1
00:00:00,000 --> 00:00:03,792
[AUDIO LOGO]

2
00:00:03,792 --> 00:00:08,070

3
00:00:08,070 --> 00:00:11,010
So let's jump in and
take a closer look

4
00:00:11,010 --> 00:00:14,970
at VPLS, or Virtual
Private LAN Service.

5
00:00:14,970 --> 00:00:18,600
The really cool advantage to
this feature from a customer's

6
00:00:18,600 --> 00:00:22,350
perspective is they really
see the service provider's

7
00:00:22,350 --> 00:00:26,670
environment as like one
big Ethernet switch.

8
00:00:26,670 --> 00:00:29,280
And that can be very
advantageous for them

9
00:00:29,280 --> 00:00:31,890
because everything's
on one subnet then.

10
00:00:31,890 --> 00:00:35,370
All of their routers can be
OSPF neighbors with each other

11
00:00:35,370 --> 00:00:38,890
or whatever routing protocol
they're running and so on.

12
00:00:38,890 --> 00:00:42,630
So let's, in this video,
jump in and take a look

13
00:00:42,630 --> 00:00:45,120
at the details of VPLS.

14
00:00:45,120 --> 00:00:46,890
And then in the
next video, we'll

15
00:00:46,890 --> 00:00:50,370
be jumping onto the command line
and putting this into action.

16
00:00:50,370 --> 00:00:53,220
But for now, let's go ahead
and jump in and take a look

17
00:00:53,220 --> 00:00:55,050
at some of the details of VPLS.

18
00:00:55,050 --> 00:00:57,920

19
00:00:57,920 --> 00:01:03,880
So at its core, VPLS is a
multipoint Ethernet connection

20
00:01:03,880 --> 00:01:04,959
service.

21
00:01:04,959 --> 00:01:07,330
So, essentially,
what that means is

22
00:01:07,330 --> 00:01:10,960
it gives us a way to connect
remote sites together

23
00:01:10,960 --> 00:01:13,540
as if they were
connected to a single LAN

24
00:01:13,540 --> 00:01:18,880
segment, so on the same VLAN
or the same layer 2 connection.

25
00:01:18,880 --> 00:01:23,530
And we can create either
one or more emulated LANs

26
00:01:23,530 --> 00:01:25,090
for each customer.

27
00:01:25,090 --> 00:01:28,180
They might have a need to
have different connections

28
00:01:28,180 --> 00:01:31,810
effectively the same as
having different VLANs.

29
00:01:31,810 --> 00:01:35,350
Each of these emulated
LANs is going to simply

30
00:01:35,350 --> 00:01:37,810
look like a giant
Ethernet switch

31
00:01:37,810 --> 00:01:40,120
as far as the
customer is concerned.

32
00:01:40,120 --> 00:01:43,180
Again, just like on a
regular Ethernet switch,

33
00:01:43,180 --> 00:01:46,330
all of these devices will be
able to talk to each other

34
00:01:46,330 --> 00:01:48,250
at layer 2.

35
00:01:48,250 --> 00:01:51,970
If we create numerous
emulated LANs,

36
00:01:51,970 --> 00:01:53,860
then they're going
to be separate.

37
00:01:53,860 --> 00:01:58,270
Again, much in the same way that
VLANs would separate traffic,

38
00:01:58,270 --> 00:02:00,040
the different emulated
LANs are going

39
00:02:00,040 --> 00:02:02,470
to separate the traffic as well.

40
00:02:02,470 --> 00:02:06,280
For each of these
LANs, the PE routers

41
00:02:06,280 --> 00:02:10,090
are going to have a
full mesh of pseudowires

42
00:02:10,090 --> 00:02:12,650
because, of course,
that's how this works.

43
00:02:12,650 --> 00:02:14,650
So to keep the
traffic separated,

44
00:02:14,650 --> 00:02:19,090
we're going to have separate
meshes of pseudowires.

45
00:02:19,090 --> 00:02:24,650
Functionally speaking, when
the PE receives a frame,

46
00:02:24,650 --> 00:02:27,670
it will look at the
destination MAC address

47
00:02:27,670 --> 00:02:32,050
and forward it to the
correct egress PE.

48
00:02:32,050 --> 00:02:35,480
Now, that, of course, is going
to bring up the question, well,

49
00:02:35,480 --> 00:02:39,550
how does it know what is
the correct egress PE?

