1
00:00:00,000 --> 00:00:03,776
[AUDIO LOGO]

2
00:00:03,776 --> 00:00:07,080

3
00:00:07,080 --> 00:00:11,220
So let's begin here by
jumping in and just getting

4
00:00:11,220 --> 00:00:16,560
an introduction or an overview
of the IS-IS protocol.

5
00:00:16,560 --> 00:00:19,840
Hopefully, some of this
is a review for you.

6
00:00:19,840 --> 00:00:23,388
But if not, it'll at least
get us to a good foundation

7
00:00:23,388 --> 00:00:24,930
where we can begin
to talk about some

8
00:00:24,930 --> 00:00:28,650
of the more detailed
information about IS-IS.

9
00:00:28,650 --> 00:00:31,320
So even if this
is a review, it's

10
00:00:31,320 --> 00:00:33,900
probably worth watching
just to refresh yourself

11
00:00:33,900 --> 00:00:36,240
on all this information
and make sure

12
00:00:36,240 --> 00:00:40,320
that we're all ready to go with
some of our more advanced IS-IS

13
00:00:40,320 --> 00:00:41,140
topics.

14
00:00:41,140 --> 00:00:43,440
So again, let's just
jump in and cover

15
00:00:43,440 --> 00:00:47,170
some of those IS-IS basics.

16
00:00:47,170 --> 00:00:49,650
So the first thing to
mention about IS-IS

17
00:00:49,650 --> 00:00:55,050
is this is a routing
protocol designed by the OSI,

18
00:00:55,050 --> 00:00:57,600
or Open Systems Interconnection.

19
00:00:57,600 --> 00:01:02,910
And this is designed to be
an interior routing protocol.

20
00:01:02,910 --> 00:01:06,360
It's not meant to scale to,
say, the size of the internet,

21
00:01:06,360 --> 00:01:08,280
like BGP version 4.

22
00:01:08,280 --> 00:01:12,190
It's really in direct
competition, if you will,

23
00:01:12,190 --> 00:01:13,920
with OSPF.

24
00:01:13,920 --> 00:01:18,570
In fact, we'll be comparing
this a lot to OSPF as we go.

25
00:01:18,570 --> 00:01:20,320
They're very, very similar.

26
00:01:20,320 --> 00:01:23,080
Now, a little bit of terminology
to get out of the way--

27
00:01:23,080 --> 00:01:29,010
we basically have two different
types of nodes in IS-IS.

28
00:01:29,010 --> 00:01:32,190
The first is the end system.

29
00:01:32,190 --> 00:01:37,900
The end system is really just
any host that's not a router.

30
00:01:37,900 --> 00:01:40,500
So this could be a
printer or a workstation

31
00:01:40,500 --> 00:01:44,340
or a server, really just
anything other than a router.

32
00:01:44,340 --> 00:01:48,930
Then, of course, there's the
Intermediate System, or the IS,

33
00:01:48,930 --> 00:01:51,450
which, of course, is where
the routing protocol gets

34
00:01:51,450 --> 00:01:52,350
its name--

35
00:01:52,350 --> 00:01:55,450
intermediate system to
intermediate system.

36
00:01:55,450 --> 00:01:59,140
And that is simply
referring to a router.

37
00:01:59,140 --> 00:02:02,760
So basically, this is the
router-to-router protocol.

38
00:02:02,760 --> 00:02:06,030
That's essentially
what IS-IS is.

39
00:02:06,030 --> 00:02:08,520
Now, a lot of people aren't
really all that familiar

40
00:02:08,520 --> 00:02:11,790
with IS-IS, just
because it's not

41
00:02:11,790 --> 00:02:16,050
as popular as its
counterpart, again, OSPF.

42
00:02:16,050 --> 00:02:22,560
However, it is very popular in
large-scale service provider

43
00:02:22,560 --> 00:02:26,710
networks, which of course, is
why we need to talk about it.

44
00:02:26,710 --> 00:02:29,580
And we'll see coming
up why that's the case.

