Predefining with Symmetrical Gaps


The next logical step is to take an address space, presubnet it, and add symmetrical gaps between the preassigned blocks. This is one of the more elegant and sophisticated approaches to managing an address space, but it is difficult to describe. The goal of such an approach is to maintain some semblance of symmetry and order. Scanning back through the tables presented in previous sections, you can see that there's a bit of a "rat's nest" look to the address assignments. In other words, blocks were created as they were needed, and there is no rhyme or reason to the juxtaposition of different block sizes.

One tactic you can use to maintain a more orderly appearance is to preallocate blocks based on prefix size. While I'll admit that neatness might not be something everyone aspires to, it definitely helps keep an address space more manageable and might even prove to be the more efficient approach over time.

This approach is based on the binary mathematics of the IP address space. It should be intuitive by now that in a binary number system, for each numeral you move to the right, you are halving the value of the previous numeral. Conversely, for each numeral you move to the left, you are doubling the value of the previous numeral. Applying this mathematical fact, a /25 network is exactly half of a /24. Thus, there are two /25s in a /24. You can put this concept to use in managing an IP address space by predefining blocks.

If you are managing a large space, such as a /16, you might find it easier to manage space by predefining a pair of /25s whenever you need to create one. That way, a new /25 is ready and waiting should the need arise for a new one. Assigning the lower range of numbers first also creates room for the /25 to grow into a /24. To illustrate this point, consider Table 14-5.

Table 14-5. Demonstrating Symmetry

Network Number

CIDR Mask

Base Address

Base

/24

192.168.125.0

1

/25

192.168.125.0

2

/25

192.168.125.128


In Table 14-5, you can see how the /25s neatly cleave the /24 into equal halves. The first one contains host addresses numbered 0 through 127, and the second contains host addresses 128 through 255. Thus, the two /25s are sometimes referred to by their relative values as the "upper half" and the "lower half" of the /24. Should you need to create a network or subnetwork with more than 64 devices but fewer than 128, you would use a /25. To keep things symmetric and enable future growth, you would record both halves being created in your database but assign only one. Ostensibly, you would start with the lower half of the /24. The other would remain free for future use.

Similarly, you can create four /26 networks from within a /24-sized address space. Alternatively, as you can see in Table 14-6, you can take a /24 and carve it into four /26 network blocks. Another neat twist shown in this table is to take one of those /26 blocks and subdivide it into a pair of /27s. The symmetry is apparent even in decimal, because the initial host addresses of each of those smaller blocks starts on a familiar power of 2 or sums of powers of 2.

Table 14-6. Demonstrating Symmetry

Network Number

CIDR Mask

Decimal Address Value

Base

/24

192.168.125.0

1

/26

192.168.125.0

2

/26

192.168.125.64

3

/26

192.168.125.128

4

/27

192.168.125.192

5

/27

192.168.125.224


Strengths and Weaknesses

This approach excels in creating a logical and symmetrical framework for the long-term management of your address space. If you maintain your records properly, you can tell at a glance how much address space remains unassigned, and in what size chunks. It is important to remember that you aren't necessarily creating "routes" that your routers will have to track when you predefine address blocks in this manner. A network prefix doesn't become a route until it is implemented. Even then, your network would aggregate network prefixes to deal with the largest block possible without compromising its ability to deliver all datagrams successfully within your network.

Another nice feature of this approach is that you can supernet unused contiguous blocks to create spaces that are larger should the need arise. As you saw in Chapter 6, supernetting absolutely requires numerically contiguous address space. Two or more smaller spaces can be integrated into a single larger space just by virtue of changing a network mask. Predefining address space creates the potential for supernetting in the future.

Perhaps the only weakness in this approach is that the predefined space is inherently idle. However, it isn't necessarily wasted. Remember: This is space that would be idle regardless of how you did your network assignments. Predefining it just puts the available address space into sized blocks that are immediately recognizable and creates room for growth for neighboring networks. If those neighbors don't require growth, the block can be put to another use.

Realistic Context

To better demonstrate the concept of predefined address spaces, consider Table 14-7. In this table, you can see how an address space is parsed across the various departments of a moderate-sized company. The company enjoys a /22 network block, which is equal in size to four Class C networks (1024 total addresses). At a high level, the fictitious company's /22 CIDR block has been sectioned into a pair of /24s and a /23. These blocks are then carved into even smaller blocks for assignment to individual departments within the company.

Table 14-7. Symmetrical Assignments in a 23-Bit Network

Network Number

Status

CIDR Mask

Base Address

CIDR block

/22

192.168.125.0

Base

/23

192.168.125.0

Base

/24

192.168.125.0

1

HQ

/25

192.168.125.0

2

Free

/25

192.168.125.128

Base

/24

192.168.126.0

1

Sales

/26

192.168.126.0

2

Finance

/26

192.168.126.64

3

Free

/26

192.168.126.128

4

HR

/27

192.168.126.192

5

Network interfaces

/27

192.168.126.224

Base

Free

/23

192.168.127.0


Did you notice that the free spaces were created to maintain symmetry in the address space? We could have taken the /25 free space that immediately follows the HQ space and allocated it to both the Sales and Finance Departments (their two /26s equal a /25), but that would have left three important groups of users landlocked without any way to expand except for renumbering. This isn't a very attractive option. You can expand a subnet in one of two ways. If there's adequate space that is numerically contiguous, you can supernet to create a larger block. Your only other alternative is to move the affected department to a larger space that must be carved from the free /23 space.

In this example, because HQ gets a /25, we retain symmetry by creating a free /25. This effectively consumes the first /24 worth of addresses in our block. The next block of addresses assigned comes from the top of the next /24.

This is perhaps the best way to carve up an address space. Maintaining symmetry through the use of carefully sized and predefined free spaces ensures that you won't run into any problems like the ones you saw earlier in this chapter.




IP Addressing Fundamentals
IP Addressing Fundamentals
ISBN: 1587050676
EAN: 2147483647
Year: 2002
Pages: 118
Authors: Mark Sportack

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net