QUESTION: what makes multiple hostnames possible?
S Claus
here is a perhaps dumb question about hostnames on a company network.
At this place where I work we have servers that can be pinged using:
ping SMTAPPU01
...yet the ping starts with the statement:
Pinging ldncsr0988.intra... [nn.nnn.n.n]
And I can also using:
ping ldncsr0988
which gives the same reply.
My question about this is, is it NAT that is used to have a server
have two hostnames?
And another question is what is the benefit of this? Is there any
benefit other than being able to use more than one hostname?
Thanks in advance,
Argon
> Hi all,
>
> here is a perhaps dumb question about hostnames on a company network.
>
> At this place where I work we have servers that can be pinged using:
>
> ping SMTAPPU01
SMTAPPU01 is the windows name
> ...yet the ping starts with the statement:
> Pinging ldncsr0988.intra... [nn.nnn.n.n]
>
> And I can also using:
>
> ping ldncsr0988
ldncsr0988.intra... is the DNS name of the same machine
> which gives the same reply.
>
> My question about this is, is it NAT that is used to have a server
> have two hostnames?
Nope. It is that windows networks do not use DNS as their native name
resolution system.
> And another question is what is the benefit of this? Is there any
> benefit other than being able to use more than one hostname?
M$ machines may be used in IP networks. Remember Bill Gates saying
something like "the internet is just a intermediate thing" or
something like this ... Correct me.
Jorgen Grahn
On Wed, 12 Aug 2009 00:58:15 -0700 (PDT), S Claus <sa...@temporaryinbox.com> wrote:
> Hi all,
>
> here is a perhaps dumb question about hostnames on a company network.
>
> At this place where I work we have servers that can be pinged using:
>
> ping SMTAPPU01
>
> ...yet the ping starts with the statement:
> Pinging ldncsr0988.intra... [nn.nnn.n.n]
>
> And I can also using:
>
> ping ldncsr0988
>
> which gives the same reply.
>
> My question about this is, is it NAT that is used to have a server
> have two hostnames?
No, it has nothing to do with NAT. It's caused by either DNS+whatever
Windows uses, or pure DNS. Each IP address has one or more names, and
one of them is the canonical name. Ping shows you the canonical name,
just to be helpful.
The Wikipedia article on DNS gives more details.
> And another question is what is the benefit of this? Is there any
> benefit other than being able to use more than one hostname?
Yes. You can have one set of names on your machines which are fixed
(refer to the actual machine) and one set which refers to the services
they provide. Like 'www', 'mail', 'pop' ... then you can move these
latter names around along with the actual services, and the users do
not have to care.
I am sure there are other uses too.
/Jorgen
--
// Jorgen Grahn <grahn@ Oo o. . .
\X/ snipabacken.se> O o .
Barry Margolin
<477d448f-55e3-4454...@i18g2000pro.googlegroups.com>,
S Claus <sa...@temporaryinbox.com> wrote:
> Hi all,
>
> here is a perhaps dumb question about hostnames on a company network.
>
> At this place where I work we have servers that can be pinged using:
>
> ping SMTAPPU01
>
> ...yet the ping starts with the statement:
> Pinging ldncsr0988.intra... [nn.nnn.n.n]
>
> And I can also using:
>
> ping ldncsr0988
>
> which gives the same reply.
>
> My question about this is, is it NAT that is used to have a server
> have two hostnames?
It's the same way that two entries in the phone book can have the same
phone number.
> And another question is what is the benefit of this? Is there any
> benefit other than being able to use more than one hostname?
If a machine is used for multiple services, it can be useful to give it
names that reflect each service. So if it's both a mail server and a
web server, it's more mnemonic to refer to them by the names
mail.company.com and www.company.com.
--
Barry Margolin, bar...@alum.mit.edu
Arlington, MA
*** PLEASE don't copy me on replies, I'll read them in the group ***
Martijn Lievaart
> Hi all,
>
> here is a perhaps dumb question about hostnames on a company network.
>
> At this place where I work we have servers that can be pinged using:
>
> ping SMTAPPU01
>
> ...yet the ping starts with the statement: Pinging ldncsr0988.intra...
> [nn.nnn.n.n]
>
> And I can also using:
>
> ping ldncsr0988
>
> which gives the same reply.
>
> My question about this is, is it NAT that is used to have a server have
> two hostnames?
No, it's simple. There is DNS to map a name to an IP address. Multiple
names can legitimately map to one IP address.