45
00:02:29,580 --> 00:02:34,530
IS-IS supports a lot of features
that service providers find

46
00:02:34,530 --> 00:02:36,280
quite attractive and useful.

47
00:02:36,280 --> 00:02:39,510
Now, we mentioned that this
is very similar to OSPF.

48
00:02:39,510 --> 00:02:45,630
Technically, IS-IS was
actually started before OSPF.

49
00:02:45,630 --> 00:02:47,100
But we're not
really too concerned

50
00:02:47,100 --> 00:02:48,570
about which came first.

51
00:02:48,570 --> 00:02:52,450
The most important thing
here is that there are many,

52
00:02:52,450 --> 00:02:54,330
many similarities.

53
00:02:54,330 --> 00:02:56,940
They both use the
same algorithm.

54
00:02:56,940 --> 00:03:00,420
They're both link-state
routing protocols, and so on.

55
00:03:00,420 --> 00:03:03,990
We'll be talking more about
this as we go through IS-IS.

56
00:03:03,990 --> 00:03:07,770
But the point is, yes,
there are many similarities.

57
00:03:07,770 --> 00:03:09,900
In fact, the way I
usually say it-- they

58
00:03:09,900 --> 00:03:12,570
actually have more in common
than they have different.

59
00:03:12,570 --> 00:03:15,600
But that said, there are
fundamental differences,

60
00:03:15,600 --> 00:03:18,010
of course, that we
need to discuss.

61
00:03:18,010 --> 00:03:22,110
So as we just mentioned, this is
a link-state routing protocol,

62
00:03:22,110 --> 00:03:24,580
just like OSPF here.

63
00:03:24,580 --> 00:03:28,620
And as such, it has a lot
of those same attributes.

64
00:03:28,620 --> 00:03:33,540
Again, it exchanges
databases, not routes.

65
00:03:33,540 --> 00:03:37,320
Exchanging routes is
distance vector behavior.

66
00:03:37,320 --> 00:03:41,130
This is link state,
again, like OSPF.

67
00:03:41,130 --> 00:03:43,890
Routes are then
calculated on each

68
00:03:43,890 --> 00:03:47,490
IS based on that
exchange database.

69
00:03:47,490 --> 00:03:51,150
Once again, we run
Dijkstra's algorithm,

70
00:03:51,150 --> 00:03:55,440
or the shortest path first
algorithm against the database

71
00:03:55,440 --> 00:03:58,110
and come up with our own routes.

72
00:03:58,110 --> 00:04:01,710
This, just like other
modern routing protocols,

73
00:04:01,710 --> 00:04:07,270
fully supports VLSM, or
Variable-Length Subnet Masking.

74
00:04:07,270 --> 00:04:10,440
So in other words, it
is a classless protocol.

75
00:04:10,440 --> 00:04:14,130
It doesn't rely on the
traditional class A, class B,

76
00:04:14,130 --> 00:04:16,620
class C subnet structure.

77
00:04:16,620 --> 00:04:21,190
It can support any class
network with any mask.

78
00:04:21,190 --> 00:04:22,800
Now, one of the
big features making

79
00:04:22,800 --> 00:04:25,740
this very attractive
to service providers

80
00:04:25,740 --> 00:04:29,550
is this has a very, very
proven track record.

81
00:04:29,550 --> 00:04:33,750
It's very reliable
and very scalable,

82
00:04:33,750 --> 00:04:37,020
along with fast convergence.

83
00:04:37,020 --> 00:04:40,290
All of these things make this
very attractive to a service

84
00:04:40,290 --> 00:04:41,430
provider.

85
00:04:41,430 --> 00:04:43,200
We want reliability.

86
00:04:43,200 --> 00:04:44,730
We want scalability.

87
00:04:44,730 --> 00:04:46,870
And it has to be fast.

88
00:04:46,870 --> 00:04:49,890
Remember, we're going to be
using this in the service

89
00:04:49,890 --> 00:04:54,570
provider as the underlay
routing protocol to whatever

90
00:04:54,570 --> 00:04:56,200
other structure we're running--

91
00:04:56,200 --> 00:04:58,350
so for example, maybe MPLS.

