AretNCarlsen

When running a server instance on Amazon/Rackspace/etc, where your server relies on /dev/random (e.g. for SSL and other cryptography),

4.) you have implemented your own RNG (e.g. *haveged* or a WAN-accessed entropy source).

1.) you haven't ever thought about whether /dev/random is reliable in a VM environment.

3.) you regularly run entropy testers against /dev/random to verify randomness (and hope/trust that the entropy you observe is unique, not copied to any other instances).

2.) you have thought about it, and chosen to hope/trust that the VM provider has implemented a reliable RNG (and that your server image/snapshot accesses it properly, especially restarted snapshots).

1.) you haven't ever thought about whether /dev/random is reliable in a VM environment.

2.) you have thought about it, and chosen to hope/trust that the VM provider has implemented a reliable RNG (and that your server image/snapshot accesses it properly, especially restarted snapshots).

3.) you regularly run entropy testers against /dev/random to verify randomness (and hope/trust that the entropy you observe is unique, not copied to any other instances).

4.) you have implemented your own RNG (e.g. *haveged* or a WAN-accessed entropy source).

3.) run entropy testers against /dev/random to verify entropy (and hope/trust that the entropy you observe is uninot copied to any other instances).

1.) haven't ever thought about whether /dev/random is reliable in a VM environment.

2.) have thought about it, and chosen to hope/trust that the VM provider has implemented a reliable RNG (and that your server image/snapshot accesses it properly, especially restarted snapshots).

4.) implement RNG independently in software (e.g. *haveged*).

Peachy Dandy [peachydandy.com], the web hosting service for two of my domains, ceased to be sometime on Sunday 2011-02-13. Nameserver domains unregistered, control panels inaccessible. No warnings, no email…