OpenTLD is actively involved in the research and development efforts of EDNS as it allows deeper data analysis and further optimization of its services.
EDNS allows Content Delievery Networks (CDNs) to receive more information about the type and attributes of devices requesting DNS data, instead of retaining this information at resovler level.
While still under development, the technology could allow useful traffic and behaviour data to be aggregated and promises to add desirable features to the normal attributes of DNS. It is described in RFC2671 and RFC6891.
In order to categorize DNS resolving information, OpenTLD recognizes three different classes:
Class 1 – This is traffic from devices that use OpenTLD’s resolvers. The information from these devices is usually the most reliable becayse OpenTLD provides the resolving and can detect which device which wants to resolving which domain.
Class 2 – This is traffic that comes through other resolving parties, but where OpenTLD has access to “EDNS client subnet ” information. It is less accurate than Class 1 traffic because only the subnet of the device is visible.
Class 3 – This is traffic from unknown resolving parties and the most difficult to estimate.
Hundreds of millions of DNS requests go through OpenTLD’s Cloud Anycast DNS each day, and this massive data is parsed and organized in real time. This allows further optimization of the information from web traffic, email traffic, etc.
After obtaining the data, it must also be analysed in real time, which requires significant computing power.