In article <(E-Mail Removed) .com>,
(E-Mail Removed) wrote:
>I understand that with Zeroconf there is no need to do any type of
>configuration before adding a device on a network.
Well, not exactly.
http://www.ietf.org/internet-drafts/...nklocal-17.txt
Abstract
To participate in wide-area IP networking, a host needs to be
configured with IP addresses for its interfaces, either manually by
the user or automatically from a source on the network such as a DHCP
server. Unfortunately, such address configuration information may not
always be available. It is therefore beneficial for a host to be able
to depend on a useful subset of IP networking functions even when no
address configuration is available. This document describes how a
host may automatically configure an interface with an IPv4 address
within the 169.254/16 prefix that is valid for communication with
other devices connected to the same physical (or logical) link.
IPv4 Link-Local addresses are not suitable for communication with
devices not directly connected to the same physical (or logical)
link, and are only used where stable, routable addresses are not
available (such as on ad hoc or isolated networks). This document
does not recommend that IPv4 Link-Local addresses and routable
addresses be configured simultaneously on the same interface.
Now, translating that out of marketing talk, microsoft has always tried
to make computers easy to use. If security gets in the way, that's not
something to worry about. So the idea of 'plug it in, and it works' was
introduced in windoze for workgroups (win3.11) and carried forward when they
invented TCP/IP or something.. What this means is that the burden of
configuring the computer was transferred from the workstation to the server.
The minor little fly in this ointment is that a proper DHCP configuration
does take a bit of understanding. Microsoft realized this was lacking in the
average person who was configuring their systems, so in windoze98, they
added this 'zero-conf' service. Briefly, if a system can't find a DHCP server
for any reason (network down, servers misconfigured, you name it), the host
will reach up between it's legs and pull an address out of there. It will
then gracious-arp (is anyone using this address?) and if no one complains
it will use that (otherwise, it will try a different random number).
The 'draft-ietf-zeroconf-ipv4-linklocal-17.txt' is as the number infers, the
17th try to get the internet community to adopt this concept. Drafts expire
after six months (this draft expired in early January), so microsoft has only
been trying for eight years to get this adopted as any form of internet
standard. So far, no success. DHCP, and it's predecessor BOOTP were developed
in 1995 so that organizations with a limited number of IP addresses and a lot
of computers that were not all on at the same time to dynamically reuse the
limited number of addresses.
This brings up a minor problem - people normally use hostnames, rather than
IP addresses (and there are 65534 different addresses that this protocol
might randomly use). So how to you tie names to IP addresses? DNS is built
around the concept of a fixed name to address relationship. Now, you can give
fixed names to addresses, but we still don't know which computer is the one
we want to talk to. There are a number of ways around this in DHCP such as
allowing the client to ask for a specific address, or by assigning the IP
address based on the MAC address (both of which destroy the Dynamic aspect
of DHCP). Another way is the microsoft concept (adopted from their earlier
NETBEUI protocol) of allowing the clients to announce their name, and
believing that. ("I am the President of the USA" "OK, I'm Bill Gates and
I'm pleased to meet you.) Perhaps that might be a major security problem,
but who cares - it's easy to use.
In fairness to microsoft, all of these protocols are designed not to pass
a router, so everything should be local. If you trust everyone who is going
to connect to your network, there really shouldn't be much of a security
problem. If you don't trust them, why are you allowing them on your net?
Section 5 of the draft does admit that using this "may open a network host
to new attacks." - a typical microsoft understatement of a problem.
>Like, a printer can be just plugged in and used. Am I correct?
Not in *nix. You have to somehow configure the print server.
>Is there a Domain server in a Zeroconf network?
Prohibited on page 7 of the draft noted above.
>If there is none then how does the machines on the network come to know
>about existence of a printer?
There is a another draft proposal (mentioned in the draft above) called
http://www.ietf.org/internet-drafts/...xt-mdns-38.txt (notice
that this is the 38th revision).
Abstract
Today, with the rise of home networking, there are an increasing
number of ad-hoc networks operating without a Domain Name System
(DNS) server. The goal of Link-Local Multicast Name Resolution
(LLMNR) is to enable name resolution in scenarios in which
conventional DNS name resolution is not possible. LLMNR supports all
current and future DNS formats, types and classes, while operating on
a separate port from DNS, and with a distinct resolver cache. Since
LLMNR only operates on the local link, it cannot be considered a
substitute for DNS.
HOWEVER, this is only name resolution. The document says
Service discovery in general, as well as discovery of DNS servers
using LLMNR in particular, is outside of the scope of this document,
as is name resolution over non-multicast capable media.
>What is tmdns? There is so little documentation on this.
Mandrake indicates that "Tmdns is tiny/trivial Multicast DNS Responder for
Linux. It should allow you to take part in a zeroconf environment." and
says it's from
http://zeroconf.sourceforge.net/
Old guy
Similar Threads
Linear Mode
