 |
|||
     |
|||
|
Annex Class C Supernet SupportLast Modified: April 16, 1996
1. IntroductionThe Annex Product Line will support Supernetting of Class C internet addresses in revision X10.1.3 and a post Release of Mckinley2. Supernetting, will allow an annex to be configured with a subnet mask shorter than the intrinsic subnet mask( i.e. 255.255.255.0) for a Class C address. With Annex support of supernetting, a Class C subnet mask may be set to a range of 255.255.0.0 to 255.255.255.252, excluding 255.255.255.128. With this new functionality, a block of class C addresses can be addressed as a "domain" or a single destination address with more than 254 hosts. The reasoning for this is the exhaustion of the class B network address space, and the allowance of networks to have more than 254 host nodes with multiple Class C blocks being referenced as if it is a single "subnet". This also allows Internet Service Providers (ISP's) to assign a client a block of contiguous IP addresses and identify them as a single "network domain". These contiguous blocks are usually of sizes that are based on the power of 2. For instance it is very logical to only allow sizes of 1, 2, 4, 8, 16, 64, and 128 (etc.), depending on what subnet mask is to be derived. On a public network, this is very important for having unique subnet ID's throughout the internet. It is the goal of this document to explain the concept of supernetting of the C address and provide examples of how it is used. It is not the goal of this document to explain the concept of CIDR (not to say that Annex does not support CIDR), or the ability to supernet Class B or A networks (which Annex does not support). There are references to CIDR in this document for explanation and reference only.
2. References
3. Annex Configuration for SupernettingWith proper configuration, the Annex will create a supernetted route and place it into the routing table. Only RIP Version 2 will support supernetting. Please consult the Annex Network Administrator's Guide for setting up RIP Version 2 Routing.
3.1. ROM ConfigurationAll ROMS 0803+ allow configuration of class C Subnet which is up to one byte smaller than its intrinsic subnet value. The intrinsic subnet value for class C is 255.255.255.0. Therefore a range of 255.255.0.0 to 255.255.255.252 (excluding 255.255.255.128) are legal and is permitted via ROM Monitor configuration (i.e. addr command). These ROMS do not allow the relative configurations for Class B or A. For example, a subnet address of 255.0.0.0. is not permitted for Class B, and a class A subnet address less than 255.0.0.0 is not allowed. The ROMS do allow the subnet mask of 0.0.0.0, which will default to intrinsic masks for A (255.0.0.0), B ( 255.255.0.0) and C(255.255.255.0) at boot time.
3.2. Administration ConfigurationThe Subnet Mask can be set using host na (Network Admin. program), or admin. within the annex. Note that the selecting of the subnet mask is not as strict as the ROM Monitor Configuration. After rebooting, it is necessary to check for the actual subnet mask via a stats command, as well as well as looking at the routing table (netstat -r) NOTE: (More explanation will be provided in later sections).
4. How Supernetting Works
4.1. RIP 2 Table EntryA RIP 2 Table Entry contains the following representation when displaying an IP net address:
---------------------------------------------------
|BYTE1.BYTE2.BYTE3.BYTE4/{Subnet Mask # of 1 bits}|
|-------------------------------------------------|
|IP Address | Subnet Mask Bit Count |
---------------------------------------------------
The Subnet mask bit count indicates the consecutive 1 bits used for the network address in the table. Using the intrinsic mask for a class C network of 255.255.255.0, the table entry will contain 24 as the mask count. To set up a supernet for instance, using the subnet mask of 255.255.254.0 yields a count of 23. NOTE: with a subnet of 255.255.254.0, a class C network can now have (256*2)-2 host nodes or 510 total hosts, instead of the default of 254.
