High Availability

---

High availability (HA) is ensuring systems remain operational and accessible with minimal downtime.

• involves designing and implementing infrastructure for fault tolerance and redundancy
• concept of availability can be measured over a defined period (e.g. one year)
  • can be measured as:
    • Uptime
      • the amount of time client data and resources are available on the servers
    • Downtime
      • time or percentage that a system is unavailable
      • maximum tolerable downtime (MTD) metric expresses the availability requirement for a particular business function
      • is calculated from the sum of scheduled service intervals plus unplanned outages over the period
  • usually loosely described as
    • 24x7
    • 24x365
  • Availability is often measured in the number of nines (including the whole number) found in a percentage
    • E.g., 99.999% uptime, it is stated as “five nines”

Nines Value	Availability	Annual Downtime (hh:mm:ss)
Six	99.9999%	00:00:32
Five	99.999%	00:05:15
Four	99.99%	00:52:34
Three	99.9%	08:45:36
Two	99%	87:36:00

Scalability and Elasticity

• Scalability and elasticity

Fault Tolerance and Redundancy

Fault tolerance is protection against system failure by providing extra (redundant) capacity.

• fault tolerant systems identify and eliminate single points of failure

Redundancy is overprovisioning resources at the component, host, and/or site level so that there is failover to a working instance in the event of a problem.

Site Considerations

• Disaster Recovery Sites

Cloud as Disaster Recovery (DR)

• Cloud as Disaster Recovery

Testing Redundancy and High Availability

• Load testing
  • incorporates specialized software tools to
    • validate a system’s performance under expected or peak loads
    • and identify bottlenecks or scalability issues
• Failover testing
  • focuses on validating failover processes to ensure a seamless transition between primary and secondary infrastructure
• Testing monitoring systems
  • validate effective detection and response to failures and performance issues