Topology load balancing distributes DNS name resolution requests based on the proximity of the client to the data center housing the resource that responds to the request. When Topology load balancing is enabled, the BIG-IP system uses topology records to make load balancing decisions.
A topology record is a set of characteristics that maps the origin of a DNS name resolution request to a destination. This screen capture shows the Topology record creation screen in the Configuration utility.
Each topology record contains the following elements:
When Topology load balancing is configured, the order of the topology records is vital and affects how the BIG-IP® system scores the pools or pool members to which it load balances DNS name resolution requests. By default, the BIG-IP system prioritizes topology records using Longest Match sorting. As a result, topology records are automatically sorted based on a specific criteria each time the BIG-IP system configuration loads. Alternatively, you can disable Longest Match sorting and customize the order of the topology records in the list.
When Longest Match is enabled, the BIG-IP system sorts the topology records by the LDNS request source statement, the destination statement, and the weight of the record.
The system first sorts the topology records by the type of LDNS request source statement using this order from highest to lowest:
If the type of LDNS request source statement is the same in multiple topology records, the BIG-IP system then sorts these records by the type of destination statement using this order from highest to lowest:
If the type of LDNS request source statement is the same in multiple topology records and the type of destination statement is the same in those records, the system then uses the value of the weight from lowest to highest to sort the records.
The example shows a list of topology records sorted automatically using Longest Match. Note that the fourth and fifth records have the same LDNS subnet and the destinations are both of type State; therefore, the weight determines the position in the list, thus the record with the lower weight is first.
When you configure a wide IP for Topology load balancing, you can route DNS requests to the data center that is closest to the client making the request. With this configuration, the BIG-IP® system load balances DNS name resolution requests to the pools in a wide IP based on the topology records. (The members of each pool must be in the same data center.)
This example illustrates and describes how DNS name resolution requests are load balanced when a wide IP is configured for Topology load balancing. An administrator configures the wide IP www.siterequest.net for Topology load balancing. The wide IP contains three pools: Pool1 and Pool3 are located in the North America data center; Pool2 is located in the South America data center. Next, the administrator creates topology records, as shown in the following figure, and ensures that Longest Match is enabled on the BIG-IP® system.
When you configure a pool for Topology load balancing, you can route DNS requests to the data center that is closest to the client making the request. With this configuration, the BIG-IP® system load balances DNS name resolution requests to the members of the pool.
You can configure wide IPs for Topology load balancing. You can also configure a pool for Topology load balancing. Additionally, you can configure both a wide IP and the pools in the wide IP for Topology load balancing.
When you configure both a wide IP and the pools in the wide IP for Topology load balancing, the BIG-IP® system uses topology records to load balance DNS name resolution requests first to the pool assigned to the wide IP, and then to a member of the pool.
The BIG-IP® system uses an IP geolocation database to determine the origin of DNS requests. The database included with the BIG-IP system provides geolocation data for IPv6 addresses at the continent and country levels. It also provides geolocation data for IPv4 addresses at the continent, country, state, ISP, and organization levels. The state-level data is worldwide, and thus includes designations in other countries that correspond to the U.S. state-level in the geolocation hierarchy, such as, provinces in Canada.
The BIG-IP® system uses an IP geolocation database to determine the IP addresses that match the geographic names that you define in a topology record, such as continent and country.