If MyGov was an eCommerce website the repetitive outages would have been cured in about six hours or less.
The probable reason the Human Services department can’t ramp up capacity to meet this anticipated demand is either a contractual or risk-based cause. There are almost no technical impediments to instantaneously scaling capacity to meet increased demand in the age of Cloud computing.
Like most of the government, the typical mid-level public servant’s primary objective is to cover their own arse. They prove this by outsourcing even the most basic decisions to consultants and the big accounting firms. Alternatively, they layer every change in onerous policy and use that as a fallback mechanism to reduce the risk to their career trajectory.
Your typical Mandarin sees their role as mitigating change-based risk. It’s hard to break systems or processes if they never change. Legislation to enhance sovereignty or privacy are also handy obfuscation tools in the public servants’ arsenal of deferral, delay or deny.
The quickest method of fixing myGov website outages is adding more load-balancing technology to the department’s technology stack. Load balancing is performed by a series of “Application Delivery Controllers”, a system that takes the incoming web requests – in this case newly unemployed Newstart applicants – then spreads the load across more Webservers. If unexpected workloads place a greater burden on limited capacity Server infrastructure, the system can quickly adapt using Cloud based “Capacity on Demand”.
The process works well in commercial ecosystems but in government environments, inherent limitations restrict how quickly this happens. The desired outcome demands a brave public service administrator willing to override the layers of outsourcing red tape in contracts with behemoths like IBM and Fujitsu. These big IT players provide this type of technical system wrapped with inflexible contracts governed by service level agreements and penalty clauses. The actual mechanics needed to add capacity on demand is easy and requires a few mouse clicks. The delays created by indecision are the limiting factor.
There’s another important thing to remember in this scenario. In the Canberra technology bubble, legacy technology still plays a massive role in service delivery. The stalwart Mainframe is still chugging away in the corner for a simple reason. Keeping it there may cost taxpayer’s $10 million per annum in support charges, but is a low-risk strategy, albeit expensive. The alternative is to replace those old boxes with state of the art technology. The catch is that the functionality required to do this involves a massive software rewrite that could cost anything up to $1 billion dollars.
This is why recent health system computer upgrades have failed dismally. These projects are incredibly complex, require skills not readily available because the experts have retired from the field and scope-creep creates project delays and budget blow-outs.
The simple solution to this dilemma is to lower expectations or outsource more services to private enterprise. Otherwise, accept the inflexibility and security that snuggling up to the public service bosom provides.
The Catch-22 is that taxpayers expect more from the public service and every investment leads to demands for more.
We need government entities to stop competing with commercial operators and the public service to provide a bare minimum of services to reduce the size of the bureaucracy.
Look no further than the ABC to see how excessive government funding can distort expectations and destroy competing commercial operators.
Mike Ryan is a curmudgeonly technical copywriter who lives in the glorious Hunter Valley.
Got something to add? Join the discussion and comment below.