Download Clarification of book page 224

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts
no text concepts found
Transcript
Why subnet? The IPv4 address space is very constrained. We want to get the most out
of our limited address space. Assume we currently have a class C network. The entire
address space for a single class C network has 256 possible values. This does not take
into account IP addresses that cannot be assigned – like .0 and .255. These are
reserved addresses, and so practically we only have 254 addresses that can be
assigned to individual hosts in the network.
I want you to clearly see that 256 = 28. The 8 (exponent) refers to the fact that there are
8 bits in the standard class C suffix (the last octet). Recall the diagram below.
When subnetting a class C IPv4 address, we are only manipulating bits in the last octet.
There is nothing we can do with the first three octets, as they are part of the standard
class C network identifier.
There may be administrative reasons or IP address allocation reasons for subnetting.
The result of subnetting is essentially breaking apart, or subdividing one of the standard
IPv4 address blocks into multiple separate logical networks (that come from the same
standard IPv4 address block). Each subnet will have its own network reference (a
specific IP address that denotes the specific network) and broadcast domain (and
specific IP broadcast address for that particular network).
Some examples: SBA could place each classroom on a separate subnet. This would
allow each classroom to exist on a logically separate network. The traffic we create or
exercises we perform in one class would be isolated on that subnet and would generally
not impact other subnets (this is a simplification, but illustrates how subnetting can
beused in practice).
You CANNOT arbitrarily break up IP address space within a given IP class (A, B, or C)
and assign some addresses to one logical network and some to another logical network
without subnetting. Why not?
A single IP network address block has only one IP address that denotes the network
and one IP address that is used for broadcast messages within that network.
Take any class C address block. The last octet of the each class C address block
ranges from .0 to .255. For example, a standard class C IPv4 address block might be
200.152.75.0, .1, .2, .3 - .255 (The address block includes ALL 256 addresses in that
1
block where 200.152.75.0 is the network reference IP and 200.152.75.255 is the
broadcast reference IP for that particular network.
Different physical / logical networks cannot share these network and broadcast IP
addresses. For example, 200.152.76.0-.255 is a completely different class C address
block with a network reference IP of 200.152.76.0 and a broadcast reference IP of
200.152.76.255.
Different organizations typically do not share physical or logical networks. You can
also think back to discussions of bridges, routers, LANs, and VLANs and think about
instances where separate logical and / or physical networks are used. In addition to
the hardware and supporting software (like configuring a switch to support different
VLANs) used, you must account for IP addresses and how these addresses are
used. It could also be the case that the networks are on separate physical networks,
each with their own administrator.
Subnetting is a way of creating multiple SEPARATE logical networks from a single IPv4
network address space associated with a specific address class. Once you subnet a
given address (like a class C address), it is generally divided that way for good, although
it is POSSIBLE to recombine the subnetting scheme without applying the changes to
ALL addresses for that particular network block.
Keep in mind that some of the subnetting schemes are not common or may not even be
possible. In general, you cannot use subnetted bit combinations of all 0s or all 1s in
either the prefix or the suffix for the various subnetting schemes. For example, if you
subnet a class C address with 2 bits in the prefix and 6 in the suffix, you end up with 22 =
4 subnets and 26 = 64 hosts per subnet. Traditionally, the subnets that include IP
addresses in the range .0 - .63 or .192 - .255 cannot be used (the first all begin with two
zeros in the prefix and the second begin with all ones in the prefix (00 and 11 are not
allowed while 01 and 10 are). Likewise, valid HOST addresses (IPs you assign to
devices like PCs and routers will not have all zeros or ones in the suffix. In this example,
(.0, .63), (.64, .127), (.128, .191), (.192, .255) would represent the network reference and
broadcast IP for each of the 4 subnets.
Say, you subnet a class C address with 4 bits in the prefix and 4 bits in the suffix. Keep
in mind with class C, we are ONLY concerned about the last 8 bits of the 32 bit IPv4
address space – the standard suffix of a class C. We do not mess with the 24 bit
standard prefix of the class C. Why?
At a minimum, class C addresses have 24 bits assigned to the prefix. We can “add”
bits to the standard prefix via subnetting, but cannot take bits away from the standard
prefix.
We should have: 24 = 16 subnets and 24 = 16 hosts per subnet, but the book and notes
say 14 subnets and 14 hosts per network. WHY?
In this scheme, there are only 14, valid 4-bit subnet IDs for class C address. We
CANNOT use any combinations of 0000 or 1111 in EITHER the subnetted prefix or
suffix. So, if we subnet the class C address block 195.70.115.0 by placing 4 bits of
the last octet in the prefix and 4 in the suffix, we create the following IP address
blocks:
2
1) 195.70.115.0 – 195.70.115.15
2) 195.70.115.16 – 195.70.115.31
3) 195.70.115.32 – 195.70.115.47
4) 195.70.115.48 – 195.70.115.63
5) 195.70.115.64 – 195.70.115.79
6) 195.70.115.80 – 195.70.115.95
7) 195.70.115.96 – 195.70.115.111
8) 195.70.115.112 – 195.70.115.127
9) 195.70.115.128 – 195.70.115.143
10) 195.70.115.144 – 195.70.115.159
11) 195.70.115.160 – 195.70.115.177
12) 195.70.115.176 – 195.70.115.191
13) 195.70.115.192 – 195.70.115.207
14) 195.70.115.208 – 195.70.115.223
15) 195.70.115.224 – 195.70.115.239
16) 195.70.115.240 – 195.70.115.255
This range would not be used
This range would not be used
We cannot use bolded blocks 1) and 16) because the valid subnet addresses and host
IPs cannot contain all 0s or all 1s in the four bits for the network ID OR in the 4 bits for
the host ID (0000 and 1111). The valid combinations for both the prefix and suffix are:
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
Since any number in block 1) MUST start with 0000 (in the subnetted suffix), that block
cannot be used. Example: .1  0000 | 0001  the network ID portion of the subnet (the
suffix) is all 0s
Example .6  0000 | 0110  the network ID portion of the subnet is all 0s
Example .15  0000 | 1111  the network ID portion of the subnet is all 0s AND the
host portion of the subnet is all 1s
Likewise any number in block 16) MUST start with 1111, that block cannot be used.
Example: .245  1111 | 0101  the network ID portion of the subnet is all 1s
Example .255  1111 | 1111  the network ID and host ID portions of the subnet are all
1s
So, the first value for an IP address is .17. Where does this come from?
Using a subnetting scheme of 4 bits to the prefix and 4 bits to the suffix, we cannot
use the first address block for the reason stated above. So, anything from .0 to .15
cannot be used. .16 is the network reference for block 2) and CANNOT be assigned
to a host. The first host address that can be assigned is .17. The last IP address that
can be assigned to a host from block 2) is .30. Why not .31? .31 is the network
broadcast IP for block 2). So, the first IP that can be assigned to a host from block 3)
is .33.  .32 is the network reference for block 3). .47 is the broadcast reference for
block 3). These IPs (and the upper and lower bound for EVERY subnet block) are
3
RESERVED. Lower bound is the network reference (or name) and the upper bound
is the broadcast address for the respective network.
Why does each subnet need its own network reference and broadcast? The subnet
is a SEPARATE logical network. Each network MUST have its own network and
broadcast reference. Recall, a LAN is basically defined by its broadcast domain.
4