our EFTPOS use to the highest in the world.
Since 1989, the company, owned by four leading Australian and New Zealand banks — ASB, Bank of New Zealand, Westpac and ANZ National — claims to have processed more than six billion transactions.
Paymark has an Auckland CBD office with processing centres in Albany and Mt Wellington. These sites are also linked to 65,000 merchants and 80,000 checkout terminals.
The system works through a web-facing and switch solution connecting through third parties such as banks, financial institutions and retailers.
Systems programmer Stuart Hogan says the company has a high availability (HA) design, a good reflective quality assurance environment and extremely robust processes.
This means there is much redundant hardware at sites, such as network gear having two routers, two switches, two servers and so on.
The processing centres are also designed that if one site fails totally, the main/core services can be carried out at the other site. Some smaller services must still be carried out at one site, but as their demand increases the case for extra DR/HA capability increases.
Paymark is guarded about the brands and some of the systems it uses, citing commercial confidentiality, but claims to meet and exceed industry standards and the demands set by the finance sector.
The company has a dedicated 24/7, 365 days-a-year operations team that monitors all services and which acts as the initial escalation point should any incidents arise.
The team includes two network specialists, two e-commerce specialists, two core switch specialists and non-stop hardware systems programmers like Hogan.
In addition, on-call support technicians who are specialised in platforms, applications and network issues are also available.
Paymark stresses the use of industry best practices, such as off-site storage, and researches the latest industry systems and white papers, etcetera, to further improve reliability.
Hogan claims Paymark experiences few outages, and these are usually caused by configuration of software, rather than hardware, resulting in an “always there service” to customers.
Indeed, it is the customers, and integrating with them when issues arise — such as integrating their systems and their various releases of Java, .Net, etcetera — that help here.
“Given the fact they are shared, certain events/situations are beyond our control — such as hosting and outsourcing. [So], we need to have a clear understanding of the integration issues of customers and, as such, a part of the support staff’s work is to troubleshoot customer sites. Automated threshold monitoring and tools greatly assist in reducing incidents,” he says.
Indeed, transaction flows are modeled and this information is fed into Paymark’s capacity management programme to ensure resources are always available.
Hogan says 24/7 operation is now ingrained in the company psyche and new staff have to be quickly brought up to speed. New technology is always an issue, but skilled staff can mitigate issues.
“Even the most robust design can be hit by catastrophic events beyond our control and, as such, we have a network of strategies and processes in place to ensure any service is returned to normality as quickly as possible. These include risk planning, escalation procedures, vendor maintenance agreements, security testing and close inter-departmental liaison,” he says.
Paymark is currently upgrading its systems, implementing a public-facing (in-bound transaction) IP infrastructure, to replace the existing X25 network feeds as such technology is now unsupported and will be switched off by the telcos. It has also migrated the existing point solution IP feeds into this core gateway.
Hogan says the project has been delivered on time, with customers being unaffected and continuity of service ensured by Paymark’s HA and DR design.
“Thorough QA, load stress and destructive testing right through to the core transactional switch has provided a high level of comfort and confidence in the design, and ‘fit for use’ of the infrastructure,” he adds.