policy change: static verification methods
IP ip at sonera.se
Tue Feb 23 17:04:17 CET 1999
Hi Paula et al. I, and as far as I know, a couple of other LIR:s, have no objections for the proposed policy, on the contrary. The proposal has a pretty fine tune in balancing the end-customer need as stated as one of our assigning goals in RIPE-185, and the not-that-good policy comment of CATV-network customers as beeing as any dial-up customers, just becouse they are using a modem. As far as I recon, one have to use a 'modem' when going for a Leased Line as well, though signalling would not be analog. Any way, I think it perfectly ok to implement the policy right away. Best regards Per Lundberg/Sonera Registry/RSS11-RIPE > -----Original Message----- > From: Paula Caslav [mailto:paula at ripe.net] > Sent: Tuesday, February 23, 1999 4:44 PM > To: lir-wg at ripe.net > Cc: local-ir at ripe.net > Subject: policy change: static verification methods > > > > Hello all, > > A few weeks ago I send the following proposal to the lir-wg mailing > list for comments. Since no comments were received, we are assuming > that nobody has objections to change the policy. We have therefore > decided to implement it starting immediately. > > Kind regards, > > Paula Caslav > RIPE NCC > > ------- Forwarded Message > > Date: Wed, 10 Feb 1999 14:38:52 +0100 > From: Paula Caslav <paula at ripe.net> > Sender: owner-lir-wg at ripe.net > To: lir-wg at ripe.net > Subject: static verification methods > > > Hello all, > > At the Local IR working group at RIPE 32, the issue was raised about > changing the policy for applying verification methods to cable > connections. Here is our proposal on how to change this policy. > > First some background: A few years ago, the IANA asked the Regional > Registries to keep an eye on "very small assignments to > individuals". This means cases like static dial-up where a registry > has to use one IP address per dial-up customer. Since these kinds of > assignments could potentially use up a lot of address space, it was > decided to strongly discourage using static addresses for this and > that that we should monitor these assignments more closely. To do > this, any Local Registry that assigns a lot of addresses for this > purpose (more than 1024 addresses), was requested that they would have > to enter each of these customers in the database individually (as is > the case with all assignments to end-users). Later when some Local > Registries had a problem with entering all their dial-up customers in > the database for confidentiality problems, it was decided to add an > alternative method that they could instead send us a weekly report > showing the customers they added that week. This is so that the > Regional Registries could still keep track of these kinds of > assignments. > > In the last few years, more and more technologies have come along that > fall under this procedure. Aside from static dial-up, other services > such as web hosting (when not using http 1.1), cable connections, ADSL > connections, some ISDN connections, etc.. often need to assign one IP > address per customer. In many of these cases, using one IP address per > customer is necessary because the host is usually permanently > connected to the Internet. > > The issue was raised that applying these verification methods to > technologies that are permanently connected to the Internet and need > one IP address per customer, is too cumbersome and adds a lot of extra > administrative overhead for the provider. We also realise that there > are inconsistencies between the Regional Registries in dealing with > these technologies. > > For these reasons, we would like to change the policy and make a > distinction between technologies that are usually connected to the > Internet permanently > and those that are connected only temporarily. Here is the > proposal for > the new policy, please give us your comments by the end of next week. > > Permanently online: > > This includes technologies such as cable, ADSL, and various server > (web, ftp, etc..) connections that are usually permanently > connected to the Internet. It can also include analogue or digital > (ISDN) dial-up connections if they are used for a service that needs > to be permanently online and for services like dial-out on demand. > > For most of these connections, you need to assign one IP address per > connection, even if using DHCP (which we recommend because it makes > renumbering easier). We propose to change the policy and to not apply > the verification methods to these types of connections. > > For these technologies, we still need to keep track of the amounts of > addresses being assigned, however we would do this whenever the > customer requests a new assignment for this purpose. Since these > services are permanently connected to the Internet, we will probe > these addresses to verify whether the assigned addresses really are in > use. It must therefore be made clear to the RIPE NCC exactly which > block of IP addresses is used for this purpose. We would expect to see > that a high percentage are reachable. We might at this point request > additional information to show that they have really used the > addresses that were assigned for this purpose in the past- this could > be in the form of a list of all URLs assigned, their DNS records, > config files, etc.. something to show that the IP addresses are all in > use. > > Virtual web hosting is the only one that has an alternative available > (http 1.1) that would allow assigning one IP address for several > hosts. Since we want to encourage that people use http 1.1, we propose > to continue applying the special verification methods to virtual web > hosting that doesn't use http 1.1. > > Temporarily online: > > This includes technologies that only need to be connected to the > Internet temporarily, such as most telephone dial-up connections. We > recommend the use of DHCP and assigning IP addresses based on the > number of modem ports. We will therefore continue to apply the special > verification methods to any organisation that assigns IP addresses > based on the number of customers instead. > > > > > > > > ------- End of Forwarded Message > > >
[ lir-wg Archives ]