Summary routes

[fasalmehmood]

Please recently I am finding it very difficult to crack the IPV6 summarization but happy with the ipv4 summarization. I do know that each Hextet has 16 bits i.e. every Hexa number has 4 bits and I am not missing ANY zeros in my calculation. Please can I request you to explain the following summarization to me. Cisco ENCORE book pages are 229-234 &  273-280. 

Q1: 3 loopbacks to be summarized. 

2001:db8::1/128 Lo0 on R1

2001:db8::2/128 Lo0 on R2

2001:db8::3/128 Lo0 on R3

On page 234 the summarized loopback addresses advertised to R4 by R3 is 2001:db8:0:0::/65 network. Should that not have been 2001:db8:: /126 prefix because before :: they are all zeros and we are only looking for the matching bits and any different bits are zeroed out which I did? Please explain his logic and correct my approach?

Q2: On page 273 3 more Loopbacks on 3 router R1, R2 & R3 are to be summarized by R2

2001:db8::1/128 Lo0 on R1

2001:db8::2/128 Lo0 on R2

2001:DB8::3/128 Lo0 on R3

But on page 279 he is referring the same addresses as 2001:db8:0:1/128, 2001:db8:0:2/128 & 2001:db8:0:3/128 , Is he missing double colons before 1,2 & 3 and If not then why has the 1,2 & 3 are moved into the subnet ID portion? I have checked the errarata but nothing is mentioned about this?

But now he has summarized them to 2001:db8:0:0::/62. Again my answer would have been 2001:db8::/126. Please explain and correct my approach?

Q3: And lastly on page 279 again the same 3 router links need to be summarized again

Firstly he wanted to summarize this link between R1 & R2. Link ipv6 address 2001:db8:0:12::/64. But why is he saying to consider the 4th hextet value of 18 in decimal rather than in binary format to break them down and look for the matching bits here?

Summarized answer is 2001:db8::/59

Later on he included the link between R2 & R3 to summarize as well. IPv6 Link address 2001:db8:0:23::/64. Again 23 Hexadecimal decimal conversion is done first?

Summarized answer is 2001:db8::/58

Any explanation about all 3 questions will be highly appreciated.

[Harold Ritter]

Q1: You are correct in saying that these 3 routes could be summarize to 2001:db8:: /126, but it is also correct to summarize to 2001:db8:0:0::/65, which would include a lot more than these 3 routes. It just depends on what you are trying to achieve.

Q2:

> But on page 279 he is referring the same addresses as 2001:db8:0:1/128, 2001:db8:0:2/128 & 2001:db8:0:3/128 , Is he missing

> double colons before 1,2 & 3.

You are correct. It should be 2001:db8:0::1/128, 2001:db8:0::2/128 and 2001:db8:0::3/128.

> But now he has summarized them to 2001:db8:0:0::/62. Again my answer would have been 2001:db8::/126. Please explain and correct > my approach?

If the goal was to make the aggregation prefix the most specific, the /126 would be better, but it is necessarily incorrect to decide to summarize to a /62.

Q3:

> But why is he saying to consider the 4th hextet value of 18 in decimal rather than in binary format to break them down and look for the

> matching bits here?

Converting to decimal normally makes it easier to tell how many bits are required to summarize (0x12 == 18 requires 5 bits, 0x23 == 35 requires 6 bits).

Regards,

[fasalmehmood]

Thank you Harold for clarifying. I am confident with the 3rd question now and can resolve it but please if you can give me a little breakdown of how the author have solved the Q1 that will be much appreciated and then I can use the same approach to solve the Q2 then as I am not familiar with his approach/method?

Thank you again for your reply.

[Harold Ritter]

Hi,

For both Q1 and Q2, the aggregate addresses provided in the book are not optimal but they encompass the 3 mentioned addresses (i.e. 2001:db8::1/128, 2001:db8::2/128 and 2001:db8::3/128).

2001:db8:: /126 includes 2001:db8::/128, 2001:db8::1/128, 2001:db8::2/128 and 2001:db8::3/128

2001:db8:0:0::/65 includes 2001:0db8:0000:0000:0000:0000:0000:0000 through 2001:0db8:0000:0000:7fff:ffff:ffff:ffff (therefore including the 3 /128 mentioned above even if not optimal) 

2001:db8:0:0::/62 includes 2001:0db8:0000:0000:0000:0000:0000:0000 through 2001:0db8:0000:0003:ffff:ffff:ffff:ffff (therefore including the 3 /128 mentioned above even if not optimal)

Hope this help,

[fasalmehmood · Accepted solution]

Thank you Harold, it is much clearer now.