As your organization and its trading partners become more and more integrated, it is becoming critical that file transfers are performed without delays or disruptions. For instance, a document containing a batch of transactions could traditionally be delivered within a window of several hours without causing any issues. But today, in the effort to make business processes as efficient as possible, that same document must now be delivered within seconds.
Organizations are therefore taking a closer look at how they can provide the best high availability for their systems to minimize any potential disruptions to their file transfers.
Comparing High Availability Strategies
Many of the secure file transfers from your trading partners are probably going through an SFTP server in your organization's network. If that SFTP server were to go down (for example, due to a CPU or drive failure), then you would need to fail over to a redundant backup system to continue to service your trading partners, thus maintaining high availability.
Two common approaches for providing high availability for SFTP servers and most other applications are Active-Passive or Active-Active.
With an Active-Passive approach, only one SFTP server will be active at a time to service your trading partners. A backup copy of the SFTP server would exist on your network as a "passive" system, meaning that it is installed and configured, but it is not actively running.
To prepare in the event of a failure of the active SFTP server, it is important that you frequently replicate all settings and configuration files from the active SFTP server to the passive system. If the active SFTP server fails, then the passive SFTP server could be launched and your network configured to point to this new system.
In an Active-Passive configuration, the downtime for your trading partners (when a failure occurs) can be a few seconds or several hours depending on how the passive system is started.
The least efficient and often slowest implementation of an Active-Passive approach is to rely on human intervention to detect the failure and then manually start up the passive system. This could take several hours depending on when the outage is reported, the process to start the passive system, and the complexities of configuring the network to route traffic to the new system.
A much better approach would be to have a third-party system monitoring tool that would immediately detect when the SFTP server fails, and then would automatically start up the passive system. The result should be a a much shorter disruption for trading partners of only a few seconds.
Active-Active, or Clustering
The next level in high availability is to use an Active-Active approach, also referred to as "clustering."
With Active-Active, two or more installations of the SFTP server can be running concurrently, sharing the same set of configurations and trading partner accounts. The SFTP servers in the cluster are in constant communication with each other, so if one of the SFTP servers were to fail, the remaining systems in the cluster will continue to service the trading partners. This configuration will provide the maximum high availability since it is not dependent on human interaction or third-party tools to start up other systems.
If you need maximum up-time for your SFTP server, GoAnywhere Services now offers clustering.
Another advantage of an active-active configuration is that you can load balance the traffic over multiple systems, which is important when you need to service a large number of trading partners. This will require that you install a load balancer like GoAnywhere Gateway in front of the cluster. Typically this load balancer will be in your DMZ and will be your trading partners' initial point of contact.
The Bottom Line
Both Active-Passive and Active-Active methods provide high availability for your SFTP server environment if configured properly. However, Active-Active will provide the maximum up-time because it keeps multiple SFTP servers running concurrently in a cluster, along with the added benefit of load-balancing.
How critical up-time is to your bottom line will be the best guide to determining which high availability approach best fits your organization.