Re: Lists and EPP
Klaus Malorny <[email protected]>
| Newsgroups | gmane.ietf.provreg |
|---|---|
| Message-ID | <[email protected]> |
On 08/06/11 11:16, Bernie Hoeneisen wrote: > Hi Ulrich > > As I understand this functionality is out-of-scope for EPP. Such functionality > has been addressed and standardized by the CRISP working group as IRIS protocol. > > For reference, see RFC 3981 ff. > > http://tools.ietf.org/html/rfc3981 ff. > > cheers, > Bernie > Hi, hmm, for some reasons, CRISP never really took off. And are there any alternative transport mechanisms besides BEEP, which does not seem to have widespread library support? Besides this, an implementation with CRISP has probably the same issues as an EPP implementation, namely that you have potentially to deal with hundred thousands domains, which may eat up your resources server-side. However, generally, a reporting mechanism could be done in an incremental way, so that it even could be implemented in EPP: If one considers the list of domains being lexicographically sorted, one can give a starting point from which the names shall be reported. The server returns only the next n names according the given sort order (n ranging from a few tens to a few hundreds). To continue the list, one specifies the last name of the last response as the new starting point. While this seems to be less efficient, it has the benefit that the returned list is nearly up-to-date (compared to generated reports which may be up to 24 hours old), and one can do one's synchronization stuff in a continuous mode if need be. If this is too slow, the client can even use multiple starting points in parallel. We have implemented such methods in different places and it works sufficiently enough. regards, Klaus _______________________________________________ provreg mailing list [email protected] https://www.ietf.org/mailman/listinfo/provreg