Active Standby and Active Active are two different configurations for high availability (HA) and load balancing in computer systems. In both configurations, there are multiple servers available to process requests, but they differ in how those servers are utilized. In an active-standby configuration, there is a single active server processing requests, while the other servers are in standby mode. In contrast, an active-active configuration distributes the workload across multiple active servers. Each configuration has its own advantages and disadvantages and is chosen based on the specific needs of the system. In this response, I will elaborate on the differences between active standby and active-active configurations.
Active-active clustering and active-passive clustering are the two most common configurations for high availability (HA). What is the difference between them? What you should know about the differences between active-active and active-passive nodes.
High-Availability Active-Active Cluster
A cluster that is active-active typically consists of two or more nodes running the same service at the same time. A cluster that is active-active has as its main goal load balancing. Load-balancing spreads workloads over all nodes to avoid overloading any one node. There will be an improvement in response time and throughput because there are now more nodes to serve.
This set-up, which includes a load balancer (ex. This type of cluster configuration is HA. Web clients connect to the load balancer instead of directly connecting to the HTTP server. The load balancer then connects the client to the HTTP servers that are behind it.
It’s not a random process to assign clients to nodes within the cluster. It’s determined by the load balancing algorithms set up on the load balancer. In a “Round Robin”, for instance, the client that connects first is sent to the server. The second client then goes to the server. In a subsequent post, we’ll go into more detail about these algorithms.
High-Availability Active-Passive Cluster
A cluster configured as active-passive also has at least two nodes. As the name “active/passive” suggests, not all nodes will be active. If there are two nodes and the first one is already active, then the second must be passive, or in standby.
Active-passive failover is used when one of the nodes fails. The passive server (failover server) serves as a backup server that can take over immediately if the primary server (active server) becomes disconnected or unable to function.
Difference Between Active Standby and Active Active
- The capacity of each node can be utilized up to 50% for the Active/Active configuration. This is because one node will be able take on the entire load in the event of a failure.
- If more than 50% of the nodes are active in Active/Active mode, then performance will be affected in the event that one node fails.
- With Active/Active configurations, failures in one path do not cause service outages, but with Active/Standby, the situation can be different depending on failure identification time and time to shift from active to standby.
- The Active/Active configuration is useful for temporary capacity and throughput expansions in the event of unexpected scenarios. However, performance can be affected by a failure.
- With Active/Standby, this option is unavailable even in a temporary situation.
- The Active/Standby Method is easier to troubleshoot, as only one path remains active, compared to the Active/Active Method, which keeps nodes and both paths active simultaneously.
- Load balancing is typically supported by Active/Active configuration, but not Active/Standby.
- Active/Active configurations allow for momentary capacity increases, but in general they are more complex than Active/Standby configurations.
- In the case of failures, there is virtually no outage in Active/Active configuration. This can be a problem in Active/Standby.
Comparison Table about Active Standby and Active Active
| Active-active | Active-passive |
|---|---|
| Setup : Configure two Adaptive servers as companions, with each having independent workloads. These companions are run on both the primary and secondary servers as separate servers, until one of them fails. | Setup : One Adaptive Server can run on either the primary or secondary node. The Adaptive Server is run on the primary node prior to a failover and the secondary after the failover. |
| Failover : If failover occurs, the secondary partner takes over all devices, client connections and so on, from the primary companion. The secondary companion will service the clients who have failed over, and any new clients until the primary companion returns to its normal activities. | Failover : If a system fails, then the Adaptive Server, along with its resources, is relocated and restarted at the secondary node. |
| Failback : Failback occurs when the primary companion returns its devices to the secondary companion and reconnects the client connections. | Failback : Failback refers to a planned failover, or relocation, of the Adaptive Server with its resources on the primary node. Failback isn’t required but can be done to serve administrative purposes. |
| Client connection failover: in failover time, consumers commune with secondary companions to resubmit uncommitted transactions. During failback clients connect to their primary companion to resubmit transactions. Clients who have the failover property automatically reestablish connections. | Client connection failover : During failover or failback, clients reconnect to the same Adaptive Server in order to resubmit transactions that have not been committed. Clients that have the failover property automatically reestablish connections. |
Finale
Active-standby and active-active are two different configurations for high availability and load balancing in computer systems. Active-standby provides redundancy with only one server actively processing requests, while active-active provides better utilization of resources with multiple servers processing requests simultaneously. The choice between the two configurations depends on the specific requirements and constraints of the system, with active-standby being simpler and easier to manage, and active-active providing better performance and utilization at the cost of added complexity.