[Dev] oid coinage problem

[petite_abeille]

On Saturday, Apr 5, 2003, at 00:42 Europe/Zurich, rys mccusker wrote:

> what if the clients do not use globally unique ID algorithms for
> inventing new object IDs?

Why not?

>
> let's ignore the possibility that using globally unique IDs alone
> would actually be attractive to someone.

Why?

>  Let's assume people want
> simple ascending integers

Why?

>  for simplicity,

How is that simple?

>  or want low-cost ID
> generation

How is centralizing id creation "low-cost"?

>  -- perhaps because of high frequency object creation.

What is "high frequency object creation"? 1 per second? 10? 100? 1000? 
A million? Per clients? Is that for a PIM or a telecom switch? How is 
centralization going to help?

"One of the main reasons for using UUIDs is that no centralized
   authority is required to administer them (beyond the one that
   allocates IEEE 802.1 node identifiers). As a result, generation on
   demand can be completely automated, and they can be used for a wide
   variety of purposes. The UUID generation algorithm described here
   supports very high allocation rates: 10 million per second per
   machine if you need it, so that they could even be used as
   transaction IDs."



>
> I have not yet added to that page the additional material due to
> and all flight interchange with a friend who suggested some ID
> allocation schemes which allocate blocks of IDs to permit them
> to be some allocated by clients after getting an entire block.

What is the scope of your schema? One client, one server? Many clients, 
one server? Many clients, many servers? Is the system centralized or 
not? To which extend? Is the osa foundation going to run an ultimate 
object id issuance authority?