Active Standby vs Active Active
Active/Standby and Active/Active are two failover mechanisms used widely all over the world to improve the reliability of the systems. Also, these two methods can be considered as high availability implementation methods. Each mechanism has its own method to determine and perform failover. Different systems use these methods to achieve the required level of redundancy depending on the level of the critical nature of the instance.
Active/Standby Configuration
In Active/Standby Configuration, only one node is in active mode while the other is in standby mode. When an issue is identified on the Active system, the standby node will take the place of the active node without any changes on the last state until such time the issue is resolved. However, in this case, whether to switch back to the original node after restoration of the issue or not may depend on the configuration of the two nodes. Also in general, there should be some kind of synchronization between the active and standby nodes in order to switch instantly on a failure. In most cases, heartbeat signals between active and standby nodes are used to identify the failure of the active node as well as for the real time synchronization between nodes. Here, always only one set of equipment is active all the time hence, simplifies the routing and troubleshooting. Also a failure in heartbeat link, leads both nodes to independent mode where use of shared resources might get inconsistent depending on the configuration. In Active/Standby configuration there is no need to implement a load balancing method before the nodes in order to share the load, since only one node will be active at any given time unless there is inconsistency.
Active/Active configuration
In Active/Active configuration both nodes are in active mode while handling the same function on the same state. If there is a failure in one active node, then the other active node automatically handles the traffic and function of both nodes until the issue is resolved. Here, both nodes should have the capacity to handle the total traffic individually in order to work independently in a failure situation without any performance or quality degradation to the ultimate function. After restoration of issue, both nodes will go to active mode, where load will be shared between nodes. As a general practise in this configuration, there should be a mechanism to share the load between nodes using some kind of load balancing method to keep both nodes in active mode simultaneously. Also, failure identification should take place at the load balancing point in order to shift the whole load to available node.
|
What is the difference between Active/Standby and Active/Active Configuration? – In the Active/Standby configuration, utilization of the standby node is virtually zero even though it is up and running all the time, whereas with Active/Active configuration capacity of both nodes can be used up to a maximum of 50% in general for each node, since one node should be able to take the whole load in case of a failure. – Therefore, if more than 50% is utilized for any active node under Active/Active mode then there will be a performance degradation in case of a failure in one active node. – In Active/Active configuration, failure in one path will not lead to service outage, whereas with Active/Standby configuration, it can vary depending on the failure identification time and shifting time from active node to standby node. – Active/Active configuration can be used as a temporary throughput and capacity expansion in case of unanticipated scenarios, even though, it leads to performance degradation during a failure. – Whereas, with Active/Standby such option is not available even for a momentary situation. – Even though Active/Active configuration have this capacity expansion advantage, there should be a load balancing method before the nodes, which is not required under Active/Standby configuration. – Active/Standby method is less complex and easy to troubleshoot the network, since only one path is active all the time compared to Active/Active method, which keeps both paths and nodes active at the same time. – Active/Active configuration typically supports load balancing, whereas, with Active/Standby configuration no such solution is available. – Even though, Active/Active configuration allows momentary capacity expansion, in general, it provides additional complexity to the network than the Active/Standby configuration. – Since both paths are active under Active/Active configuration, outage time is virtually zero in case of failure, which can be higher in case of Active/Standby configuration.
|