We all know what is used for (loopback).

What are uses cases for the rest of the reserved loopback space?

  • 1
    Yes, it's used for loopback too. Jun 28, 2012 at 4:34

4 Answers 4


It's also reserved for loopback, so no, it's not widely used for anything.

In practice, is usually used as "the" loopback address, but the rest of the block should loopback as well, meaning it's just generally not used for anything. (Though, for example, larger Cisco switches will use 127.0.0.xx IPs to listen for attached cards and modules, so at least some of other addresses are in use.)

From RFC3330: Special-Use IPv4 addresses - This block is assigned for use as the Internet host loopback address. A datagram sent by a higher level protocol to an address anywhere within this block should loop back inside the host. This is ordinarily implemented using only for loopback, but no addresses within this block should ever appear on any network anywhere [RFC1700, page 5].


In practice, I have seen other addresses used in two places:

  1. As responses to DNSRBL lookups. Different responses can encode the reason the IP address (or domain) was listed. Wikipedia has some details, as does RFC5782. SORBS list their return codes. Project Honeypot encode data in the three available octets.
  2. In the Ubuntu /etc/hosts file. I don't remember the details but there was a conflict so they added another localhost-something entry with a different IP address in the range. It's and it's a bug workaround.
  • As for #1, that is probably related to the fact that DNS RBLs have to list something at each name in use (otherwise it doesn't exist), and A records are fairly small. There's nothing magical about using 127/8 for such purposes, it's just guaranteed to not be used for anything real of significance that cannot easily be detected immediately.
    – user
    Jun 28, 2012 at 9:04
  • The range is used in DNSBLs specifically so that it doesn't cause inadvertent and unwanted network traffic in the case that the returned IP address is actually used.
    – Ladadadada
    Jun 28, 2012 at 9:28
  • Which could also be achieved by using, for example, The downside of that is that you wouldn't be able to differentiate results based on resolved-to IP address, but for that I'm not sure what is the egg and what's the omelette.
    – user
    Jun 28, 2012 at 9:30
  • is not a bug workaround, it's a workaround for inability to specify DNS port in resolv.conf, so dnsmasq uses some unusual loopback IP address to avoid conflicts with other DNS servers taking socket.
    – darkk
    Aug 14, 2017 at 10:33

As already stated whole block is used as loopback so i'm only adding one example for regular desktop use.

Loopback other than is required if you want to secure RDP, or some other restricted connection, with local proxy. For example, using RDP through SSH tunnel requires that you setup local side for port forwarder to listen on This is because RDP client that comes with Windows refuses to connect to localhost or

That's right, normally you would not connect RDP client to same computer that you are using (and not allowed to do so even if wanted to see nice mirror effects :).

  • You gave me an idea, but it didn't work. I tried to connect to myself using, but it gives the same access denied error as if I used
    – Steven Lu
    Feb 22, 2015 at 15:21
  • @StevenLu With ip addresses that ends in zero like you could always expect more or less strange problems. Maybe try It could be that modern windows systems will recognize that you are logging in from same machine but it still should try to connect you first. Feb 22, 2015 at 17:19

Adding to to the other answers:

There are use cases, for example in development and testing. Instead of creating dedicated networks or interfaces sometimes it can be simpler just to use some unused addresses in the range.

You could spin up a development "server" listening on and develop a client that connects to This works out of the box and all these are separate addresses, so there could be another server process listening on and they would not interfere.

This works on linux, I am not sure about other operating systems.

As an example create a "server" process:

sudo mknod -m 777 fifo1 p
cat fifo1 | netcat -l -k 1234 > fifo1

In another terminal on the same host:

$ netstat -tulpn|grep 1234 # check that server is listening:

tcp        0      0*               LISTEN      28043/netcat        
$ echo abc | netcat 1234 # wrong ip, no result

$ echo abc | netcat 1234 # matching ip, receives echo:

For completeness, you could remove fifo1 after stopping the server from the first terminal, using the rm command.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .