From: Kenneth Wygand (KWygand@customonline.com)
Date: Sat Jul 03 2004 - 15:07:32 GMT-3
Well, Dan Shechter got it... all of the following addresses are valid on point-to-point links using a /31 mask as per RFC 3021. Cisco supports this as of IOS 12.2(2)T.
1.0.0.0 /31
1.255.255.0 /31
1.0.0.255 /31
10.255.255.255 /31
Essentially, point-to-point links don't need to know of a "network" or "broadcast" address, because there is only a single host on the remote side. As a result, a /31 allows only two addresses in the host portion. By process of elimination, each host knows what the address is of the host on the remote side.
For more details, check out RFC 3021 at the following link:
http://www.faqs.org/rfcs/rfc3021.html
Stay tuned for the next question of the day... :)
Ken
________________________________
From: Kenneth Wygand
Sent: Fri 7/2/2004 5:44 PM
To: Tom Martin; ccielab@groupstudy.com
Subject: Quiz question of the day 20040702 (Hints)
Tom,
The 0.x.x.x and 255.x.x.x addresses -theoretically- are valid IP addresses, provided IOS were designed to parse through addresses that do not conform to the inherent "rules" we as engineers have designed into the TCP/IP protocol. So for the sake of our discussion, I would say these addresses are always invalid.
Now that leaves us with the following addresses:
1.1.1.1
1.0.0.0
1.255.255.0
1.0.0.255
10.255.255.255
We know 1.1.1.1 is valid with any subnet mask value, so let's ignore that one. That leaves us with:
1.0.0.0
1.255.255.0
1.0.0.255
10.255.255.255
Now, we've already mentioned that these addresses can be valid with an assortment of subnet mask lengths as you originally listed, as well as the /32 routes that you can easily apply as loopback interfaces. However, now we specified the restriction is imposed that each subnet mask must be at least 24 bits with the /32 excluded.
This means that we can only use values in the last octet to indicate a "host" portion of our IP addresses. The binary value in the last octet of each IP address listed is as follows:
1.0.0.0 = 00000000
1.255.255.0 = 00000000
1.0.0.255 = 11111111
10.255.255.255 = 11111111
So no matter where we choose to specify the host portion, we will have all 0's or all 1's in that field. Conventional thinking indicates these addresses must be either the network address or broadcast address for the particular network portion.
But... does an all 0's or all 1's in the host portion -HAVE- to be a network address or broadcast address?? ;-)
Hint Hint... IOS 12.2(2)T...
Anyone??? ;-)
Oh, and by the way, I include the date in the subject for those that sort by subject and/or particular threads. I have not compiled an archive of these "question of the day" emails, but would consider doing so if I receive enough requests to do so.
Ken
________________________________
From: Tom Martin [mailto:tig@wiltecinc.com]
Sent: Fri 7/2/2004 5:09 PM
To: ccielab@groupstudy.com
Cc: Kenneth Wygand
Subject: RE: Quiz question of the day 20040702 (this one's for you Jamie!)
Okay, then I'd have to go with only one valid IP: 1.1.1.1
Now you've got me curious as to the point you're making... Am I wrong?
BTW, I just noticed that you have your question of the day identified
with a unique date identifier. Is there a web site (or other archive)
where you have previous questions posted?
-- Tom
-----Original Message-----
From: Kenneth Wygand [mailto:KWygand@customonline.com]
Sent: Friday, July 02, 2004 4:53 PM
To: Tom Martin
Cc: ccielab@groupstudy.com
Subject: RE: Quiz question of the day 20040702 (this one's for you
Jamie!)
OK, Tom. Now let's say single host assignments (/32 subnets) are not
allowed.
Now what... ?? Are your choices still valid?
(There is a point here...)
Ken
________________________________
From: Tom Martin [mailto:tig@wiltecinc.com]
Sent: Fri 7/2/2004 4:25 PM
To: Kenneth Wygand
Cc: ccielab@groupstudy.com
Subject: RE: Quiz question of the day 20040702 (this one's for you
Jamie!)
Nope. All of the addresses (with exception of 0.X.X.X and 255.X.X.X)
remain valid with a 32-bit subnet mask. I am testing on a loopback after
all... :)
-- Tom
-----Original Message-----
From: Kenneth Wygand [mailto:KWygand@customonline.com]
Sent: Friday, July 02, 2004 3:12 PM
To: Tom Martin
Cc: ccielab@groupstudy.com
Subject: RE: Quiz question of the day 20040702 (this one's for you
Jamie!)
Tom,
Consider the following requirement to put a spin on things: "There can
exist no subnet mask less than 24 bits".
Now what do you think? Does your answer change?
Ken
________________________________
From: Tom Martin [mailto:tig@wiltecinc.com]
Sent: Fri 7/2/2004 2:30 PM
To: Kenneth Wygand
Cc: ccielab@groupstudy.com
Subject: RE: Quiz question of the day 20040702 (this one's for you
Jamie!)
On a Cisco router the following work (tested on a loopback interface).
I've also included a subnet mask since the "default" classful subnet
mask won't necessarily work.
1.1.1.1 255.0.0.0
1.0.0.0 255.255.255.255
1.255.255.0 255.0.0.0
1.0.0.255 255.0.0.0
10.255.255.255 248.0.0.0
These don't work for obvious reasons:
0.45.34.23
255.223.234.24
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
Kenneth Wygand
Sent: Friday, July 02, 2004 1:59 PM
To: ccielab@groupstudy.com
Subject: Quiz question of the day 20040702 (this one's for you Jamie!)
Quiz question for the day:
How many of the following are possible valid, addressable IPv4
addresses:
1.1.1.1
0.45.34.23
1.0.0.0
1.255.255.0
1.0.0.255
255.223.234.24
10.255.255.255
Good luck to all!
Ken
This archive was generated by hypermail 2.1.4 : Sun Aug 01 2004 - 10:11:46 GMT-3