Talk:Supernetwork

Made corrections to article

Author indicated supernetting requires link-state routing protocol such as RIP2 or OSPF. RIP is a distance-vector routing protocol. Supernetting only requires that contigious address blocks be used and a routing protocol which is capable of carrying subnet information in its updates (also added). Author also states "Supernetting is a networking technique which allows the use of multiple IP networks on the same network interface" which is false. Supernetting mainly allows contigious blocks to be assigned to an organization when larger blocks, such as class A or B ranges are unavailable. Instead multiple blocks of address, for instance class C, are assigned. It also allows for smaller internet routing tables as now only one entry need be placed in the upstream router representing the networks below it. This is my first Wikipedia contribution, I'll add more to this article once I learn the customs around here :-).

Supernet size

The author claims:

Supernets always have masks that are smaller than the classful mask, otherwise it isn't a supernet.

I am not sure this is the case. It is common, for instance, to create 16-bit aggregations in 10.0.0.0/8 address space, with 24-bit subnets beneath the 16-bit aggregations:

• 10.0.0.0/8
  • 10.1.0.0/16
    • 10.1.1.0/24
      • Hosts

I consider this use of 16-bit aggregation to be "supernetting", regardless of the fact that the 16-bit chunks have longer prefixes (larger masks) than the classful /8 prefix.

76.20.226.15 (talk) 21:58, 5 December 2008 (UTC)

Clay

No, this is just subnetting a subnet. Supernetting is just the act of breaking the Classful rules of IP addressing. Just because you have created a subnet beneath one of your subnets does not make it a supernet. I see where you are coming from though, but its not correct.Jwjkp (talk) 22:10, 5 December 2008 (UTC)

Want to add on this a little. Just think of it this way, if the bits for the network portion are larger than the classful bits for the IP, then its subnetting. If the bits used is less than the classful bits for the IP, its supernetting.

The only a class A address can be supernetted is if the mask is 1-7; for class B 1-15; and for class C 1-23. Jwjkp (talk) 01:35, 23 December 2008 (UTC)

Merge

I've merged Route summarization & Route aggregation into this article because I believe they are all the same thing, but this needs a cleanup & better examples.--DanielSiva (talk) 11:05, 4 April 2009 (UTC)

Supernet - summarized route

Hi wiki community,

I am studying inter networking and I was reading up on the SuperNet article. I came accross the line:

'The summarized route is 192.168.98.0/20. The subnet mask is 255.255.240.0.'

And I was confused and could not follow. Then I realized that in the next paragraph it is mentioned that the summary route could be modified to 192.168.98.0/20 to exclude certain networks, which I do understand, but I think it is incorrect to say that (from the summary table alone) the aggregated route would be 192.168.98.0 /20 but should be 192.168.96.0/20 in the first place. Can somebody please confirm this and maybe edit the article if I am right? Otherwise I appologize for being pedantic or simply wrong.

Regards

Michael K123

Here are some other examples/references I found that appear to indicate that the example in the article has problems: [1], [2], [3]. --Kvng (talk) 18:32, 20 February 2011 (UTC)

On dubious issue was resolved. I still wonder whether 192.168.98.0/20 is a legit route specification. I find no indication in the refs I provide above that it is. --Kvng (talk) 16:04, 23 February 2011 (UTC)

Why would it not be? I removed the new dubious tag because there is nothing dubious about it, the example route is logical and does what is explained, and therefore it is a valid example. CIDR works on the principles of vlsm and masks may be set on any bit. This is the flexibility of CIDR vs. classful addressing. Of course it is always the job of the administrator to make sure the configuration is practical. The problem here was that someone who didn't understand the example modified it so it appeared wrong. Kbrose (talk) 05:16, 3 March 2011 (UTC)

192.168.98.0/20 translates to 192.168.98.0 mask 255.255.240.0 or 11000000.10101000.01100010.00000000 mask 11111111.11111111.11110000.00000000. What looks dubious to me is that we've got a network bit set outside the mask. In normal context this is not a legit route. I can't find any evidence that it is legit in the context of supernetting. --Kvng (talk) 15:36, 5 March 2011 (UTC)

I agree with Kvng, who just beat me to editing this page :-) I believe that the statement "However, the summarized route may be modified to 192.168.98.0/20 to exclude the first two networks and the last, such that the first actually routed network is specified." is confusing at best.

