Fiber optic cable of Microsoft Azure data centers is defective due to storm Netherlands – IT Pro – News

Nearly. Couple of minor notes. What you name an AZ is definitely a Area. And a Area is subdivided into separate AZs. Merely put, a Area is a DC or group of DCs shut along with lightning-fast interconnects (fiber from the operator itself). Every AZ typically has its personal energy and networking/core routers, and so can fail independently of different AZs.

For instance with Google Cloud Platform you may have the Area eu-west4 (NL/Groningen) And there are 3 AZs: eu-west4-aeu-west4-beu-west4-c. This terminology can be utilized in Azure and AWS. A Area will not be restricted to three AZs, by the way in which. For instance, eu-west1 (BE/St. Ghislain) has 4 AZs, so additionally an eu-west1-d.

As a rule, it’s smart to divide your servers among the many completely different AZs of a Area. Community visitors inside a Area is usually very low-cost and even free. This manner you possibly can cluster a number of servers with out having to pay for visitors inside your cluster. Mainly you have to be protected from issues like fireplace and damaged uplinks that approach.

A multi-region setup is usually harder and complicated. That is the place issues like latency come into play, which is not supreme in lots of conditions.

If I learn the message from Micro$oft accurately, this issues damaged interconnect(s?) between numerous AZs throughout the West Europe Area. In that case, it’s typically greatest to modify off the workload within the remoted AZ, in order that the opposite two AZs can proceed with out a hitch. Nevertheless, the message from MS may be very quick… Not completely clear which traces/AZ’s are damaged precisely..

AWS: https://docs.aws.amazon.c…s-availability-zones.html
GCP: https://cloud.google.com/compute/docs/regions-zones
Azure: https://study.microsoft.c…ailability-zones-overview