Original subtitles

1

So I want to teach you a trick

2

now this doesn’t always apply, it only works in certain situations

3

but it saves you a lot of a time

4

if you remember back to your Binary, this bit is 128

5

this bit is 64, this is 32, this is 16

6

this is 8, this is 4, this is 2 and that is 1

7

so 255 in decimal and an IP address would be an octet populated with binary 1's

8

please refer back to the ICND 1 course if you can’t remember binary

9

but hopefully, at the point, you're fairly comfortable with it.

10

If you were given subnets where for instance

11

the third octet was in the range 4 to 7

12

in other words, from 4 to 1 less than 8 so 7

13

you could summarize that automatically as 4

14

so for example, let’s say you’re given 172.16.4.0/24 up to 172.16.7.0/24

15

so notice in the third octet the range is from 4 to 7

16

so in other words, from 4 to 1 less than 8

17

you could immediately write the answer as 172.16.4.0

18

now to work out the subnet mask you just remember that the first octet is 8 bits

19

the second octet is 8 bits and that’s 16

20

and then you need to work out where binary value of 4 is

21

so let's count 1 2 3 4 5 6, so it’s in binary bits 6

22

so 8 + 8 = 16 + 6 binary bits which we’ve not counted to see where 4 is

23

gives you 22, so the mask would be 22 8 + 8 + 6

24

and it’s a simple as that

25

to work out the answers to a question likes this

26

by the same token if you were given an example

27

where the values was from 8 to 15

28

in other words 8 to 1 less than 16

29

you could summarize that immediately as 8.

30

So let’s say for example it was 10.8.0.0/16 up to 10.15.0.0/16

31

in other words, from 8 to 1 less than 16

32

you could summarize it automatically as 10.8.0.0

33

so in other words, were saying if it's from this binary value 8

34

up to 1 less than the next binary value

35

you just summarize it down to this binary value of 8

36

finally, to work out the subnet mask you need to remember

37

that the first octet is 8 bits and then work out where 8 is

38

so 8 is 1 2 3 4 5

39

so 8 + 5 will give you 13

40

8 binary bits + 5 binary bits gives you 13

41

so the mask is 13, by the same token 16 to 31

42

so 1 less than 32 can be summarized to 16

43

32 to 63 in other words, 1 less than 64

44

so 32 to 63 can be summarize to 32

45

64 to 1 less than 128 in other words 127

46

so 64 to 127 can be summarize as 64

47

now I’ve already shown you those examples by working it out in binary

48

just to remind you 64 up to 127

49

we work out in binary and work out the answer as 172.16.64.0

50

so once again, 64 to 127 can be summarized as 64

51

and then you count the number of common bits

52

so 8 + 8 + 2 because 64 is in the second binary bit position

53

giving you a total of 18

54

so, therefore, you can work out this answer

55

in a matter of seconds rather than minutes

56

this example with 172.16.32.0 up to 172.16.63.0

57

can quickly and easily be summarized as 172.16.32.0

58

19 bits are in common and the way we work that out

59

is 8 bits in the first octet + 8 bits in the second octet

60

is 16 + 32 is in the third binary bit position

61

so 3 bits gives you a total of 19

62

so I’m hoping this trick will save you quite a bit of time

63

when working out summarization please be careful though

64

if you are given an example of let say 16 to 35

65

you're going to have to split up your summary

66

the 16 to 31 subnets can easily summarize very quickly

67

but if the question asks you to summarize subnets

68

that go across this bit boundaries

69

then you would have to work it out in binary

70

but this will hopefully save you a bit of time

71

also be careful if you're given an example

72

where you're asked to summarize from 16 to let say 19

73

and you use this example that I’ve explain

74

you’ll be summarizing more than just those subnets

75

so it will be better, in that case, to do it in binary

76

So what are the advantages of VLSM and summarization?

