Multi-Master DNS

DNS is a foundational element of IP communications. To communicate over an IP network, an IP device needs to send IP packets to the intended destination; and each IP packet header requires both source and destination IP addresses. DNS provides the translation from a user-entered named destination, e.g., web site www address, to its IP address such that the sending device may populate the destination IP address with the address corresponding to the entered domain name. Thus, DNS provides the means for improved usability of IP applications by insulating end users from typing IP addresses directly into applications like web browsers and enabling web servers to serve web pages compromised of diverse linked content.

Given its criticality in network communications, it’s crucial that DNS services not only provide accurate information but that they be highly available such that DNS resolution services are available at all times. DNS administrators are responsible for the accuracy of the DNS data published in their DNS servers. This data corresponds to a particular DNS domain within an internal or the Internet domain tree for which the administrator is authoritative. Resolvers internally or on the Internet will seek resolution data for this domain from servers to which you’ve published this authoritative data. A major challenge arises as to how a given DNS data change from an accepted source is propagated to all DNS servers that are authoritative for that DNS data (i.e., configured with the corresponding zone data) as quickly and reliably as possible. Changes to DNS data can include additions, modifications, or deletions of DNS zones and servers, zone and server parameters, and DNS resource records. We’ll examine the sources of these data changes, how the Internet-standard DNS architecture accounts for propagating changes and an alternative approach to provide rapid and reliable change management with resiliency.