Implementing IP Addressing

 

 
Implementing a Hierarchy.  We've briefly touched on subnetting and how addresses may be organized hierarchically so that they are easier to find and easier to packet filter.  The first thing we need to do is work towards something called route summarization.  Route summarization is also called route aggregation or supernetting.  Making use of route summarization means that one route in an routing table represents many other routes.

Route summarization is based on using the bits associated with the network to route the packets to a particular network and then using the host bits once that network is reached.  This is not all that different from prefix routing used in telephony with area codes and local exchanges.  To make route summarization more effective CIDR - Classless Inter-Domain Routing is important due to its ability to account for the ever growing number of networks on the Internet.  With CIDR we are no longer limited to classes A, B or C.

In order for route summarization to be effective, the following criteria must be met.

  • Multiple IP addresses must share the same leftmost bits.
  • Routers must base their routing decisions on 32 bit IP addresses and a prefix length of up to 32 bits.
  • Routing protocols must carry the prefix length with the 32 bit IP address.

For example, let's assume a router has the following networks behind it.

  • 192.108.164.0
  • 192.108.165.0
  • 192.108.166.0
  • 192.108.167.0
  • 192.108.168.0

Each of these networks could be advertised to other routers separately, but this would mean advertising five routes.  These routes can be summarized by the following.

192.108.168.0/21

This CIDR says that the first 21 bits should be used as the destination.

When should such a hierarchy be used?

  • Is a hierarchy needed within the addressing plan?

  • What are the criteria for dividing a network into summarization groups?

  • How is the route summarization performed and how dies it correlate with routing?

  • How many end systems does each route summarization and/or subgroup contain?

The following diagram represents how the number of devices and the relative physical locations have a lot of impact on the reasons for supernetting.

 

 

While it is a bit difficult to put into words and is likely better represented by images, you want to aggregate addresses that are
  • located reasonably close to each other
  • will readily have packets follow the close to the same paths as other devices on the aggregate
  • modularly composed
  • have greatest potential for scalability

As you can see we can have aggregations within other aggregations.

Static or Dynamic Addresses.  Finally we will discuss some of the situations and motivations for selecting whether something will have a static or dynamic IP address.

For example, you know you don't want your gateway router(s) to be constantly changing IP addresses.  This is the same for your DHCP server(s).  You also aren't likely to want these addresses to be public since it makes it easier for crackers to find them on the Internet.

Some reasons for one or the other.

  • A large number of end systems - prefer dynamic addresses
  • Improved hiding for things that don't need to be visible in making addresses change.
  • A user is less likely to be able to do something like configure their own server for less than proper purposes if their IP address changes each time they get on the network.  How will they advertise their address?
  • If you want to track address then you prefer unchanging addresses
  • Higher availability is fostered with dynamic addresses
  • Dynamic addresses are actually easier to acquire for people who get on your network - good for some things - bad for others
  • Want static IP addresses for things like servers and/or network devices