1
00:00:00,000 --> 00:00:07,170
[AUDIO LOGO]

2
00:00:07,170 --> 00:00:12,120
So now, let's jump in and
talk about multi-area OSPF.

3
00:00:12,120 --> 00:00:16,050
This is going to be used two
address our scalability issues

4
00:00:16,050 --> 00:00:18,030
with OSPF.

5
00:00:18,030 --> 00:00:21,900
Now understand, this still won't
scale as large as something

6
00:00:21,900 --> 00:00:23,640
like BGP.

7
00:00:23,640 --> 00:00:26,850
It still can't carry hundreds
of thousands or millions

8
00:00:26,850 --> 00:00:27,810
of routes.

9
00:00:27,810 --> 00:00:32,340
However, it will let us scale it
up to a much larger environment

10
00:00:32,340 --> 00:00:35,350
than we could with
just a single area.

11
00:00:35,350 --> 00:00:38,340
It also brings to
the table the ability

12
00:00:38,340 --> 00:00:40,860
to do route summarization.

13
00:00:40,860 --> 00:00:44,370
Keep in mind as we talked
about in the previous video,

14
00:00:44,370 --> 00:00:47,310
all of the routers
in an area have

15
00:00:47,310 --> 00:00:50,490
to have the same exact
link-state database.

16
00:00:50,490 --> 00:00:55,020
And, of course, a side effect of
that is I can't have one router

17
00:00:55,020 --> 00:01:00,610
have a /16 for a network while
another router has a /24.

18
00:01:00,610 --> 00:01:04,290
So since they have to
have identical databases,

19
00:01:04,290 --> 00:01:07,900
we cannot summarize
within an area.

20
00:01:07,900 --> 00:01:11,760
So we have to have
multi-area OSPF in order

21
00:01:11,760 --> 00:01:13,830
to do proper summarization.

22
00:01:13,830 --> 00:01:16,980
That also implies
that our summarization

23
00:01:16,980 --> 00:01:21,763
points in our network also have
to fall on area boundaries.

24
00:01:21,763 --> 00:01:24,180
So let's go ahead and take a
look at some of these things.

25
00:01:24,180 --> 00:01:26,040
We'll start by
taking a look at some

26
00:01:26,040 --> 00:01:28,470
of the attributes
of multi-area OSPF.

27
00:01:28,470 --> 00:01:30,090
And then, of course,
we'll jump in

28
00:01:30,090 --> 00:01:32,310
and take a look at
what our databases look

29
00:01:32,310 --> 00:01:34,680
like with multiple areas.

30
00:01:34,680 --> 00:01:36,700
Let's get started.

31
00:01:36,700 --> 00:01:39,160
So as we said during
the introduction,

32
00:01:39,160 --> 00:01:44,250
the biggest thing that
multi-area OSPF brings to OSPF

33
00:01:44,250 --> 00:01:46,860
is really that scalability.

34
00:01:46,860 --> 00:01:50,070
And again, the reason this
helps is because remember,

35
00:01:50,070 --> 00:01:52,980
that all of the
routers in a given area

36
00:01:52,980 --> 00:01:56,970
have to have identical
link-state databases.

37
00:01:56,970 --> 00:01:59,730
And by having
multiple areas, that

38
00:01:59,730 --> 00:02:05,220
means that SPF is only
run for database changes

39
00:02:05,220 --> 00:02:07,360
within that area.

40
00:02:07,360 --> 00:02:10,889
So the way that it
works in between areas

41
00:02:10,889 --> 00:02:14,910
is we only carry
prefix information

42
00:02:14,910 --> 00:02:16,770
between those areas.

43
00:02:16,770 --> 00:02:22,170
We do not carry the full
database information.

44
00:02:22,170 --> 00:02:26,330
Now, we'll talk a lot more about
these LSAs, the different types

45
00:02:26,330 --> 00:02:29,750
of LSAs within the area
and between the areas

46
00:02:29,750 --> 00:02:31,920
in an upcoming video.

