Dynamic DNS

Windows NT 4.0 includes two name-to-IP-address mapping services—WINS and DNS. One key difference between the two is that whereas WINS accommodates dynamic registration of NetBIOS names (and associated IP addresses), DNS names (and associated IP addresses) must be statically entered into the DNS service database.

Static registration of name-to-IP-address information is undesirable in the Windows NT/Windows 95 operating systems. This is because in all but very small Microsoft network installations, machines do not have static IP address assignments. The machines use the Dynamic Host Configuration Protocol (DHCP) to obtain an IP address assignment each time they initialize ("boot up").

A proposal for dynamic registration of DNS information is under consideration by the IETF (Internet Engineering Task Force). You can download the specification from "http://ds.internic.net/internet-drafts/draft-ietf-dnsind-dynDNS-09.txt". With DNS "dynamic update," a client machine, having obtained its assigned IP address from DHCP, can use a standard protocol to dynamically register its DNS name and IP address in the DNS database. In the Windows NT 4.0 time-frame, it was determined that "dynamic update" should not be implemented due to the fluid status of the specification, and scalability concerns. The current dynamic DNS proposal is based on a "pull from single master" replication model and consequently, if the single master is down or unreachable, dynamic updates cannot occur. Microsoft favors a multiple master arrangement similar to WINS that would allow registrations to continue without a single point of failure.

In the long run, using DNS dynamic update is more desirable than the
Windows NT 4.0 DNS-WINS integration for the following reasons:

DNS-WINS integration does not enable efficient DNS "reverse lookup" (resolution of IP address to DNS name). Reverse lookup is used for security purposes by Internet WWW (World Wide Web) and Firewall services. Due to the recent proliferation of such services, the need for efficient DNS reverse lookup is significantly increasing. If DNS dynamic update was used instead of WINS integration, this problem would no longer exist.
DNS-WINS integration does not enable a proper "primary-backup" relationship between Microsoft and non-Microsoft DNS name servers. This is due to the fact that non-Microsoft DNS name servers are not capable of WINS lookup. For migration purposes, customers want to install Microsoft DNS servers as backups of non-Microsoft DNS primary servers. If DNS dynamic update was used instead of WINS integration, this problem would no longer exist.
Non-Microsoft hosts do not register in WINS, and therefore can not "dynamically register" in DNS. DNS dynamic update is an IETF standard. If implemented (by Microsoft), non-Microsoft hosts [that support the DNS dynamic update standard] could dynamically register in the Microsoft DNS, and Microsoft hosts could dynamically register in a non-Microsoft DNS [that supports the DNS dynamic update standard].
WINS registration is not secure, and has no reasonable means of becoming secure. The IETF is working to complete a standard for adding security to DNS dynamic update. The DNS security standard is not scheduled to enter IETF standards track until the fourth quarter of 1996. Some companies have already decided to forego a standard and "rolled their own" security implementation.

	Microsoft-based clients can not register with any of the current versions of Dynamic DNS, and dynamic DNS servers can not replicate their dynamic data to other NON-Dynamic DNS servers.
	In current implementations of Dynamic DNS, the primary is a single point of failure. All of the clients must register their names and IP addresses with this machine. If this machine is down then no updates to the DNS database can occur.