4.2. Available Host ID AddressesWhen calculating the number of host Id's the formula (as stated in the last section is the following: The subtraction of 2 is for the designation of Host ID of all 0's (0) (network address) and a Host ID of all 1 bits (or Hex F) (broadcast address). This deduces to the common IP address rule for all addresses, which is a host address with all 1 bits or all 0 bits are illegal and therefore cannot be assigned to any host node. This brings up an added addressing issue when suppernetting two or more "predefined" actual subnets. For instance, prior to supernetting subnets 192.73.216.0 and 192.73.217.0, host addresses 192.73.216.0, 192.73.216.255, 192.73.217.0, and 192.73.217.255 are all invalid host addresses. Upon supernetting the 216 and 217 addresses together, the dot decimal displayed addresses of 192.73.216.255 and 192.73.217.0 are now legal and can be assigned to hosts as long as the subnet mask is set correctly. The network address for both nets will be 192.73.216.0/23. This pattern will continue for each consecutive subnet in the supernet that is supported using the single subnet mask. Therefore the first and last host address of the entire subnet are the only two (2) invalid addresses. NOTE: Actual host addresses as all other IP addresses are displayed in dot decimal form. The act of supernetting and subnetting, if the subnet bit count is not a multiple of 8, the dot decimal representation will be different that the actual masked address. The dot decimal notation separates each 8 bits (number) by a dot. So in the paragraph above, the user would actually assign nodes 192.73.216.255 and 192.73.217.0 to hosts that will reside on the 192.73.216.0/23 subnet.
4.3. Supernet Routing IllustrationThe following illustrations(Figures 4.1, 4.2, and 4.3) show how subnet masks can be altered for supernetting. The concept of the subnet mask moves the NetworkID/HostID area of network entries routing entries to actually "alter" the subnet domains where host node ID's can be addressed. NOTE: The actual host IP addresses are unique and static, and remain that way. Their subnet masks in conjunction with their actual network ID's create the network "domain" or a "logical" network ID in the routing table. This logical Network ID can now be used for routing to nodes where the actual network ID's are different. This is accomplished by all of the nodes in the "logical" subnet having the same subnet mask. These masks in conjunction with the actual network ID can calculate the same "logical" Network ID. Figure 4.2. illustrates the concept of the new "altered" "logical" Network ID "domain".
Figure 4.1 Subnet Mask 255.255.255.0 (Supports 254 hosts).
+-----------------------------------+
|194 |73 |217 |1 | Host IP Address.
+-----------------------------------+
|255 |255 |255 |0 | Subnet Mask
+-----------------------------------+
|11111111|11111111|11111111|00000000| Subnet Mask (Binary (24 bits))
+-----------------------------------+
| Network ID | HostID | Net/HostID.
+-----------------------------------+
|194 |73 |217 |0 | Displayed Network ID.
+-----------------------------------+
|194 |73 |217 |1 | Displayed Host ID Address.
+-----------------------------------+
|194 |73 |217 |1 | Network domain host address.
+-----------------------------------+
Figure 4.2 Subnet Mask 255.255.254.0 (Supports 510 hosts).
+-----------------------------------+
|194 |73 |217 |1 | IP ADDRESS
+-----------------------------------+
|255 |255 |254 |0 | Subnet Mask
+-----------------------------------+
|11111111|11111111|11111110|00000000| Subnet Mask (Binary (23 bits))
+-----------------------------------+
| Network ID | HostID | Net/HostID.
+-----------------------------------+
|194 |73 |216* |0 | Displayed Network ID.
+-----------------------------------+
|194 |73 |217 |1 | Displayed Host ID Address.
+-----------------------------------+
|194 |73 |216 |257 | Supernet domain host address.
+-----------------------------------+
Figure 4.3 Subnet Mask 255.255.255.192 (Supports 62 hosts).
+-----------------------------------+
|194 |73 |217 |1 | IP ADDRESS
+-----------------------------------+
|255 |255 |255 |192 | Subnet Mask
+-----------------------------------+
|11111111|11111111|11111111|11000000| Subnet Mask (Binary (26 bits))
+-----------------------------------+
| Network ID |HostID| Net/HostID.
+-----------------------------------+
|194 |73 |217 |1# | Display Network ID.
+-----------------------------------+
|194 |73 |217 |129 | Displayed Host ID Address.
+-----------------------------------+
|194 |73 |217 |1|1 | Subnet domain host address.
+-----------------------------------+
As illustrated, Figure 4.1 shows a default (24 bit subnet mask) for a
Class C Network.
Figure 4.2 Shows a Supernetted (23 bit subnet mask), which
doubles the amount of host nodes for a class C block (domain).
* - Notice that the network ID in the routing table is now 194.73.216.0
(not 194.73.217.0). The first reason is that the subnet mask disallows
the last bit in the third BYTE (217 - requires the last byte). This
also allows up to 510 nodes to use the "216" subnet, as if they are on
the same net. These include all actual 216 and 217 subnet addresses
which are also using the subnet mask 255.255.254.0.
