Comparative availabilities of resilient cloud architectures

It is challenging for IT architects to design their cloud applications to be resilient without indicative data on the availability of different architectures. In this report, Uptime Institute Intelligence uses historical provider status updates from Amazon Web Services, Google Cloud and Microsoft Azure to calculate the availabilities of several resilient cloud architectures, including a multi-cloud implementation.

KEY POINTS

• Historical data indicates hyperscale cloud providers delivered 99.99% (or better) service availability in most availability zones and regions in 2022 — substantially exceeding service level agreements and design goals.
• The historical availability of a dual-region architecture was found to be only slightly greater than that of a dual-zone architecture. Operating expenditure doubled, however.
• This expenditure might be justified, given that, in some configurations, the dual-region architecture was found to suffer a third of the cloud-service downtime of a single-zone architecture.
• Historical status data suggests dual-cloud architectures might deliver 99.9995% availability on average. However, in the worst-case scenario they are only marginally better than dual-region architectures, but substantially more expensive.