Usually, IP geolocation service providers solely rely on evidence or hint data for their geolocation estimation. The data could come from various sources, such as RIR registration data, user-submitted data, geofeed data, triangulation estimated data etc.
The problem with the method is that in most cases, regardless of how compelling and accurate the data is, it can only reflect a single point in time. As explained in our How accurate can IP Geolocation get? blog post, not all IP addresses are assigned to static devices. Therefore, even if valid, a single location is unlikely to represent the most common location for an IP address deployment.
Moreover, it is difficult to verify or validate contradictory data coming from multiple sources. This is probably the main reason why many IP Geolocation providers are still using a human-based validation process, which unfortunately does not necessarily improve the outcome.
BigDataCloud utilises a unique patent-pending technology for network routers service area estimation (confidence area). This technology allows for an extra automatic step verification, making the resulting IP Geolocation outcome much more accurate and reliable.
In a nutshell, the BigDataCloud IP geolocation estimation is an outcome of two almost entirely independent and parallel processes. One is to determine a confidence area where the IP address could be deployed if assigned dynamically, and another is to collect field evidence data about its actual locations.
Then, at the final stage, both outcomes are confronted. The ‘confidence’ value is determined to reflect the overall data quality measure, which may represent the IP geolocation quality.
It is to indicate whether at least one of the processes did not come up with a deterministic value or the outcomes were not mutually supportive.
Is when both outcomes are deterministic and mutually supportive, but at least one of them is not of excellent quality.
Both outcomes are of good quality and mutually supportive.
Please note that the low confidence value does not necessarily reflect the IP Geolocation quality. It can still be very accurate. However, the medium-high values are likely to indicate a higher accuracy or more reliable IP Geolocation estimation.