Investigate complex enclave scenarios

From discussions with stakeholders, it came to our attention that enclave scenarios can be very complex.

In this spike we would like to confirm that the design in CFE-896 can answer the requirements of several stages of disconnected environments mirroring from one disconnected environment to the next.

Such a scenario is illustrated in https://lucid.app/lucidchart/0936e002-1031-4522-9a0a-5c9327649f2d/edit?invitationId=inv_dfc82202-4050-4af8-b22e-6598c2ef81c6&page=0_0#

To confirm our findings we need to:

• Create EC2 instances on AWS 
  • Accessible from ssh
  • With Docker registry distribution (registry) installed and accessible
  • 3 instances 5000,6000,7000

• Check that oc-mirror v2 can natively use Registries.conf (skopeo)
• Use env var to make Cincinnati API URL configurable: check that oc-mirror v2 uses the configured cincinnati URL
• Check if we can have one cache per enclave (no blobs)
• Create some sort of audit file for each enclave
• Generate diff tars based on the audit files
• Be able to have several chunks of diff tar (configurable size)
• Simulate cincinnati server and data
• Small release to test easily
• Test an oci index => digest needs to be preserved

https://docs.google.com/drawings/d/1FQ1XuXbKBkqVOzxTWkbYeiAPUcR2Ziq3nCDLFsX5ujs/edit

This issue is timeboxed to 3 days