Discussion on cloud storage

[UsulofDoom]

Anyone truly know the ins and outs of it? Let’s say I create a 1G byte zipped file. I then send to the cloud. How much space does that take up in the digital world?

It has to be backed up to other server farms I would think. Do they have air gap backups like tape or optical? Do they compress their backups?

Does the file take up more storage every year forever? Can it every be truly deleted? How does this affect the European web laws?

[malchior]

Caveat IMO most of these questions are too broad to be answered absolutely correctly. Each of these topics could need hours of clarification to reach a specific design for a user application but an attempt was made at a superficial level.

Anyone truly know the ins and outs of it? Let’s say I create a 1G byte zipped file. I then send to the cloud. How much space does that take up in the digital world?

Likely 1GB of paid storage. There is typically no additional compression possible. You could choose to store it in multiple places though and it'd be multiples of that.

It has to be backed up to other server farms I would think.

It depends heavily on what 'storage tier' is being paid for and how the user configures it. Data can be configured to be stored in one place or multiple places. You have a lot of control at a broad level where data is stored.

Do they have air gap backups like tape or optical?

Probably not. The AWS tech is proprietary/trade secret stuff. It is thought to range from the fastest SSDs possible ranging down to cheap hard drives on custom controllers to control the main cost (electricity). Disks of the spinning and solid state size are much more space/cost efficient. Microsoft/GCP/Oracle/Digital Ocean all are probably using similar tech. If you want to create analogs of air gap you possibly can but the design details are myriad here. AWS for instance has WORM functionality options in their storage service offerings.

Do they compress their backups?

Define backups. Loosely speaking you have to think differently about design in the cloud. A chief difference is that you are paying for storage. The application defines the parameters and costs for that storage. If you call it a backup it is still data being stored somewhere and you pay for that usage. If some data is defined as a 'backup' and the application can tolerate waiting hours to get it back then you can save money by storing that data on cheaper and slower to restore storage. If you want it faster, you pay more. If you want it to be hot ready to go all the time you might pay the most. You have a lot of cost control but it comes with a lot of complexity.

Does the file take up more storage every year forever?

I think I am following this. It effectively doesn't matter. If you store 1GB of data then you pay for 1GB of data storage at some rate. If it takes up 5GB inside the AWS black box...that is AWS's business.

Can it every be truly deleted?

This is another where the variables are many and the question is just too broad. I can give you an example though - S3 is the AWS object based storage service. You put a file in there (in what they call a bucket). You can either overwrite it or delete it. There is another option called versioning which I won't dive into since it isn't important here. When you delete, AWS deletes the reference of the object from a global registry and designates the system to write to that portion of the disk in the platform first. Meaning it is physically overwritten almost immediately. AWS also physically destroys storage devices when they decommission hardware.

How does this affect the European web laws?

This is where you drop into a large universe of design/legal questions. If you are wondering about these things for a business purpose at any substantial risk you need to engage expensive experts (e.g., cloud architects and probably lawyers) to get an answer.