Network Registry FAQ for Net Admins (draft)

Questions and answers are being developed from the discussion on the Net-Admin-L list.

Also see CIT's Host Registration FAQ for related questions and answers.

1. Exactly what will happen on June 1st from the point of view of users who have not yet been registered? Will they still be able to function?
2. What is the difference between self-update and self-registration?
3. What tools are available for MAC/IP mapping?
4. Can the NOC get the jack number from the IP/MAC list?
5. How will visitors who don't have NetIDs be able to register?
6. Why are host names not required in the network registry?

1. Exactly what will happen on June 1st from the point of view of users who have not yet been registered? Will they still be able to function?

   On June 1, users should not notice any difference due to changes made by CIT. The changes are taking place on the database end to accommodate new functionality. One exception: On the 400 subnets that use CIT's DHCP service, some users may notice that the self-registration screens look slightly different than in the past, but the experience will be the same.

   In general, the netadmins of each subnet can decide how they wish to manage their subnets and comply with the new policy. In a timely fashion that takes into consideration both the policy and the needs of their department, netadmins may need to make changes and announce those changes to their users. But those changes do not need to occur by June 1.

2. What is the difference between self-update and self-registration?

   The user self-registration feature is only available to subnets that choose to use CIT's DHCP registration service. The user is automatically directed (via a restricted DNS) to a web page that registers them with DNS/DHCP and Network Registry as appropriate. See the sample self-registration forms.

   The user self-update feature is still being designed but will be available to subnets that use CIT's DHCP service and to other subnets that allow self-update. The user needs to visit the update web page (not yet available); if a DNS registration (IP address and host name) exists, the user self-update web page will detect and update the user's MAC address and NetID for that registration.

3. What tools are available for MAC/IP mapping?

   The web site http://berring.cit.cornell.edu/cgi-bin/welcome.cgi has some tools for netadmins. The "IP/MAC address tracking tools" link has several search mechanisms. Halfway down the page under "Search IP to MAC to switch port," you can type in the CIDR notation for your subnet, e.g. 128.253.180.0/24, and get the complete mapping (if your subnet is on CIT network switches).

   What is CIDR notation?

   Classless Inter-Domain Routing is a method for assigning IP addresses without using the standard IP address classes like Class A, Class B or Class C. In CIDR notation, a subnet's IP address is represented as A.B.C.D/n, where "/n" is called the IP prefix or network prefix. An IP address is 32 bits long, and the prefix specifies the number of bits used to identify a network. The remaining bits are used to identify hosts on the network. A prefix of /24 means there are 8 bits available to identify up to 256 hosts, while a prefix of /28 means there are 4 bits available to identify up to 16 hosts. For example, 192.9.205.22/18 means the first 18 bits are used to represent the network and the remaining 14 bits are used to identify hosts.

4. Can the NOC get the jack number from the IP/MAC list?

   Technically they can and do for security cases (if the MAC is seen on CIT network switches). But the process is manual and it takes significant time to process a list, so the NOC won't be able to do this for all subnets. One of the tools we are investigating is the possibility of providing online access to such data but that is still in the investigation stage.

5. How will visitors who don't have NetIDs be able to register?

   Visitors will be able to request a temporary guest code to use for the duration of their stay. The guest code can be entered in the network registry in place of a NetID. The procedure for requesting a guest code has not been established yet, but information about this is expected to be available soon, tentatively by mid-June.

6. Why are host names not required in the network registry?

   The policy does not specifically require host names because host names would be difficult to register on subnets using dynamic DNS. The network registry does have a field to record a host name, however, and for subnets using static IPs or static DHCP, supplying a host name is still considered a "best practice" and a de facto standard at Cornell.