Choose your Cloud Regions Wisely so you can live Happily : )
After I read above news on 2023–01–20, I think it is good to write more about the Choice of a Cloud Region. It is complex to change your selected region (or move to a new region) AFTER you set up your workloads (e.g. Compute VMs, Database, Networking, Storage, Security) in a cloud region. Hope below tips can help you SELECT the RIGHT Region at the START. I will use Google Cloud as an example but the same concepts can be applied to other Clouds (e.g. AWS, Azure, OCI).
1/ Which Region has your desired services ?
I often go to the cloud provider status page to see if they got the services I need in a particular region, e.g. GCP status page
2/ Default Region or not ?
You can see, e.g. GCP doc, “When you create a project, Compute Engine automatically selects a default region and zone based on the location from which you create the project.“ Remember to change it if you do NOT want to use Default Region/Zone.
3/ Cost ?
You can easily compare the costs with each cloud provider cost tool, e.g. GCP Pricing Calculator, e.g. US East Virginia can cost 10+% more than US West Oregon.
4/ Performance (e.g. Latency) ?
You can easily see the latency difference in each region with GCP Network Intelligence center performance dashboard. Try to set up your Computer and storage closer to your customers and partners, so they can enjoy a faster web/mobile user experience. This is particular important if you are serving latency sensitive workloads, e.g. Games and Streaming.
5/ Limits and Quota ?
Make sure you increase the limits/quotas in the selected region so you can scale up and use the services successfully. Learn more at GCP Quota doc.
“Within these categories, some quotas are global and apply to your usage of the resource anywhere in Google Cloud. Others are regional or zonal and apply to your usage of the resource in a specific Google Cloud region (both quota types) or zone (allocation quotas only). For example, there are separate limits for how many Compute Engine VM instances that you can create in each Google Cloud region.”
6/ Growth Capacity ?
Try to work with Google Cloud Account team and/or support to know if you can ADD or Scale up Compute, Storage, Database, etc…easily later in your desired region . This will avoid the Y Combinator post that unable to add new nodes in Azure Germany region when needed.
7/ Compliance and Data residency/sovereignty?
Some countries and industries have very strict compliance and data residency requirements, e.g. Germany, “Data residency regulations expect businesses with operations in a specific country to keep and secure consumers’ personal information in data servers within that country’s borders. In the same guise, companies that want to operate in Germany must be able to comply with Germany data privacy laws.” .
GCP has tools, e.g. Assured Workload to help this.
8/ Go Green and pick the Low CO2 Region ?
Global warming is a real issue, so please choose the low CO2 cloud regions whenever possible, e.g. in GCP, you can easily see which region is low CO2 as in the image below. Below will help you choose the Low CO2 region and save the Earth.
a/ see which region is low CO2
b/ Carbon free energy for Google Cloud region
c/ How to incorporate carbon free energy in your location strategy e.g. Pick a cleaner region for your new applications. If you are going to run an application over time, running in the region with the highest CFE% will emit the lowest carbon emissions.
d/ Reduce your cloud carbon footprint with new Active Assist recommendations
9/ Do NOT put ALL your Eggs in 1 Basket? So, Do NOT put ~ALL your Cloud Workloads in 1 Region !
Often we will just skip the region choice and then it use the Default region, e.g. AWS US East (N. Virginia) (us-east-1) or GCP us-central1. Then the default region will become very crowded and the biggest region. It will cause a lot of customer impact if the default region goes down, e.g. AWS US-East-1 cloud region Lambda outage caused global issues on 2023–06–13 (due to a subsystem responsible for capacity management for AWS Lambda).
List of AWS outages in US-EAST-1 regions below per https://awsmaniac.com/aws-outages/ , so may be good to spread out some workloads to other regions ? You can read more at this post and comment.
