In this section:
High Availability, in a general sense, means that systems should be available when needed.
Customers must define high availability for their systems, which can be 24/7 or for a specified period. This leads us to the topic of a maintenance window and unplanned downtime.
Any system can undergo disruptions that are within a maintenance window or unplanned. In all cases, procedures and control measures must be put in place to prevent, detect, and address these situations.
Maintenance Window Downtime
A maintenance window is a critical consideration when dealing with high availability systems. For instance, when applying Service Packs or Hotfixes, you will need a plan that minimizes disruptions to the end user. In this example, measuring the length of planned downtime has to take into account:
Unplanned events can occur at any time. You can have a hardware failure or a power outage. Regardless of how high availability is defined, the deployment plan must have built-in control measures for preventing, detecting, and addressing unplanned events.
For systems that have high availability defined as 24/7 or for a specified period, there may be some flexibility with the system being unavailable. However, for mission-critical systems, control measures take on an even greater significance. A Disaster Recovery Plan (DRP) must be in place. This is covered in Planning for Disaster Recovery.