77

We get more efficient use of the IP address space

78

so rather than for instance having to use a /24 mask

79

on a serial link which consumes 254 host addresses

80

we can use a /30 mask which only needs to, there are fewer updates

81

because we can hide network changes

82

or topology changes by sending a summary root

83

rather than individual networks or subnets to other devices

84

it also allows us to implement hierarchical levels

85

for better route summarization, so in the real world VLSM and route summarization

86

are used very heavily to conserve IP addresses and reduce routing table sizes

87

so here’s an example of address hiding and topology change hiding

88

the router on the right-hand side only receives 1 route

89

from the route from the left-hand side 10.1.0.0/16

90

so if a more specific subnet like 10.1.12.0/24 went down

91

the router on the right-hand side is oblivious to that fact

92

because it only has 10.1.0.0/16 in its routing table

93

and that’s all that's been advertised to it that route state has not changed

94

and thus the router on the right-hand side

95

does not have to reprocess or re-compute its routing table

96

it is oblivious to the fact that this subnet 10.1.12.0 has gone down

97

because all it sees is the super net or summary of 10.1.0.0/16

98

thus there are major advantages to implementing summarization

99

including topology change hiding

100

however, it’s important that you realize that there’s a difference

101

between what are called classful routing protocols

102

and classles routing protocols

103

classful routing protocols do not include the subnet mask

104

when advertising the network, that means other devices do not know

105

what subnet mask is being used

106

so router assumes and we all know how bad it is to assume

107

but they assume that within the same network

108

there is consistency of the subnet mask

109

in other words, everyone within the same network

110

is using the same subnet mask as everyone else

111

so in other words, when a router's received on an interface

112

the subnet mask for the received route is implied

113

by the subnet mask on the local interface

114

as the router does not know what subnet mask was used by the other routers

115

so it assumes that they are using the same subnet mask as itself.

116

routes will automatically be summarized when going across a classful boundary

117

so summary routes are exchanged when crossing a classful boundary

118

in other words, as an example when going from a 10 network

119

to a 192.168 network or from 10 to 11 and so forth and so on

120

examples of classful routing protocols

121

includes RIP version 1 and IGRP

122

IGRP is no longer supported on the Cisco IOS

123

and RIP version 1 shouldn’t be used in today’s networks

124

but just for completeness, it's mentioned here.

125

Classless routing protocols do include the subnet mask

126

with the network in routing advertisements

127

in other words, classless routing protocols advertise

128

not just the network like 10.1.1.0 but also the associated mask like /24

129

because the subnet mask is included in the routing updates

130

classless routing protocols support Variable Length Subnet Mask or VLSM

131

summary routes can be manually configured

132

so unlike in classful routing protocols

133

where automatic summarization takes place across classful boundaries

134

in classless routing protocols summarization in some cases, for example

135

with EIGRP can be configured on any interface anywhere in the network

136

examples of classless routing protocols include

137

RIP version 2, EIGRP, OSPF and ISIS

138

in this course, we'll concentrate mainly on RIP v2, EIGRP and OSPF

139

but just be aware that there are other routing protocols out there

140

be careful EIGRP and RIP v2 act as classful routing protocols by default

141

you need to use the command no auto summary within the routing process

142

to disable this default behavior

143

so that they act like a classless routing protocol.

144

So let’s look at some of the issues regarding discontiguous networks

145

or discontiguous subnets, the router on the left

146

has a network of 10.1.1.0/24 connected to it

147

this if you remember is a class A subnet, the router on the right

148

has a subnet of 10.1.2.0/24 connected to it also a class A subnet

149

they are both connected to the router at the top

150

with class C addresses of 192.168.1.0 and 192.168.2.0

151

so please note we are going from a class A, to class C, to class A subnet

152

when traversing these routers

153

the problem here is classful routing protocols like RIP v1 and IGRP

154

will automatically summarize this subnets their classful network

155

so 10.1.2.0 Will automatically be summarize as 10.0.0.0

156

the same will take place here, on this router 10.1.1.0

157

will automatically be summarize to 10.0.0.0

158

this causes an issue for the router in the middle

159

because when it wants to go to 10.1.1.0

160

it believes it can send traffic to the left, as well as to the right

161

because it's receiving the same route from multiple routers

162

If this router was pinging a device over here 10.1.1

163

it would only be a 50% success rate because half of the packets

164

will be sent to this network on the right-hand side

165

be careful of routing protocols like EIGRP and RIP v2

166

even though they are classless they act as classful

167

and thus have the same issue, where they automatically summarize

168

a classful boundaries, don't forget to use the command

169

no auto summary under the routing process to disable this behavior

170

Once you've typed that command, the routers will not summarize the networks

171

and they will be advertised in EIGRP in RIP v2

172

as 10.1.1.0/24 as well as 10.1.2.0/24

173

so the router in the middle will be able to correctly route

174

to the various networks OSPF does not have this issue

175

because OSPF does not automatically summarize

176

you have to manually summarize networks.