92
00:04:58,350 --> 00:05:01,260
Or maybe we're just
running native BGP routing.

93
00:05:01,260 --> 00:05:02,820
But either way, we
need to know how

94
00:05:02,820 --> 00:05:05,830
to get to the next hops
of the other devices.

95
00:05:05,830 --> 00:05:08,310
So this is usually
going to be used--

96
00:05:08,310 --> 00:05:12,930
again, or OSPF-- to be the
underlay routing protocol.

97
00:05:12,930 --> 00:05:17,160
This is why we need things
like reliability, scalability,

98
00:05:17,160 --> 00:05:19,560
and of course, fast convergence.

99
00:05:19,560 --> 00:05:23,520
Now, when IS-IS was
originally designed,

100
00:05:23,520 --> 00:05:26,760
it was actually designed
to be the routing protocol

101
00:05:26,760 --> 00:05:29,820
for something called
Connectionless Network

102
00:05:29,820 --> 00:05:33,210
Service, or CLNS.

103
00:05:33,210 --> 00:05:37,770
CLNS and, of course,
therefore IS-IS

104
00:05:37,770 --> 00:05:40,890
use something called
Connectionless Network

105
00:05:40,890 --> 00:05:45,990
Protocol, or CLNP,
as the transport.

106
00:05:45,990 --> 00:05:47,880
So essentially, what
we're saying here

107
00:05:47,880 --> 00:05:51,150
is that unlike most
other routing protocols

108
00:05:51,150 --> 00:05:56,400
that we use for IP,
either IPv4 or IPv6,

109
00:05:56,400 --> 00:06:01,870
IS-IS doesn't itself
actually run on top of IP.

110
00:06:01,870 --> 00:06:04,680
Now, we have a small
environment set up right now.

111
00:06:04,680 --> 00:06:06,630
We don't really need to
look at the topology.

112
00:06:06,630 --> 00:06:09,810
We're just going to look
at the Wireshark capture

113
00:06:09,810 --> 00:06:13,980
off of one router that's
running IS-IS so we can take

114
00:06:13,980 --> 00:06:17,040
a look at how this
does not run over IP

115
00:06:17,040 --> 00:06:19,830
but rather runs over CLNS.

116
00:06:19,830 --> 00:06:21,840
So let's hop into
Wireshark and take a look

117
00:06:21,840 --> 00:06:24,600
at a capture off
of our router 1,

118
00:06:24,600 --> 00:06:26,370
which is in fact running IS-IS.

119
00:06:26,370 --> 00:06:28,930

120
00:06:28,930 --> 00:06:31,540
So here we have our
Wireshark capture.

121
00:06:31,540 --> 00:06:34,450
We're filtering based on IS-IS.

122
00:06:34,450 --> 00:06:38,350
And we can see
various IS-IS packet

123
00:06:38,350 --> 00:06:40,060
types going back and forth.

124
00:06:40,060 --> 00:06:41,890
Here we see some hellos.

125
00:06:41,890 --> 00:06:44,260
We see some CNSP packets.

126
00:06:44,260 --> 00:06:47,650
We'll get into these different
types of protocol packets

127
00:06:47,650 --> 00:06:49,760
in a future video.

128
00:06:49,760 --> 00:06:53,860
But for now, what we want to
see is here in the main capture

129
00:06:53,860 --> 00:06:58,210
window, we see that we have an
Ethernet header, an LLC header.

130
00:06:58,210 --> 00:07:03,070
And then notice it goes right to
ISO, ISIS InTRA Domain Routeing

131
00:07:03,070 --> 00:07:04,600
Information Exchange Protocol.

132
00:07:04,600 --> 00:07:08,110
If we open this up, we can see
that the protocol identifier

133
00:07:08,110 --> 00:07:10,480
is in fact ISIS.

