How To Summarize IP Networks

When we think about ways to efficiently manage IP networks and their addressing, subnetting is often the first technique that comes to mind. Subnetting is an extremely important topic in modern IP management, but it only focuses on carving assigned network space into smaller and smaller parts. Address summarization, often referred to as supernetting, does the opposite – representing several blocks of IP addresses with a single, comprehensive prefix.

IP Address summarization is useful in a number of ways, but is particularly helpful when manually summarizing routes to reduce routing table entries and overhead. Let’s walk through an example to teach you the steps.

Network Summarization Example

Networks to be summarized:

Step 1. Find the interesting octet

The first step in summarizing a list of networks is to identify the first octet where the decimal value is not the same.

In this example, the first octet for all networks is 172. No difference there. The second octet is 16 for every network in the list. Again, no difference either. The third octet is where we see different values, which is what we’ll call the interesting octet. That is where we need to focus our attention.

Step 2. Convert the interesting octet values to binary > 172.16.00000011.0/24 > 172.16.00000100.0/24 > 172.16.00000101.0/24 > 172.16.00000110.0/24 > 172.16.00000111.0/24 > 172.16.00001000.0/24 > 172.16.00001001.0/24

Step 3. Identify the common bits and convert to decimal


The common bits in the third octet are shown in red above. Now we simply need to add trailing zeros to the end and convert it to decimal. This gives us the summary network address.

172.16.00000000.0 > = Summary Address

Step 4. Count the number of leading common bits to find the mask

To find the summary mask, we need to count the number of bits from left to right until they no longer match. We already determined that in our example the first and second octets match, each with 8 bits. The third octet contains four matching bits (in red). Combined, this gives us a mask of 20 (8+8+4).

172.16.00000000.0 > 10101100.00010000.00000000.00000000 (20 leading common bits) = Summary Address

This was a simple summary example, but the same process will work for any networks. If you have other IP summary questions or need more examples, feel free to leave a comment below.

Author Aaron

Aaron knows networks. He's been involved in building and supporting world-class data networks for the past 10 years - from international cloud service providers to Fortune 50 data centers. Aaron consults independently and is focused on building the best training platform available.

More posts by Aaron

Join the discussion 2 Comments

  • Incroyable HULK says:

    Hello Aaron,

    I am a little confused with this article. I was under the impression that summarization had the goal of simplifying IP management and could be practically used for ACL and other configuration scenario.

    In your example, I believe the result summarization is too large ?! would result in 16 Class C subnets: to to to to
    …. to

    So, the subnet you wanted: to are included in this summarization but how efficient is it to write rules that would allow traffic from 9 other subnets?

    Maybe I am missing something here? I’m no networking expert!

    I was also reading this article from Cisco as a reference:

    Anyway, thanks for your awesome website!

    • Aaron says:

      Great question! Your are exactly right – a /20 summary address covers more than the list of networks presented. The steps above summarize a list of prefixes into the smallest single summary prefix. The example highlights the fact that in some scenarios this process can be rather address ineffiecient – covering well more than the required networks. Breaking summaries down to an exact match for a list of networks is a more complicated effort. You essentially have to piece together many smaller prefixes or summery prefixes to come up with no extra addressing space.

      For example, the network list used here when broken down into an exact match looks like the following:

      That hardly seems like it reduces complexity, but it is more concise. Generally, a single superset address is preferred if the addressing is flexible enough to accommodate extra address space that may (or may not) come with the smallest single summary prefix. Hope that helps!

Leave a Reply