For reference, Figure 4.3 shows how to increase the subnets to
2 which support 62 hosts each.
NOTE: Illustration 4.3 show "subnetting" as an opposite to "supernetting",
and is only used for reference in this document.
# - Assumes 1 as the identifier of subnet 1 (of a possible 2).
5. Setting up a "Supernet" NetworkThe following is needed to "supernet" 2 or more Class C networks to act as a single Class C network domain.
See the Annex Network Administrator Guide Rev.A p A288 Table A-22 "Values for Bits Field with Corresponding Subnet Masks". This table shows a list of subnet masks which can be used, and the hexadecimal representation
6. ExamplesThe following shows examples of supernetted network domains.
6.1. Successful ScenarioThe following is a bit more explanation to the Supernetting Example in the Annex Network Administrator's Guide Rev A. p A-273. The ISP provides a client with a block of 4 Class C network addresses. They are 192.24.8.0 through 192.24.11.0. The client wants to supernet all 4 into one domain of 1,022 hosts. How can this be done?? Using the book example, the destination network address is 192.24.8.0, and the subnet mask is 255.255.252.0. Below is the "proof" that this example shall work for the client. The subnet mask "conversions" for 255.255.252.0 are as follows:
255.255.252.0 Dot Decimal Notation.
FF FF FC 00 Hexadecimal Notation.
11111111 11111111 11111100 00000000 Binary Notation.
A. Network 192.24.8.0.
1) Convert network address for routing purposes.
192.24.8.0 Dot Decimal Notation.
C0 18 08 00 Hexadecimal Notation.
11000000 00011000 00001000 00000000 Binary Notation.
2) Mask Network Address and Subnet Mask.
11111111 11111111 11111100 00000000 Subnet Mask.
11000000 00011000 00001000 00000000 Network ID.
=====================================
11000000 00011000 00001000 00000000 New Network ID
C0 18 08 00 Network ID-Hexadecimal.
3) Network Entry in the routing table.
192.24.8.0/22 22 bit network address.
4) Network host ID range for actual address 192.24.8.0.
This is the same network ID - so the host ID range will be
192.24.8.1 - 192.24.8.255.
192.24.8.0 is the network address.
Total 255 nodes.
B. Network 192.24.9.0.
1) Convert network address for routing purposes.
192.24.8.0 Dot Decimal Notation.
C0 18 09 00 Hexadecimal Notation.
11000000 00011000 00001001 00000000 Binary Notation.
2) Mask Network Address and Subnet Mask.
11111111 11111111 11111100 00000000 Subnet Mask.
11000000 00011000 00001001 00000000 Network ID.
=====================================
11000000 00011000 00001000 00000000 New Network ID
C0 18 08 00 Network ID-Hexadecimal.
3) Network Entry in the routing table.
192.24.8.0/22 22 bit network address.
4) Network Host ID range for actual address 192.24.9.0.
Since the actual network address is 192.24.9.0 and the
binary representation is:
11000000 00011000 00001001 00000000
With the mask the Host id section now contains:
01 00000000
This host ID range of the network domain will effectively
contain.
192.24.8. (256 - 511).
Note - The Dot Decimal Notation could never display this, and the
annex will never display this. But internally and logically, this is
what would be assumed.
Total 256 Nodes.
C. Network 192.24.10.0.
1) Convert network address for routing purposes.
192.24.8.0 Dot Decimal Notation.
C0 18 0A 00 Hexadecimal Notation.
11000000 00011000 00001010 00000000 Binary Notation.
2) Mask Network Address and Subnet Mask.
11111111 11111111 11111100 00000000 Subnet Mask.
11000000 00011000 00001010 00000000 Network ID.
=====================================
11000000 00011000 00001000 00000000 New Network ID
C0 18 08 00 Network ID-Hexadecimal.
3) Network Entry in the routing table.
192.24.8.0/22 22 bit network address.
4) Network Host ID range for actual address 192.24.10.0.
Since the actual network address is 192.24.10.0 and the
binary representation is:
11000000 00011000 00001010 00000000
With the mask the Host id section now contains:
10 00000000
This host ID range of the network domain will effectively
contain.
192.24.8. (512 - 767).
Note - The Dot Decimal Notation could never display this, and the
annex will never display this. But internally and logically, this is
what would be assumed.
Total 256 Nodes.
