If we setup 10 ECS instances in 10 different regions and asked them to communicate with each other. What will be the differences to setting up 10 nodes in those exact same regions and performing the same task? Hardware specs will be the same.
如果我们在10个不同的区域设置10个ECS实例并要求它们相互通信。在这些完全相同的区域中设置10个节点并执行相同任务会有什么不同？硬件规格将是相同的。
In other words, when tests are done on aws region to region, is this a good indicator of what tests in the wild will be like, or does amazon do optimisation tweaks that do not allow you to setup an environment that will be indicative of the real world, without AWS?
换句话说，当在aws区域到区域进行测试时，这是一个很好的指示，表明野外测试会是什么样的，或亚马逊做的优化调整不允许你设置一个指示现实世界，没有AWS？
One test parameter could be latency for example
例如，一个测试参数可以是延迟
1 个解决方案
#1
0
If you set up 10 different EC2 instances and access it from outside AWS, the traffic will be routed through the internet. So its fair to say you can simulate the latencies between regional locations.
如果您设置10个不同的EC2实例并从外部AWS访问它，则流量将通过互联网路由。所以可以说你可以模拟区域位置之间的延迟。
However, if you use services like AWS CloudFront, the results would be different since from the CloudFront edge location (more edge locations are there than the regions) onwards there is the possibility of routing traffic internally to an AWS origin.
但是，如果您使用AWS CloudFront等服务，则结果会有所不同，因为从CloudFront边缘位置（比区域更多的边缘位置）开始，可以将流量内部路由到AWS源。