47
00:02:31,920 --> 00:02:37,250
But for now, just realize that
the LSAs sent between areas

48
00:02:37,250 --> 00:02:39,260
carry less information.

49
00:02:39,260 --> 00:02:44,690
And we don't actually run
Dijkstra against those LSAs.

50
00:02:44,690 --> 00:02:49,450
And we'll see this in action
in just a few moments.

51
00:02:49,450 --> 00:02:53,850
Now, one of our requirements
when using multiple areas is we

52
00:02:53,850 --> 00:02:58,670
do have to follow a
very strict hierarchy.

53
00:02:58,670 --> 00:03:01,700
The reason for this
is as we mentioned

54
00:03:01,700 --> 00:03:04,580
during the introduction,
the only place

55
00:03:04,580 --> 00:03:09,920
we can summarize in OSPF
is at these area borders.

56
00:03:09,920 --> 00:03:14,750
And because of that, this has
to follow our IP structure

57
00:03:14,750 --> 00:03:15,690
as well.

58
00:03:15,690 --> 00:03:18,680
So that where our
areas cross, this

59
00:03:18,680 --> 00:03:20,420
is also the place
that we can actually

60
00:03:20,420 --> 00:03:23,180
do our network summarization.

61
00:03:23,180 --> 00:03:27,380
So this does require a
very specific design.

62
00:03:27,380 --> 00:03:29,260
Now, not only does
this have to follow

63
00:03:29,260 --> 00:03:32,260
our IP addressing scheme,
but this hierarchy

64
00:03:32,260 --> 00:03:34,190
does have some rules of its own.

65
00:03:34,190 --> 00:03:39,350
The first, of course, is that
there must be a backbone area.

66
00:03:39,350 --> 00:03:43,990
This is an area that all of the
other areas have to connect to.

67
00:03:43,990 --> 00:03:45,970
This has a few guidelines.

68
00:03:45,970 --> 00:03:48,430
First, it must be area zero.

69
00:03:48,430 --> 00:03:51,460
This is not an option in OSPF.

70
00:03:51,460 --> 00:03:55,000
As soon as you have more
than one area, one of them

71
00:03:55,000 --> 00:03:57,400
has to be area zero.

72
00:03:57,400 --> 00:04:00,250
It also must be contiguous.

73
00:04:00,250 --> 00:04:04,930
In other words, you cannot
have two separate area zeros.

74
00:04:04,930 --> 00:04:09,190
So you cannot have, for
example, an area zero here,

75
00:04:09,190 --> 00:04:12,580
have a net router between them,
have this connected to, say,

76
00:04:12,580 --> 00:04:15,670
area 21 with another router.

77
00:04:15,670 --> 00:04:19,240
And then over here,
do another area zero.

78
00:04:19,240 --> 00:04:22,330
This would be a split area zero.

79
00:04:22,330 --> 00:04:24,020
This is not allowed.

80
00:04:24,020 --> 00:04:26,800
Now, there are ways we can
make this work, which we will

81
00:04:26,800 --> 00:04:29,260
talk about in future skills.

82
00:04:29,260 --> 00:04:32,380
But in making it work,
what we're really doing

83
00:04:32,380 --> 00:04:35,500
is fixing the fact
that it's split

84
00:04:35,500 --> 00:04:37,810
and making it so
that it's not split.

85
00:04:37,810 --> 00:04:39,640
That's the actual fix for it.

86
00:04:39,640 --> 00:04:43,370
And the last point here is not
so much a rule as a guideline.

87
00:04:43,370 --> 00:04:48,640
We don't generally put end user
devices or resource networks,

88
00:04:48,640 --> 00:04:51,850
such as servers, or
printers, or things

89
00:04:51,850 --> 00:04:55,210
like that in the backbone area.

90
00:04:55,210 --> 00:04:57,370
It certainly can be done.

91
00:04:57,370 --> 00:05:00,670
Area zero is a full
blown OSPF area.

92
00:05:00,670 --> 00:05:02,170
So you can do this.

93
00:05:02,170 --> 00:05:06,580
But area zero is very often
reserved to just be a backbone

94
00:05:06,580 --> 00:05:10,300
and, therefore, not actually
have any resources in it.