134
00:07:10,480 --> 00:07:12,860
And if we scroll down
a little bit more here,

135
00:07:12,860 --> 00:07:16,640
we see that this, in fact, is
an ISIS HELLO that we captured.

136
00:07:16,640 --> 00:07:18,940
And we can see some
of this information.

137
00:07:18,940 --> 00:07:22,060
And we'll be covering this
in some more detail, again,

138
00:07:22,060 --> 00:07:23,720
as we go through.

139
00:07:23,720 --> 00:07:26,290
But the main point we
want to see right now

140
00:07:26,290 --> 00:07:28,600
is if we shrink
these back down, what

141
00:07:28,600 --> 00:07:31,900
we do not have is an IP header.

142
00:07:31,900 --> 00:07:35,380
This is IS-IS running
directly on top

143
00:07:35,380 --> 00:07:37,580
of Ethernet in this case.

144
00:07:37,580 --> 00:07:40,510
So now let's hop back
over to our discussion.

145
00:07:40,510 --> 00:07:44,410
And we'll pick back up here with
another big advantage to IS-IS.

146
00:07:44,410 --> 00:07:48,130
And that is that it
is very extensible

147
00:07:48,130 --> 00:07:52,870
because it uses something called
a TLV, or Type Length Value

148
00:07:52,870 --> 00:07:56,710
in order to exchange
its routing information.

149
00:07:56,710 --> 00:07:59,080
This is actually a very,
very important feature

150
00:07:59,080 --> 00:08:01,660
we have with IS-IS.

151
00:08:01,660 --> 00:08:06,190
This extensibility means
that we can support a variety

152
00:08:06,190 --> 00:08:07,660
of different protocols.

153
00:08:07,660 --> 00:08:10,630
Not only can it of
course support CLNS,

154
00:08:10,630 --> 00:08:13,870
which is what it was designed
for, although we don't use it

155
00:08:13,870 --> 00:08:15,760
for that very often these days.

156
00:08:15,760 --> 00:08:20,050
Of course, it supports
IPv4 and IPv6,

157
00:08:20,050 --> 00:08:23,570
or we probably wouldn't really
be talking about it at all.

158
00:08:23,570 --> 00:08:26,710
But it can also do things
like exchange MAC address

159
00:08:26,710 --> 00:08:28,550
information and
things like this.

160
00:08:28,550 --> 00:08:32,409
So we can use this for
other technologies such as--

161
00:08:32,409 --> 00:08:35,710
well, Cisco used it for
their FabricPath technology.

162
00:08:35,710 --> 00:08:37,960
Cisco ACI can use it.

163
00:08:37,960 --> 00:08:40,090
We can use it for VXLAN.

164
00:08:40,090 --> 00:08:41,799
So there's a lot
of different things

165
00:08:41,799 --> 00:08:45,130
that we can do with
IS-IS because it

166
00:08:45,130 --> 00:08:48,280
is so extensible
and easily updated

167
00:08:48,280 --> 00:08:50,620
to support additional
information.

168
00:08:50,620 --> 00:08:54,310
About the only other protocol
that's this flexible--

169
00:08:54,310 --> 00:08:57,070
and also why it
gets used so much--

170
00:08:57,070 --> 00:08:58,750
is BGP.

171
00:08:58,750 --> 00:09:02,210
So as far as extensibility
and flexibility,

172
00:09:02,210 --> 00:09:04,150
we can do a lot
of the same things

173
00:09:04,150 --> 00:09:09,010
both with IS-IS and BGP
because of this ability

174
00:09:09,010 --> 00:09:12,340
to just exchange
information without it

175
00:09:12,340 --> 00:09:16,360
being in such a strict format,
such as something like OSPF.

176
00:09:16,360 --> 00:09:19,570
Just to contrast,
think of it this way.

177
00:09:19,570 --> 00:09:24,910
When OSPF wanted to
go from IPv4 to IPv6,

178
00:09:24,910 --> 00:09:26,590
we needed a whole new version.

