In 2004 I moved to Brighton. First network administration, linguistic quality control. In 2007 I moved to Cork, Ireland: Product Support Specialist EMEA at Apple Operations Europe. In 2009, back to Brighton: project lead, Tech Support EMEA at NCsoft Europe, until 2011.
Two very different companies. Apple makes hardware and software for consumers. NCsoft makes online games for millions of concurrent users. The common denominator: both operate global systems where failure and inconsistency have directly visible consequences.
Apple and the Cost of a Standard
At Apple I learned what a standard really means. A standard with exceptions isn't a standard. It's a guideline with room to negotiate, and therefore worthless when it matters.
In EMEA support, that meant: every process documented, every escalation justified, every deviation traceable. Apple isn't a bureaucratic company. But a support system with ten million customers and twenty languages only holds together if every decision is consistent. An inconsistent decision in Dublin creates an expectation in Munich that becomes an exception in Amsterdam.
I saw how quickly a system tips over once exceptions become normal. And how stable it stands as long as exceptions don't happen.
„A standard with exceptions is a guideline. A guideline with exceptions is a recommendation. A recommendation solves nothing when it matters."
Learned at Apple Operations Europe, Cork 2007
NCsoft and Global Systems Under Load
NCsoft is a Korean company that operates online games for Europe. As project lead for Tech Support EMEA, I coordinated support infrastructure for titles like Guild Wars: millions of concurrent users, multiple languages, different time zones, full availability even during peak play.
That taught a different dimension of system reliability: not the consistency of a single process, but the resilience of a distributed system under real-world conditions.
When a patch introduces a bug, players in Germany, France, and Poland feel it at the same time. And they expect a response that makes sense to all three. The system has to manage multiple truths at once, and stay consistent while doing it.
| Criterion | Corporate Definition | Web Infrastructure Reality |
|---|---|---|
| Consistency | The same answer in DE, FR, PL, at the same time | HTTP headers, markup, DNS, identical across every region |
| Reliability | Predictable outcome, regardless of timing | A crawler gets the same result today as in three months |
| Transparency | Every decision documented and traceable | Structured markup matches actual content |
| Fault tolerance | One node failing doesn't break the whole system | A missing header doesn't block the whole scan |
| Recognizability | The system identifies itself clearly and correctly | Domain identity matches across every layer |
The common denominator: Both Apple and NCsoft operate systems where inconsistency becomes directly visible, and has direct consequences. The system delivers completely, or it counts as down. Fully correct is the only accepted state.
The Transfer
I spent four years working in structures that treat reliability as a design principle. Then I moved into the field that does that the least: web infrastructure.
Most websites grow out of layers added one after another: each with its own logic, its own priorities, its own blind spots. The result is a system that hides its own overall picture, because no framework has read all the layers at once until now.
SOVP is that framework. 265+ parameters, six clusters, binary result. The same standard Apple applies to every support process, transferred to web infrastructure. And the same question NCsoft asks of every system component: does it deliver under load? Does it deliver consistently? Does it deliver the same thing in every region?
„Four years inside systems that treat reliability as a design principle sharpened a way of seeing that has stayed precise ever since."
„England gave me a different way of thinking: more direct, more pragmatic, more results-driven."
On the years in Brighton and Cork