Network IP Addressing - Anycast vs. Unicast
In CDNs built around traditional Unicast DNS architecture, authoritative DNS servers (ADNS) are deployed in single-server units or clusters in dispersed geographic locations. Here, each IP address maps to a single physical location, with one or more servers providing ADNS service. Load balancing and redundancy is achieved within the site using traditional server load balancing techniques. Traditional CDNs use internal algorithms to map an end-user to a particular geolocation and, specifically, an IP address, for optimal mapping and performance.
Limitations of unicast:
- Failover/load balancing is via multiple A records
- If a server fails, the service is unavailable for the period that the DNS records are cached on the browser (this is why traditional CDNs make use of low TTLs on A records)
- Possibility of sub-optimal mapping
Figure 1: Unicast routing
The modern Anycast DNS architecture implemented on the Instart platform has ADNS "instances" deployed globally, with multiple copies of each ADNS server located in many geographically dispersed locations. Instances are referenced by IP address, but instead of a single server in a single location, there are multiple copies of the ADNS servers in multiple locations all over the world.
Anycast provides flexibility to have the same IP prefix advertised from multiple locations. The network then decides which location to route a user request to, based on routing protocol costs and possibly the "health" of the advertising servers.
Advantages of using Anycast:
- A single IP address can represent many servers, globally
- Load balancing/failover is maintained via IP address, rather than DNS
- The routing infrastructure directs packets to nearest instance of the service
- No special capabilities or algorithms are required for effective end-user mapping.
Figure 2: Anycast routing
Figure 3: Diagram showing how Anycast IP tables are maintained
If you have any questions, please contact Support.