Comparison of LDP and global privacy GDP
| Aspect | LDP | GDP |
|---|---|---|
| Data collection | Each participant perturbs their data before sending it to the aggregator | Data collector introduces noise into the aggregated data before releasing it to a third party |
| Privacy guarantee | Provides better privacy guarantee by perturbing individual data points before aggregation | Privacy protection relies on noise introduced by the data collector, ensuring privacy from third-party attacks |
| Utility | Significant utility loss due to perturbation of individual data points, especially with a large number of participants | Improved privacy/utility trade-offs compared to LDP, particularly in scenarios with a large number of participants |
| Performance | May face convergence difficulties with a limited number of participants | Requires a large number of participants for satisfactory performance, making it unsuitable for applications with a limited number of participants |
| Trust assumption | Does not require a trusted aggregator; suitable for scenarios with untrusted data collectors | Assumes a trusted aggregator, which may not be practical in distributed contexts and introduces a single point of failure |
| Aspect | LDP | GDP |
|---|---|---|
| Data collection | Each participant perturbs their data before sending it to the aggregator | Data collector introduces noise into the aggregated data before releasing it to a third party |
| Privacy guarantee | Provides better privacy guarantee by perturbing individual data points before aggregation | Privacy protection relies on noise introduced by the data collector, ensuring privacy from third-party attacks |
| Utility | Significant utility loss due to perturbation of individual data points, especially with a large number of participants | Improved privacy/utility trade-offs compared to LDP, particularly in scenarios with a large number of participants |
| Performance | May face convergence difficulties with a limited number of participants | Requires a large number of participants for satisfactory performance, making it unsuitable for applications with a limited number of participants |
| Trust assumption | Does not require a trusted aggregator; suitable for scenarios with untrusted data collectors | Assumes a trusted aggregator, which may not be practical in distributed contexts and introduces a single point of failure |
Sharing content requires targeting cookies to be enabled. Please update your cookie preferences to use this feature.