Clustering, Load-Balancing and Failover

This article describes how to use clustering with  Airlock to provide failover capability and load balancing. This may be realized with Airlock own functions or with dedicated load balancers in front of or behind  Airlock.

Three sceneries

Two Airlocks together may be operated as a failover cluster. If so, one Airlock is active, receives data and is monitored by the passive one. If the passive Airlock detects an error on the active one, it takes over automatically. The monitoring is done on a separate channel by using the cluster talk protocol.
For loads that cannot be managed with one single hardware anymore, Airlock supports upstream load balancing.

Single Airlock

This kind of scenery is recommended only for testing and integration environment.

The following list shows the major positive and negative aspects.

• + disk mirroring is possible
• + failover and load balancing of back-end application
• + for testing and integration environment
• + cheap
• + any hardware
• + internal process monitoring with self-healing
• + disk monitoring included
• + session stickiness

• - no blackout protection
• - single point of failure

Failover Cluster

This kind of scenery is used most for productive Airlock environments.The following list shows the major attributes of a failover cluster Airlock system.

• active/passive Airlock
• availability even during update process
• two hot standby systems
• only one system active
• high availability thanks to mutual monitoring
• health checks by passive Airlock system
• identical hardware needed
• note the dimensioning for blackout!

Load balanced Airlock farm

This kind of scenery is used for high performance and high redundant system requirements.The following list shows the major attributes of a load balanced Airlock farm.

• two or more systems
• external load balancer needed
• high availability thanks to application related health checks on load balancer
• highest capacity thanks to any scalability
• note dimensioning for blackout!