D. Network 192.24.11.0.
1) Convert network address for routing purposes.
192.24.8.0 Dot Decimal Notation.
C0 18 0B 00 Hexadecimal Notation.
11000000 00011000 00001011 00000000 Binary Notation.
2) Mask Network Address and Subnet Mask.
11111111 11111111 11111100 00000000 Subnet Mask.
11000000 00011000 00001011 00000000 Network ID.
=====================================
11000000 00011000 00001000 00000000 New Network ID
C0 18 08 00 Network ID-Hexadecimal.
3) Network Entry in the routing table.
192.24.8.0/22 22 bit network address.
4) Network Host ID range for actual address 192.24.11.0.
Since the actual network address is 192.24.11.0 and the
binary representation is:
11000000 00011000 00001011 00000000
With the mask the Host id section now contains:
11 00000000
This host ID range of the network domain will effectively
contain.
192.24.8. (768 - 1023).
192.24.8. (1024) - is the broadcast address for the entire supernet.
Note - The Dot Decimal Notation could never display this, and the
annex will never display this. But internally and logically, this
is what would be assumed.
Total 255 Nodes.
So as the above shows, the subnet mask 255.255.252.0 does work for creating a logical network domain using the 22 bit network address 192.24.8.0, with access to 1,022 hosts.
6.2. Unsuccessful ScenarioWhat if the client needed a fifth network ID of 192.24.12.0, with the subnet mask of 255.255.252.0. Let us also assume that the ISP is providing a block of 8 (next higher available size) Class C Subnet Network Addresses beginning at 192.24.8.0. Would the above subnet mask work? The subnet mask "conversions" for 255.255.252.0 are as follows:
255.255.252.0 Dot Decimal Notation.
FF FF FC 00 Hexadecimal Notation.
11111111 11111111 11111100 00000000 Binary Notation.
Network 192.24.12.0.
1) Convert network address for routing purposes.
192.24.8.0 Dot Decimal Notation.
C0 18 0C 00 Hexadecimal Notation.
11000000 00011000 00001100 00000000 Binary Notation.
2) Mask Network Address and Subnet Mask.
11111111 11111111 11111100 00000000 Subnet Mask.
11000000 00011000 00001100 00000000 Network ID.
=====================================
11000000 00011000 00001100 00000000 New Network ID
C0 18 0C 00 Network ID-Hexadecimal.
3) Network Entry in the routing table.
192.24.12.0/22 22 bit network address.
As stated above the supernet network is 192.24.12.0 and not 192.24.8.0. The last bit in the network address of 192.24.12.0 is not "masked out" with via the mask 255.255.252.0. Using this mask would not allow supernetting of subnets 8-12 inclusively.
6.2.1 Correct the ScenarioUse a 21 bit subnet mask for all 5 nets to allow them all to be included in the supernet. The perfect mask would be 255.255.248.0. The following is proof that it shall work. The subnet mask "conversions" for 255.255.248.0 are as follows:
255.255.248.0 Dot Decimal Notation.
FF FF F8 00 Hexadecimal Notation.
11111111 11111111 11111000 00000000 Binary Notation.
A. Network 192.24.8.0.
1) Convert network address for routing purposes.
192.24.8.0 Dot Decimal Notation.
C0 18 08 00 Hexadecimal Notation.
11000000 00011000 00001000 00000000 Binary Notation.
2) Mask Network Address and Subnet Mask.
11111111 11111111 11111000 00000000 Subnet Mask.
11000000 00011000 00001000 00000000 Network ID.
=====================================
11000000 00011000 00001000 00000000 New Network ID
C0 18 08 00 Network ID-Hexadecimal.
3) Network Entry in the routing table.
192.24.8.0/21 21 bit network address.
4) Network host ID range for actual address 192.24.8.0.
This is the same network ID - so the host ID range will be
192.24.8.1 - 192.24.8.255.
192.24.8.0 is the network address.
Total 255 nodes.
B. Network 192.24.9.0.
1) Convert network address for routing purposes.
192.24.8.0 Dot Decimal Notation.
C0 18 09 00 Hexadecimal Notation.
11000000 00011000 00001001 00000000 Binary Notation.
2) Mask Network Address and Subnet Mask.
11111111 11111111 11111000 00000000 Subnet Mask.
11000000 00011000 00001001 00000000 Network ID.