1) 192.168.98.0/20 cannot be applied on any mainstream router. For example, trying to add this to a Juniper:

lab@lab-jnpr02# set routing-options static route 192.168.98.0/20 discard    

[edit]
lab@lab-jnpr02# top show | compare                                          
[edit routing-options static]
     route 10.0.0.0/8 { ... }
+    route 192.168.98.0/20 discard;

[edit]
lab@lab-jnpr02# commit check                                                
re0: 
[edit routing-options static]
  'route 192.168.98.0/20'
    RT: DEST: 192.168.98.0 MASK: 255.255.240.0 mask too short
error: configuration check-out failed

[edit]
lab@lab-jnpr02# 

2) The phrase "modified ... to exclude the first two networks and the last" raises some questions.

Decimal notation	Binary form
192.168.97.0	11000000101010000110000100000000
192.168.98.0	11000000101010000110001000000000
192.168.99.0	11000000101010000110001100000000
192.168.109.0	11000000101010000110110100000000
192.168.110.0	11000000101010000110111000000000
192.168.111.0	11000000101010000110111100000000

The claim is that "192.168.98.0/20 means that 192.168.99.0/24 is included, but 192.168.97.0/24 is excluded. Also 192.168.110.0/24 is included but 192.168.111.0/24 is excluded, simply by setting the most significant host bits to "001" when referring to the prefix. It is not at all clear how this happens. If this scheme is correct, what range of addresses would 12.1.3.0/8 correspond to?

This is, at best, a corner case on some non-mainstream implementation, and so I think the sentence should be deleted to avoid confusion. - Ttwaring (talk) 16:25, 5 March 2011 (UTC)

No offense but did the meaning of the term "route" change since I last studied it? I would have called 192.168.98.0/20 a network as opposed to 192.168.98.1 which might very well be a route. Thus the term would be summarized network Am I wrong? Just curious Xp fun (talk) 18:45, 4 October 2012 (UTC)

It's colloquially called a summary route as you would configure a router with a command like

ip route 192.168.96.0/20  2.2.2.2 Ethernet1/1

to send it out on a particular interface. Note that 192.168.98.0/20 doesn't make any particular sense as a network, as described above. The global routing table is full of things like 4.0.0.0/8 and 207.254.128.0/21, which are interchangeably referred to as networks, routes or prefixes. Probably your teacher was more careful in his or her wording. However, I would call 192.168.98.1 a host address, unless it's floating round in a routing table as 192.168.98.1/32.

[You could configure a host with host address 192.168.98.0 and give it a netmask of /20, so the network address is 192.168.96.0 and the broadcast address is 192.168.111.255, but this is not what the original "98" guy was talking about.]

Ttwaring (talk) 21:50, 26 January 2013 (UTC)

Advertise lowest address or 0-padded common bits?

The example has this as the lowest address of the networks that are being supernetted:

Address	First Octet	Second Octet	Third Octet	Fourth Octet
192.168.98.0	11000000	10101000	01100010	00000000

And then goes on to say:

The summary route should be the lowest IP address, followed by a slash
...
The summarized route is 192.168.96.0/20.

But the lowest network address is 192.168.98.0. The address formed by taking the common bits and 0-padding on the right is the one that was given, 192.168.96.0

So it's inconsistent and I'm not sure which is right. — Preceding unsigned comment added by 151.151.109.15 (talk) 23:16, 7 December 2012 (UTC)

0-padded is correct. I've taken out the misleading "lowest IP address" terminology. - Ttwaring (talk) 21:27, 26 January 2013 (UTC)