Re: pervasiveness of surrogate (also called synthetic) keys

Craig Ringer wrote:
> On 03/05/11 11:07, Greg Smith wrote:
>
>
>> That doesn't mean you can't use
>> them as a sort of foreign key indexing the data; it just means you can't
>> make them the sole unique identifier for a particular entity, where that
>> entity is a person, company, or part.
>>
>
> Classic case: a database here has several tables indexed by MAC address.
> It's used for asset reporting and software inventory.
>
> Problem: VMs generate random MAC addresses by default. They're not
> guaranteed to be globally unique. Collisions have happened and will
> probably happen again. In this case, it wasn't a big deal, but it just
> goes to show that even the "obviously" globally unique isn't necessarily so.
>
> --
> Craig Ringer
>

Hm.. Virtual machines as assets.  Mortgage backed securities, anyone.