177

So when does automatic summarization does takes place?

178

well it only affects this routing protocols RIP v2, EIGRP, RIP v1 and IGRP

179

it occurs when you move across classful boundaries

180

in other words, when a subnet is advertised from a class A to class B

181

or B to C or any one of these combinations

182

in other words, when a router has 1 interface in a class A network for example

183

and another interface in a class B network

184

and that advertisement crosses that classful boundary going from A to B

185

the network will automatically be summarized

186

another one that people forget is when you are moving

187

across major network boundaries, automatic summarization will also take place

188

in other words, if you go from a 10 network to an 11 network

189

or to a 12 network automatic summarization will take place

190

notice the major network 10 has changed to 11 or to 12

191

these are all class A networks

192

but you are moving across a major network boundary

193

so if 1 interface on a router is in the 10 network

194

and another interface on a router is in the 11 network

195

there will be automatic summarization.

196

Remember on EIGRP and RIP v2 to type the command no auto-summary

197

because even though they are classless routing protocols

198

they act as classful routing protocols

199

when it comes to automatic summarization

200

now here’s another situation that causes a lot of confusion

201

in ICND 1 you learned about administrative distance

202

and you learned that the lower the administrative distance

203

the more preferable a route is, the administrative distance of RIP v2 is 120

204

the administrative distance of OSPF is 110

205

the administrative distance of EIGRP is 90.

206

So let's assume router 1, router 2 and router 3 have networks in the 10 range

207

connected to them, they are advertising various routes to router 4.

208

So RIP v2 is advertising 10.1.1.0/27

209

OSPF is advertising 10.1.0.0/16 EIGRP is advertising 10.0.0.0/8

210

so router 4 is receiving multiple advertisements in the 10 range

211

but if on router 4 you type the command ping 10.1.1.1

212

which way will a traffic flow, will it go to router 3

213

or will it go to router 2 or will it go to router 1?

214

Now remember EIGRP has a lower administrative distance than OSPF

215

which has a lower administrative distance than RIP

216

but please note administrative distance only comes into play

217

when the same prefix is advertised

218

a prefix is not just the network it's the network and the mask

219

router 4 will see this prefixes 10.1.1.0/27

220

10.1.0.0/16 and 10.0.0.0/8 as separate prefixes

221

these 3 routes will appear in the routing table with router 4

222

and router 4 will make its decision on the best match.

223

10.1.1.0/27 is the best match out of these 3 routes.

224

27 is the most specific, so the most specific or best match will be used

225

and not the administrative distance

226

the administrative distance would only be used

227

if the same route was advertised by multiple routing protocols

228

so in this case, the ping to 10.1.1.1

229

will go to router 1 and not router 2 or router 3

230

however, in this example, notice the same prefix is advertised

231

by the 3 routers 10.0.0.0/8 is advertised by RIP, OSPF and EIGRP

232

in this case only 1 route can be put into the routing table

233

and the choice is done via administrative distance

234

EIGRP having the lowest administrative distance

235

will have its route inserted into a routing table

236

and the ping from router 4 will now go to router 3

237

to sum this up, in this example, there are 3 separate prefixes

238

the router does not see this as the same network

239

it sees them as 3 separate prefixes or subnets

240

all 3 will be put into the routing table

241

and a decision will be made on the best match or longest prefix

242

in this case, 27 is longer than 16, just longer than 8

243

so the RIP v2 route will be chosen, however, where the route is the same route.

244

So in this example 10.0.0.0/8

245

the choice will be made on administrative distance with EIGRP winning

246

because it has the lowest administrative distance

247

please don’t forget this, a lot of engineers make the mistake

248

of assuming that administrative distance is the way choices are made

249

for choosing the best route

250

administrative distance is only chosen as a tie breaker

251

when the same route or prefix is attempted

252

to be put into the routing table by multiple routing protocols

253

So what have we covered?

254

we look at Variable Length Subnet Mask or VLSM

255

we discuss CIDR or Classless Inter-Domain Routing

256

we talked about summarization and the advantages of summarization

257

I showed you examples of how to work out summarized routes

258

I showed you routing choices and how routers will make a choice

259

firstly on most specific match and then secondly on administrative distance

260

and then I showed you some issues regarding discontiguous networks.

261

Thank you for watching!