Account for official IP range availability

[sopel]

Fortunately AWS has finally started to offer an IP ranges API, see AWS Public IP Address Ranges Now Available in JSON Form - they are expected to change several times per week, so we need to decide whether to either automate the integration or drop the respectively outdated ranges within missingcloud.

• While at it, integration of eu-central-1 should also be handled ideally.

This project's goal has been to make missing cloud-related data available in a format that you can use programmatically, which is now the case for most of the listed stuff, though some gaps remain.

@garnaat - can you judge/recommend the most reasonable way forward from here:

1. maintain missingcloud as is and automate IP ranges integration
2. maintain missingcloud as is and remove IP ranges integration in favor of the readily available AWS IP Address Ranges
3. decommission missingcloud in favor of the readily available AWS IP Address Ranges, plus somehow reusing the substantially improved 'missing data' generation meanwhile facilitated in boto/botocore (presumably from AWS internal docs/APIs) for the non IP address aspects (e.g. /botocore/data/aws/ec2/2014-06-15.api.json)
   • :exclamation: this would still lack EC2 instance type details like memory/storage/CPU I think, which could be partially scraped from the official listing though for example

[garnaat]

I've been thinking about the same thing.

My personal feeling is that there are two options:

1. Decommission this project and provide links to the places where similar data can be found.
2. Turn this into a web service that has the ability to do more interesting processing of existing data sources and expose them in a nice, consistent API.

I suspect that over time, all of the existing data presented by this project will be available programmatically by a more authoritative source but it will take time and in the process other data might arise that needs to be made available. I think a service would be the most general approach to this problem. I'm just not sure if the demand for the data justifies it.