Re: GUID's and Uniqueness.

David Fitzjarrell wrote:

> froliol_at_yahoo.com (Louis Frolio) wrote in message news:<94c28610.0409140636.2b28f6bc_at_posting.google.com>...
>

>>Greetings All, I have read many upon many articles here regarding GUID
>>data types and uniqueness. [...]
 >> From what I understand the MAC address of the NIC is
>>used as part of the algorithm to generate a GUID. 

Where did you find that documentation? All Oracle says is "host identifier on most platforms" -- which doesn't necessarily imply MAC address.

Not every platform that supports Oracle uses Ethernet, nor, I'm guessing, is it even a requirement that you have a network interface at all to run Oracle locally.

[...]
>
> Since Ethernet uses a 48 bit address space, there are potentially 2^48
> or 281,474,976,710,656 possible MAC addresses. [...]
> I hope this helps.
>
> David Fitzjarrell

I saw the same question asked on Metalink forums earlier this year (Michel Cadot replied!) -- but a quick search of Metalink turns up no details on the algorithm, so unless it's exposed to public scrutiny, we (on the outside) will never know for sure.

I suspect SYS_GUID() is unique in the same sense that modern encryption is "safe", or that computer programs generate "random" numbers. There is a non-zero chance you would generate a repeat GUID, in the same way there is a non-zero chance that someone could guess a value that encrypts to the same output as some other value. But in both cases, the chance is so small that it's well worth taking the risk for most everyday applications.

Of course, I'm pretty sure Oracle license prohibits use for operating nuclear reactors or flying airplanes, so unless your application is along those lines, you're probably fine.

--Mark Bole Received on Tue Sep 14 2004 - 18:39:26 CDT