95
00:05:10,300 --> 00:05:14,770
All of the other areas are
then non-backbone areas.

96
00:05:14,770 --> 00:05:17,890
Now this, of course, can
use any 32-bit number.

97
00:05:17,890 --> 00:05:21,820
They all, as we said earlier,
must be connected to area zero.

98
00:05:21,820 --> 00:05:26,200
And this is usually where
we provide those connections

99
00:05:26,200 --> 00:05:28,360
for the end users.

100
00:05:28,360 --> 00:05:32,450
And those resource networks
that we spoke of up above.

101
00:05:32,450 --> 00:05:34,390
So quite commonly,
what you would have,

102
00:05:34,390 --> 00:05:37,390
you would have an area
zero, connect it out

103
00:05:37,390 --> 00:05:40,240
to some other area,
say, 21 again.

104
00:05:40,240 --> 00:05:42,430
And this is where
we would actually

105
00:05:42,430 --> 00:05:44,653
connect in our resources.

106
00:05:44,653 --> 00:05:46,570
Let's take a look at
this in a little bit more

107
00:05:46,570 --> 00:05:48,020
of a structure diagram.

108
00:05:48,020 --> 00:05:52,210
So here, we have a typical
multi-area OSPF design

109
00:05:52,210 --> 00:05:55,600
where, of course, we have
our backbone area, area zero,

110
00:05:55,600 --> 00:06:00,370
as well as two other areas,
area 100 and area 200.

111
00:06:00,370 --> 00:06:02,530
So the main thing we
want to look at here

112
00:06:02,530 --> 00:06:06,190
are the different roles of the
routers when we start getting

113
00:06:06,190 --> 00:06:08,780
into a multi-area topology.

114
00:06:08,780 --> 00:06:11,500
So the first thing
we have is something

115
00:06:11,500 --> 00:06:14,170
called a backbone router.

116
00:06:14,170 --> 00:06:16,540
That would be this
router here, which

117
00:06:16,540 --> 00:06:20,380
is a router that has all of
its interfaces in the backbone

118
00:06:20,380 --> 00:06:21,040
area.

119
00:06:21,040 --> 00:06:25,690
Other routers would be things
like internal routers, which

120
00:06:25,690 --> 00:06:30,010
are routers such as these that
have all of their interfaces

121
00:06:30,010 --> 00:06:33,640
in an area other than area zero.

122
00:06:33,640 --> 00:06:37,750
And in case you're wondering
here, yes, a backbone router

123
00:06:37,750 --> 00:06:40,000
is also an internal router.

124
00:06:40,000 --> 00:06:42,250
We just give it a
more specific name

125
00:06:42,250 --> 00:06:44,110
because all of its
interfaces happen

126
00:06:44,110 --> 00:06:46,690
to be in the backbone
area, area zero.

127
00:06:46,690 --> 00:06:49,730
But yes, it is still
an internal router.

128
00:06:49,730 --> 00:06:53,170
Now, the other routers here
are a little more specialized.

129
00:06:53,170 --> 00:06:58,600
The first thing we have is an
ABR or an Area Border Router.

130
00:06:58,600 --> 00:07:02,110
And that would be this
router and this router.

131
00:07:02,110 --> 00:07:06,760
These are routers that have one
interface at least in area zero

132
00:07:06,760 --> 00:07:12,160
and one interface in some other
area, possibly more than one.

133
00:07:12,160 --> 00:07:14,800
But nonetheless, these
are then performing

134
00:07:14,800 --> 00:07:18,670
the job of an ABR or
an area border router.

135
00:07:18,670 --> 00:07:21,610
This, by the way, is
where we can actually

136
00:07:21,610 --> 00:07:23,590
do our network summarization.

137
00:07:23,590 --> 00:07:28,930
So the ABR is where we can
summarize internal routes.

138
00:07:28,930 --> 00:07:31,540
Now, we have another type
of router here bringing

139
00:07:31,540 --> 00:07:33,830
in routes from the outside.

