The debate about how a tracing app should work is about this one question. But what is the difference between a centralized and a decentralized version?
By Kristin Becker and Christian Feld, ARD Capital Studio
The hope is that the tracing app could help determine who was near a corona infected person. What health offices have worked on with pen and paper so far should be done better and faster digitally.
But how should such an app work? The key controversy is: Centralized or decentralized solution? It is an argument that is difficult to understand without knowledge of IT security and cryptography.
The basic principle is the same: If two mobile phones are close enough for a certain time with the app, they exchange so-called pseudonymous IDs via Bluetooth. These change regularly and do not allow any direct conclusions to be drawn about which specific person it is.
With or without “secrets”?
The difference: With the decentralized model, a user only sends his own IDs to the server in the event of an infection. All other app users can download it from there. The actual check whether there was a contact is only carried out locally on the cell phone.
In the central model, on the other hand, an infected person’s app also sends the codes of the contacts to the server. There is also a network of contacts there. This is sensitive information, quasi “secrets” that need special protection.
– .
