Publicação
Understanding and evaluating the Behaviour of DNS resolvers
| Resumo: | The Domain Name System is a core service of the Internet, as every computer relies on it to translate names into IP addresses, which are then utilised to communicate with each other. In order to translate the names into IP addresses, computers resort to a special server, called a resolver. A resolver is a special DNS server that knows the DNS structure and is able to navigate the huge number of DNS servers in order to find the final answer to a query. It is important for a resolver to be able to deliver the final answer as quickly as possible, to have the smallest impact on user experienced latency. Since there is a very large amount of domains and servers, and the system is highly replicated, there has to be some logic as to how a resolver selects which server to query. This brings us to the problem we will study in this thesis: how do resolvers select which DNS server to contact? If a resolver always selects the best DNS server - the one that will be able to provide the answer to the query the fastest - then resolvers can more quickly answer their clients, and thus speed up the Internet. However, if they contact different, more or less equivalent, servers they could contribute to load balancing. To understand how exactly the resolvers select the DNS servers to contact, we conducted an experimental study, where we analysed different resolvers and evaluated how they select the servers. We base the structure and parameters of our study in previous research that has been conducted on the topic, which shows that resolvers tend to use the latency of its queries to the servers as a means of selecting which server to contact. |
|---|---|
| Autores principais: | Cruz, Manuel Duarte Ribeiro da |
| Assunto: | Domain Name System Resolver Server Selection |
| Ano: | 2018 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade Nova de Lisboa |
| Idioma: | inglês |
| Origem: | Repositório Institucional da UNL |
| Resumo: | The Domain Name System is a core service of the Internet, as every computer relies on it to translate names into IP addresses, which are then utilised to communicate with each other. In order to translate the names into IP addresses, computers resort to a special server, called a resolver. A resolver is a special DNS server that knows the DNS structure and is able to navigate the huge number of DNS servers in order to find the final answer to a query. It is important for a resolver to be able to deliver the final answer as quickly as possible, to have the smallest impact on user experienced latency. Since there is a very large amount of domains and servers, and the system is highly replicated, there has to be some logic as to how a resolver selects which server to query. This brings us to the problem we will study in this thesis: how do resolvers select which DNS server to contact? If a resolver always selects the best DNS server - the one that will be able to provide the answer to the query the fastest - then resolvers can more quickly answer their clients, and thus speed up the Internet. However, if they contact different, more or less equivalent, servers they could contribute to load balancing. To understand how exactly the resolvers select the DNS servers to contact, we conducted an experimental study, where we analysed different resolvers and evaluated how they select the servers. We base the structure and parameters of our study in previous research that has been conducted on the topic, which shows that resolvers tend to use the latency of its queries to the servers as a means of selecting which server to contact. |
|---|