140
00:07:33,830 --> 00:07:38,650
So maybe, for example, this
would be something like BGP.

141
00:07:38,650 --> 00:07:41,710
Maybe this would be
something like EIGRP

142
00:07:41,710 --> 00:07:45,010
or something like IS-IS.

143
00:07:45,010 --> 00:07:47,410
The router performing
these functions

144
00:07:47,410 --> 00:07:51,940
is called an ASBR or an
Autonomous System Boundary

145
00:07:51,940 --> 00:07:53,030
Router.

146
00:07:53,030 --> 00:07:57,970
This is a router that's doing
some form of redistribution

147
00:07:57,970 --> 00:08:02,320
from these external
protocols into OSPF

148
00:08:02,320 --> 00:08:05,510
bringing external
routes into our network.

149
00:08:05,510 --> 00:08:10,220
This is where we can actually
summarize external routes.

150
00:08:10,220 --> 00:08:15,760
So technically, we can summarize
at the ABR or at the ASBR,

151
00:08:15,760 --> 00:08:18,880
depending on which type of
route we're talking about.

152
00:08:18,880 --> 00:08:24,190
The ABR, again, is for internal
routes going between areas.

153
00:08:24,190 --> 00:08:28,510
The ASBR can only summarize
the external routes

154
00:08:28,510 --> 00:08:33,549
as they're being brought into
OSPF during the redistribution.

155
00:08:33,549 --> 00:08:37,360
So now, let's wrap up with
a look at some advantages

156
00:08:37,360 --> 00:08:42,700
and disadvantages of multi-area
versus single-area OSPF.

157
00:08:42,700 --> 00:08:45,770
So let's begin with
the advantages.

158
00:08:45,770 --> 00:08:51,640
First, we have that it allows
for filtering and summarization

159
00:08:51,640 --> 00:08:53,360
between the areas.

160
00:08:53,360 --> 00:08:55,870
Now, we've been talking a
bit about the summarization

161
00:08:55,870 --> 00:08:57,010
as we go here.

162
00:08:57,010 --> 00:09:00,010
We haven't mentioned a
whole lot of filtering.

163
00:09:00,010 --> 00:09:03,520
But for the exact same
reason that we can only

164
00:09:03,520 --> 00:09:07,330
summarize between
areas, the ABR is also

165
00:09:07,330 --> 00:09:09,410
the only place we can filter.

166
00:09:09,410 --> 00:09:11,890
So again, with
multi-area we can do

167
00:09:11,890 --> 00:09:15,700
route filtering between the
areas as well as summarization.

168
00:09:15,700 --> 00:09:21,080
It also localizes the impact
of any topology changes.

169
00:09:21,080 --> 00:09:23,950
So if an interface goes
up, an interface goes down,

170
00:09:23,950 --> 00:09:26,980
or more importantly, if
an interface is flapping,

171
00:09:26,980 --> 00:09:30,640
it won't affect the
entire OSPF domain.

172
00:09:30,640 --> 00:09:34,690
It would be localized just
to the single area where

173
00:09:34,690 --> 00:09:36,070
the problem is occurring.

174
00:09:36,070 --> 00:09:39,070
Part of that also
relates to the containing

175
00:09:39,070 --> 00:09:41,860
of the flooded detailed LSAs.

176
00:09:41,860 --> 00:09:43,840
But we're really
talking about containing

177
00:09:43,840 --> 00:09:47,230
the flooding of those
detailed LSAs regardless

178
00:09:47,230 --> 00:09:49,670
of whether it's because
of a topology change,

179
00:09:49,670 --> 00:09:51,160
but just in general.

180
00:09:51,160 --> 00:09:53,830
This is what causes
that larger database

181
00:09:53,830 --> 00:09:57,830
and for the longer execution
of Dijkstra's algorithm.

182
00:09:57,830 --> 00:10:00,190
Now, these are really
good advantages.

183
00:10:00,190 --> 00:10:02,890
Very, very important that
we keep these in mind.

184
00:10:02,890 --> 00:10:05,350
And quite frankly, these
are most of the reasons

185
00:10:05,350 --> 00:10:08,290
that we're going to go
with multi-area OSPF.

