Stop Making Lecturers Upload the Same Thing Twice

This post is about a bureaucratic mess, and how a simple architecture idea could fix it.

The Bureaucratic Mess

In Indonesia, lecturers do not just teach. They also report. A lot. The government's higher education directorate (DGHEI) required lecturers to upload their academic activities, CVs, and portfolios to multiple separate systems: SERDOS for certification, SIPKD for career development, FORLAP for institutional data, and STUDI DIKTI for scholarship recipients.

Here is the absurd part: these systems asked for almost the same data. Your CV. Your portfolio. Your academic activities. Lecturers uploaded the same documents over and over into different websites.

When the Server Gave Up

With over 10,000 lecturers nationwide hitting the same central servers, the systems buckled. SIPKD crashed within its first month. The government's solution was almost comical: they scheduled access by academic rank. Professors could upload on Monday and Tuesday. Mid-rank lecturers on Wednesday and Thursday. Junior lecturers got the weekend.

Think about that. A web system so overloaded that the fix was to tell people which days of the week they were allowed to use it.

The Idea: Flip the Data Flow

The usual engineering answer to overload is load balancing: throw more servers at the problem. I proposed something different. Instead of every lecturer pushing data to one central system, let each university hold its own lecturers' data and expose it through a REST web service. The central system then pulls the data when it needs it.

I called it redefining the data provider. The lecturer maintains their data once, at their home institution. The institution becomes the data provider. The government (and even private companies scouting for research grants) becomes a data customer that fetches data on demand.

No more uploading the same CV five times. No more central server meltdowns. The load naturally distributes across hundreds of universities instead of concentrating on one point.

Why It Still Resonates

This was 2014, but the pattern is timeless. Centralized systems become bottlenecks. Distributed data with standard interfaces (REST) scales naturally and respects the source of truth. It is the same principle behind federated systems, data mesh, and modern API-first architectures.

Sometimes the best fix for an overloaded system is not more power. It is a better shape.

Read the Full Paper

Read the full paper: Redefining Data Provider: The REST Approach To Solve Indonesia Lecturer Administrative Problems →

Published at IEEE TALE 2014, Wellington, New Zealand. Part of my doctoral research at Kumamoto University.