Ping then uses reverse DNS to look up the name belonging to that IP
adress and shows that name.
These mappings are completely unrelated, it's up to the admin to make
sure that what reverse DNS returns, indeed is the canonical name for that
machine, but if misconfigured, it might return a completely different
name.
To make it a bit more complex, it is completely legitimate to reverse map
an IP address to multiple names. In fact, some argue that this was how
DNS was envisioned. But don't do that, it probably confuses a lot of
software and no one does it anyhow. Probably ping would still print the
first name in the list of names it gets back for that IP address, but
that is just guessing.
HTH,
M4
A_A_K
Go through the DNS concept. Main advantage with host name is to make
you simple to remember.
For example, take google.com, you can easily remember the name but u
cant remeber the ip address, and if you want some other webserver's IP
address you cant remember. If you have the name of that server it is
easy for us to remember.
DNS maps the NAME <--> ip address.
Some more points:
What is DNS?
-DNS is a host name to IP address translation service
-DNS is
a distributed database implemented in a hierarchy of name
servers
an application level protocol for message exchange between
clients and servers
Why DNS?
-It is easier to remember a host name than it is to remember an IP
address.
-An name has more meaning to a user than a 4 byte number.
-Applications such as FTP, HTTP, email, etc., all require the user to
input a destination.
-The user generally enters a host name.
-The application takes the host name supplied by the user and forwards
it to DNS for
translation to an IP address.
DNS Services:
-Besides the address translation service, DNS also provides the
following services:
Host aliasing: a host with a complicated name can have one or more
aliases that are simpler to remember,
e.g., relay1.west-coast.media.com -> media.com.
The longer name is the canonical hostname, the shorter the alias
hostname.
-Mail server aliasing: same as above, aliases can exist for long
canonical host names.
Load Balancing: a set of servers can have one name mapped onto
several machines.
DNS provides the full list of names to the
end user’s application which generally takes the first one in the
list.
DNS rotates the names on the list.
How does it work?
-DNS works by exchanging messages between client and server machines.
-A client application will pass the destination host name to the DNS
process (in Unix referred to as the gethostbyname() routine) to get
the IP address.
-The application then sits and waits for the response to return.
If you have any queries regarding networking feel free to ask me
(401...@gmail.com)
-Anil
glen herrmannsfeldt
< In article
< <477d448f-55e3-4454...@i18g2000pro.googlegroups.com>,
< S Claus <sa...@temporaryinbox.com> wrote:
<> here is a perhaps dumb question about hostnames on a company network.
<> My question about this is, is it NAT that is used to have a server
<> have two hostnames?
< It's the same way that two entries in the phone book can have
< the same phone number.
Not quite, as phone books don't have CNAME.
Some phone book entries, instead of giving the number will
instead say "See ..." referencing the appropriate entry,
somewhat like CNAME.
If I remember the OP, the name displayed was not always the same
name entered on the command line.
-- glen
glen herrmannsfeldt
(snip)
< No, it's simple. There is DNS to map a name to an IP address. Multiple
< names can legitimately map to one IP address.
< Ping then uses reverse DNS to look up the name belonging to that IP
< adress and shows that name.
It could do that, but I believe more use the name returned when
the orignal name is a CNAME record.
< These mappings are completely unrelated, it's up to the admin to make
< sure that what reverse DNS returns, indeed is the canonical name for that
< machine, but if misconfigured, it might return a completely different
< name.
< To make it a bit more complex, it is completely legitimate to reverse map
< an IP address to multiple names. In fact, some argue that this was how
< DNS was envisioned. But don't do that, it probably confuses a lot of
< software and no one does it anyhow. Probably ping would still print the
< first name in the list of names it gets back for that IP address, but
< that is just guessing.
Better to have one official name with both A and PTR records,
all the other names CNAME to that name. Even with more than one
A record, it works best to have only one PTR.
-- glen
Bruce Cook
> Better to have one official name with both A and PTR records,
> all the other names CNAME to that name. Even with more than one
> A record, it works best to have only one PTR.
Except of course when the Name is being used as an MX destination, then it
must be an A record.
// The following is invalid and will result in most MTAs ignoring the target
// quite a few name servers will also give a warning if they see this
foo in a 1.2.3.4
mail in cname a
@ in mx 10 mail
// the proper way of doing it is:
foo in a 1.2.3.4
mail in a 1.2.3.4
@ in mx 10 mail
There are other situations where cnames won't work (eg NS records) so you
should be careful when using them.
Bruce