186
00:10:08,290 --> 00:10:11,650
Now, of course, scalability
is also a huge advantage.

187
00:10:11,650 --> 00:10:14,000
But I was trying to go for
some more specifics here.

188
00:10:14,000 --> 00:10:16,210
Now, there are some
concerns, however,

189
00:10:16,210 --> 00:10:19,190
with running multi-area
OSPF as well.

190
00:10:19,190 --> 00:10:21,730
You could also refer
to these as advantages

191
00:10:21,730 --> 00:10:24,910
to running single area, however
you want to look at this.

192
00:10:24,910 --> 00:10:30,220
But the first one is that
the routers no longer have

193
00:10:30,220 --> 00:10:32,840
full visibility of the network.

194
00:10:32,840 --> 00:10:36,370
Now, we'll get into this more
as we go through future skills.

195
00:10:36,370 --> 00:10:40,180
But since the routers no
longer have full visibility,

196
00:10:40,180 --> 00:10:43,930
this can actually
lead to routing loops.

197
00:10:43,930 --> 00:10:49,180
Link-state routing protocols
are immune to routing loops.

198
00:10:49,180 --> 00:10:52,570
Because they have
that full visibility.

199
00:10:52,570 --> 00:10:54,700
But without that
full visibility,

200
00:10:54,700 --> 00:10:56,690
you lose that immunity.

201
00:10:56,690 --> 00:11:00,280
So as we start digging into
more detail with multi-area

202
00:11:00,280 --> 00:11:02,980
and how our ABRs
work and such, we'll

203
00:11:02,980 --> 00:11:06,760
get more into how this actually
has to do loop prevention

204
00:11:06,760 --> 00:11:10,270
and implement some
rules on our ABRs.

205
00:11:10,270 --> 00:11:12,760
And further talking
about those ABRs,

206
00:11:12,760 --> 00:11:14,920
we need to make sure
that they don't become

207
00:11:14,920 --> 00:11:17,590
a single point of
failure, which means

208
00:11:17,590 --> 00:11:20,950
we have to worry about
redundancy with those ABRs.

209
00:11:20,950 --> 00:11:24,880
Meaning, we should have at
least two ABRs between our areas

210
00:11:24,880 --> 00:11:26,450
in most cases.

211
00:11:26,450 --> 00:11:29,800
And finally, there
are some technologies

212
00:11:29,800 --> 00:11:34,300
that just don't work as well
without having a single area.

213
00:11:34,300 --> 00:11:39,790
One of those as an example would
be MPLS traffic engineering.

214
00:11:39,790 --> 00:11:43,180
Can we get it to
work between areas?

215
00:11:43,180 --> 00:11:44,320
Yes.

216
00:11:44,320 --> 00:11:47,170
Is it much more difficult
to configure and set up?

217
00:11:47,170 --> 00:11:48,070
A bit.

218
00:11:48,070 --> 00:11:51,310
So it's certainly
more preferred to have

219
00:11:51,310 --> 00:11:55,030
a single area for things
like traffic engineering,

220
00:11:55,030 --> 00:11:57,660
if it can be accomplished.

221
00:11:57,660 --> 00:11:59,970
In this video, we jumped in.

222
00:11:59,970 --> 00:12:02,670
And we took a look at
some of the advantages

223
00:12:02,670 --> 00:12:07,530
and a few of the disadvantages
of running multi-area OSPF.

224
00:12:07,530 --> 00:12:11,610
Keep in mind its main
advantage is scalability.

225
00:12:11,610 --> 00:12:13,920
Because at a certain
point, we simply

226
00:12:13,920 --> 00:12:17,910
can't run Dijkstra against
a really large database

227
00:12:17,910 --> 00:12:21,190
and still have an
efficient network.

228
00:12:21,190 --> 00:12:23,160
I hope this has been
informative for you.

229
00:12:23,160 --> 00:12:26,120
And I'd like to thank
you for viewing.

230
00:12:26,120 --> 00:12:29,000