=====================================
11000000 00011000 00001000 00000000 New Network ID
C0 18 08 00 Network ID-Hexadecimal.
3) Network Entry in the routing table.
192.24.8.0/21 21 bit network address.
4) Network Host ID range for actual address 192.24.9.0.
Since the actual network address is 192.24.9.0 and the
binary representation is:
11000000 00011000 00001001 00000000
With the mask the Host id section now contains:
001 00000000
This host ID range of the network domain will effectively
contain.
192.24.8. (256 - 511).
Note - The Dot Decimal Notation could never display this, and the
annex will never display this. But internally and logically, this is
what would be assumed.
Total 256 Nodes.
C. Network 192.24.10.0.
1) Convert network address for routing purposes.
192.24.8.0 Dot Decimal Notation.
C0 18 0A 00 Hexadecimal Notation.
11000000 00011000 00001010 00000000 Binary Notation.
2) Mask Network Address and Subnet Mask.
11111111 11111111 11111000 00000000 Subnet Mask.
11000000 00011000 00001010 00000000 Network ID.
=====================================
11000000 00011000 00001000 00000000 New Network ID
C0 18 08 00 Network ID-Hexadecimal.
3) Network Entry in the routing table.
192.24.8.0/21 21 bit network address.
4) Network Host ID range for actual address 192.24.10.0.
Since the actual network address is 192.24.10.0 and the
binary representation is:
11000000 00011000 00001010 00000000
With the mask the Host id section now contains:
010 00000000
This host ID range of the network domain will effectively
contain.
192.24.8. (512 - 767).
Note - The Dot Decimal Notation could never display this, and the
annex will never display this. But internally and logically, this is
what would be assumed.
Total 256 Nodes.
D. Network 192.24.11.0.
1) Convert network address for routing purposes.
192.24.8.0 Dot Decimal Notation.
C0 18 0B 00 Hexadecimal Notation.
11000000 00011000 00001011 00000000 Binary Notation.
2) Mask Network Address and Subnet Mask.
11111111 11111111 11111000 00000000 Subnet Mask.
11000000 00011000 00001011 00000000 Network ID.
=====================================
11000000 00011000 00001000 00000000 New Network ID
C0 18 08 00 Network ID-Hexadecimal.
3) Network Entry in the routing table.
192.24.8.0/21 21 bit network address.
4) Network Host ID range for actual address 192.24.11.0.
Since the actual network address is 192.24.11.0 and the
binary representation is:
11000000 00011000 00001011 00000000
With the mask the Host id section now contains:
011 00000000
This host ID range of the network domain will effectively
contain.
192.24.8. (768 - 1024).
192.24.8. (1024) - is the broadcast address for the entire supernet.
Note - The Dot Decimal Notation could never display this, and the
annex will never display this. But internally and logically, this
is what would be assumed.
Total 256 Nodes.
E. Network 192.24.12.0.
1) Convert network address for routing purposes.
192.24.12.0 Dot Decimal Notation.
C0 18 0C 00 Hexadecimal Notation.
11000000 00011000 00001100 00000000 Binary Notation.
2) Mask Network Address and Subnet Mask.
11111111 11111111 11111000 00000000 Subnet Mask.
11000000 00011000 00001100 00000000 Network ID.
=====================================
11000000 00011000 00001000 00000000 New Network ID
C0 18 08 00 Network ID-Hexadecimal.
3) Network Entry in the routing table.
192.24.8.0/21 21 bit network address.
4) Network Host ID range for actual address 192.24.12.0.
Since the actual network address is 192.24.12.0 and the
binary representation is:
11000000 00011000 00001100 00000000
With the mask the Host id section now contains:
100 00000000
This host ID range of the network domain will effectively
contain.
192.24.8. (1025 - 1279).
192.24.8. (1280) - is the broadcast address for the entire supernet.
Note - The Dot Decimal Notation could never display this, and the
annex will never display this. But internally and logically, this
is what would be assumed.
Total 255 Nodes.
So as the above shows, the subnet mask 255.255.248.0 does work for creating a logical network domain using the 21 bit network address 192.24.8.0, with access to 1,278 hosts. This subnet mask can also contain the 13,14 and 15 subnets. This would allow for up to a 2,046 host node subnet.
Feedback | Service Search | ERC Copyright © Bay Networks,Inc., 1997. All rights reserved. |
/tr> |
