It really depends on how they are using AWS. If they are just using them for web hosting using a public API along side database services with standard SQL (As much as there is a standard SQL), they can migrate quickly. But, the more they lean on AWS APIs, the more fucked they are.
I've used AWS in the past for projects, and I've generally elected to do the above. Use their hardware to run virtual machines and run SQL for initial buildout. When small, we didn't have to buy hardware or guess about the demand load. The nice thing is that it they way I implemented it, it was easy to pull my data off of AWS to run on our own hardware to save costs as the project matured. When transitioning to our own hardware, it was something like a 15 minute downtime while we were changing DNS records and flushing DNS cache. But it was something we were able to spend a few weeks to plan and test before flipping the switches. AWS is still a good option (speaking from a business, not freedom perspective) to ramp for unexpected demand quickly. The code is still in the project for us to buy resources from AWS in case we see unexpected demand. But, AFAIK, it's deprecated.
The biggest roadblock they would have is to get an ISP and COLO, especially with COVID restrictions going on. The second would be to physically get the servers. Fortunately, they probably have a good idea about the server load, bandwidth, and SQL load they will need to build out.
It really depends on how they are using AWS. If they are just using them for web hosting using a public API along side database services with standard SQL (As much as there is a standard SQL), they can migrate quickly. But, the more they lean on AWS APIs, the more fucked they are.
I've used AWS in the past for projects, and I've generally elected to do the above. Use their hardware to run virtual machines and run SQL for initial buildout. When small, we didn't have to buy hardware or guess about the demand load. The nice thing is that it they way I implemented it, it was easy to pull my data off of AWS to run on our own hardware to save costs as the project matured. When transitioning to our own hardware, it was something like a 15 minute downtime while we were changing DNS records and flushing DNS cache. But it was something we were able to spend a few weeks to plan and test before flipping the switches. AWS is still a good option (speaking from a business, not freedom perspective) to ramp for unexpected demand quickly. The code is still in the project for us to buy resources from AWS in case we see unexpected demand.
The biggest roadblock they would have is to get an ISP and COLO, especially with COVID restrictions going on. The second would be to physically get the servers. Fortunately, they probably have a good idea about the server load, bandwidth, and SQL load they will need to build out.
It really depends on how they are using AWS. If they are just using them for web hosting using a public API along side database services with standard SQL (As much as there is a standard SQL), they can migrate quickly. But, the more they lean on AWS APIs, the more fucked they are.
I've used AWS in the past for projects, and I've generally elected to do the above. Use their hardware to run virtual machines and run SQL for initial buildout. When small, we didn't have to buy hardware or guess about the demand load. The nice thing is that it they way I implemented it, it was easy to pull my data off of AWS to run on our own hardware to save costs as the project matured. When transitioning to our own hardware, it was something like a 15 minute downtime while we were changing DNS records, and flushing DNS cache. AWS is still a good option (speaking from a business, not freedom perspective) to ramp for unexpected demand quickly. The code is still in the project for us to buy resources from AWS in case we see unexpected demand.
The biggest roadblock they would have is to get an ISP and COLO, especially with COVID restrictions going on. The second would be to physically get the servers. Fortunately, they probably have a good idea about the server load, bandwidth, and SQL load they will need to build out.
It really depends on how they are using AWS. If they are just using them for web hosting using a public API along side database services with standard SQL (As much as there is a standard SQL), they can migrate quickly. But, the more they lean on AWS APIs, the more fucked they are.
I've used AWS in the past for projects, and I've generally elected to do the above. Use their hardware to run virtual machines and run SQL for initial buildout. When small, we didn't have to buy hardware or guess about the demand load. The nice thing is that it they way I implemented it, it was easy to pull my data off of AWS to run on our own hardware to save costs as the project matured. AWS is still a good option (speaking from a business, not freedom perspective) to ramp for unexpected demand quickly. The code is still in the project for us to buy resources from AWS in case we see unexpected demand.
The biggest roadblock they would have is to get an ISP and COLO, especially with COVID restrictions going on. The second would be to physically get the servers. Fortunately, they probably have a good idea about the server load, bandwidth, and SQL load they will need to build out.
It really depends on how they are using AWS. If they are just using them for web hosting using a public API along side database services with standard SQL (As much as there is a standard SQL), they can migrate quickly. But, the more they lean on AWS APIs, the more fucked they are.
I've used AWS in the past for projects, and I've generally elected to do the above. Use their hardware to run virtual machines and run SQL for initial buildout. When small, we didn't have to buy hardware or guess about the demand load. The nice thing is that it they way I implemented it, it was easy to pull my data off of AWS to run on our own hardware to save costs as the project matured. AWS is still a good option (speaking from a business, not freedom perspective) to ramp for unexpected demand quickly.
The biggest roadblock they would have is to get an ISP and COLO, especially with COVID restrictions going on. The second would be to physically get the servers. Fortunately, they probably have a good idea about the server load, bandwidth, and SQL load they will need to build out.