179
00:09:26,590 --> 00:09:30,420
We had to go from OSPF
version 2 to OSPF version 3.

180
00:09:30,420 --> 00:09:36,550
Protocols such as IS-IS and,
again, as an example, BGP, all

181
00:09:36,550 --> 00:09:39,050
we had to do is add
new address families.

182
00:09:39,050 --> 00:09:42,550
So again, this extensibility,
very important feature.

183
00:09:42,550 --> 00:09:45,910
Particularly moving forward
as new technologies come out

184
00:09:45,910 --> 00:09:48,490
and things like this, it
makes it a very good candidate

185
00:09:48,490 --> 00:09:53,380
if not for the end engineer to
use at least for the underlying

186
00:09:53,380 --> 00:09:57,340
technologies to use it, such
as controllers in something

187
00:09:57,340 --> 00:10:00,650
like software-defined
networking.

188
00:10:00,650 --> 00:10:02,900
Some other information
we want to get at here

189
00:10:02,900 --> 00:10:08,480
is IS-IS actually supports
two different metric types.

190
00:10:08,480 --> 00:10:13,010
Specifically, it supports what's
called the narrow-style metric.

191
00:10:13,010 --> 00:10:15,560
And this is really
the legacy style.

192
00:10:15,560 --> 00:10:17,750
We really don't use
this too much anymore.

193
00:10:17,750 --> 00:10:19,320
You'll see why in just a moment.

194
00:10:19,320 --> 00:10:22,010
It's still the default
on most platforms.

195
00:10:22,010 --> 00:10:24,770
So since we don't
really want to use it,

196
00:10:24,770 --> 00:10:26,990
we'll see that when we
get into our configuration

197
00:10:26,990 --> 00:10:30,515
later that we're going to most
likely want to change this.

198
00:10:30,515 --> 00:10:37,200
This uses a 6-bit link metric
and a 10-bit path metric,

199
00:10:37,200 --> 00:10:40,670
which in and of itself
makes it fairly limiting.

200
00:10:40,670 --> 00:10:43,310
But the more important
thing that makes it limiting

201
00:10:43,310 --> 00:10:48,440
is compared to the newer
wide-style metric, which,

202
00:10:48,440 --> 00:10:51,600
like we just said, is
the new style, of course,

203
00:10:51,600 --> 00:10:56,490
this one is required
for IPv6 support.

204
00:10:56,490 --> 00:11:00,020
And since even if today
we're not doing IPv6,

205
00:11:00,020 --> 00:11:02,250
we probably will be shortly.

206
00:11:02,250 --> 00:11:04,920
This is almost always
something we want.

207
00:11:04,920 --> 00:11:07,610
This is a 24-bit
link metric, allowing

208
00:11:07,610 --> 00:11:09,440
us to be a lot more granular.

209
00:11:09,440 --> 00:11:14,280
And it's a 32-bit path metric.

210
00:11:14,280 --> 00:11:17,460
In this video, we
jumped in and looked

211
00:11:17,460 --> 00:11:20,490
at some of the basics of IS-IS.

212
00:11:20,490 --> 00:11:23,400
We started off by looking
at some of its advantages.

213
00:11:23,400 --> 00:11:25,680
And then we saw how
it doesn't actually

214
00:11:25,680 --> 00:11:29,250
run on IP, because it
was designed for CLNS,

215
00:11:29,250 --> 00:11:36,030
but has been extended to support
not only CLNS but also IP

216
00:11:36,030 --> 00:11:39,660
versions, of course, 4 and 6.

217
00:11:39,660 --> 00:11:42,210
Then we wrapped up here
by talking about the two

218
00:11:42,210 --> 00:11:46,000
different metric styles we have
and how, generally speaking,

219
00:11:46,000 --> 00:11:49,020
we want to prefer
the wide metrics.

220
00:11:49,020 --> 00:11:51,150
I hope this has been
informative for you,

221
00:11:51,150 --> 00:11:54,200
and I'd like to thank
you for viewing.

222
00:11:54,200 --> 00:11:58,000