dns for hot standby, failover and failback

When your internet service must always be available, hot standby is your answer. The edgedirector managed dns platform helps you maintain 7x24 availability.

Topic Sections

• hot standby servers
• geographic diversity
• automatic failover
• automatic failback
• planned maintenance
• partial replicas
• network hardware savings

HOT STANDBY SERVERS
Hot standby servers are separate physical servers that are held in reserve in case of production server failures. They are distinguished from load balanced servers in that they are not used except when production servers are not available.

GEOGRAPHIC DIVERSITY
Server redundancy is just one consideration in improving service continuity. Failures may also occur in the supporting infrastructure or natural disasters may strike. Geographic diversity in the placement of hot standby servers addresses those failures. When the hot standby facility is located in a different locale, all of the supporting infrastructure is replicated. The greater the distance, the more likely that any natural influences will be limited to the original production site.

AUTOMATIC FAILOVER
Failures in the production servers can happen at any time. They can occur in the middle of the night or on holidays. Unless a site can dedicate resources to watch the service manually, automatic failover is an absolute requirement. Automatic failover ensures that website or other service users will be automatically redirected to the hot standby environment whenever necessary.

AUTOMATIC FAILBACK
Production failures may last any amount of time. But, it is often desirable that clients be served by the production servers whenever possible. In those cases, it is important that the failover mechanisms monitor the availability of the production servers. When the servers are available again, the service requests are again directed to the production servers by the failback mechanism.

However, some service configurations require synchronisation before failback can be permitted. In those situations, failback can be implemented as a manual override to the autofailover.

PLANNED MAINTENANCE
Hot standby facilities are also useful in planned maintenance operations. By triggering a manual failover, upgrades, scheduled maintenance and repairs can be performed with minimal disruption to service availability. With sufficient hot standby functionality, extended maintenance windows are more easily accommodated. Failback is manually triggered once the planned activity is completed.

PARTIAL REPLICAS
In some instances fully replicating the live servers is unnecessary. The hot standby servers may be intended to provide limited service functionality while the production site is undergoing planned maintenance. Default failover facilities are already provided by edgedirector.com servers for its customers. However, some customers prefer a destination that carries its full visual branding identity. This may require only a shared server at their hosting provider of choice.

NETWORK HARDWARE SAVINGS
Network hardware vendors such as Cisco, Juniper and Foundry Networks market devices that permit global failover as part of their product lines. They are labeled with terms such as gslb, load directors, or global load balancers. Using such hardware to build a hot standby capability can be expensive and time consuming. A managed dns hosting provider is usually a better choice for many sites. The service provider assumes all responsibility for design, implementation and maintenance. All the customer has to do is sign up for the service as a subscriber.