50
00:02:39,550 --> 00:02:44,230
And the answer to that is it
learns the MAC addresses much

51
00:02:44,230 --> 00:02:48,070
in the same way as a
traditional switch would.

52
00:02:48,070 --> 00:02:52,090
Initially, when we don't
know any MAC addresses yet,

53
00:02:52,090 --> 00:02:54,970
traffic is sent to
all the pseudowires.

54
00:02:54,970 --> 00:02:57,550
This is the equivalent
of doing flooding

55
00:02:57,550 --> 00:03:00,040
on a regular Ethernet switch.

56
00:03:00,040 --> 00:03:02,350
Also just like a
traditional switch,

57
00:03:02,350 --> 00:03:04,630
we are then going to
be learning as we're

58
00:03:04,630 --> 00:03:09,190
forwarding traffic based on
the source MAC addresses.

59
00:03:09,190 --> 00:03:14,440
As traffic comes in, the source
MAC is mapped to the interface

60
00:03:14,440 --> 00:03:15,760
that it arrived on.

61
00:03:15,760 --> 00:03:18,190
Now, that could either
be a physical interface

62
00:03:18,190 --> 00:03:22,870
if it's a locally connected
device or a pseudowire if we're

63
00:03:22,870 --> 00:03:26,050
learning it across
the MPLS environment.

64
00:03:26,050 --> 00:03:29,590
Once we've learned
that, well, then traffic

65
00:03:29,590 --> 00:03:33,190
can be forwarded just to
the appropriate interface.

66
00:03:33,190 --> 00:03:35,440
And, again, much like
a traditional switch,

67
00:03:35,440 --> 00:03:40,300
we are going to use split
horizon to prevent loops.

68
00:03:40,300 --> 00:03:43,300
Broadcast and
multicast traffic are

69
00:03:43,300 --> 00:03:48,070
going to be flooded to all
of the sites and ports.

70
00:03:48,070 --> 00:03:51,400
This, of course, is because
we filter traffic based

71
00:03:51,400 --> 00:03:53,740
on the destination MAC address.

72
00:03:53,740 --> 00:03:58,990
And a broadcast MAC address
or a multicast MAC address

73
00:03:58,990 --> 00:04:01,990
will never be the source.

74
00:04:01,990 --> 00:04:05,630
And since it's not the source,
that means we never learn it.

75
00:04:05,630 --> 00:04:07,750
And if we never
learn it, then we

76
00:04:07,750 --> 00:04:10,310
can't filter based
on that address.

77
00:04:10,310 --> 00:04:13,120
So, therefore, it's
going to get flooded.

78
00:04:13,120 --> 00:04:17,380
That means that we're going
to be relying on split horizon

79
00:04:17,380 --> 00:04:20,200
once again to make sure
that we're forwarding this

80
00:04:20,200 --> 00:04:24,970
across a loop-free tree and not
actually looping our broadcast

81
00:04:24,970 --> 00:04:26,960
or multicast traffic.

82
00:04:26,960 --> 00:04:30,190
Now, on the customer
side, the CE device

83
00:04:30,190 --> 00:04:33,700
can actually be a
switch or a router.

84
00:04:33,700 --> 00:04:35,920
Generally speaking,
we would prefer

85
00:04:35,920 --> 00:04:38,420
the customer use a router.

86
00:04:38,420 --> 00:04:40,240
This is going to be recommended.

87
00:04:40,240 --> 00:04:42,460
Now, here, we say in
a larger environment.

88
00:04:42,460 --> 00:04:44,980
But, honestly, really
in any environment,

89
00:04:44,980 --> 00:04:47,740
if they can use a router,
that would be preferred.

90
00:04:47,740 --> 00:04:50,530
And the reason for
that is we wouldn't

91
00:04:50,530 --> 00:04:55,420
have to learn all those
MAC addresses on the PEs.

92
00:04:55,420 --> 00:04:58,780
Otherwise, potentially,
the PEs may

93
00:04:58,780 --> 00:05:02,620
have to learn every MAC address
in the customer's network.

94
00:05:02,620 --> 00:05:05,230
If there's a router
between, then we just

95
00:05:05,230 --> 00:05:09,010
have to learn the MAC
address of the router.

96
00:05:09,010 --> 00:05:13,050
So in this video, we got some
more detailed information

97
00:05:13,050 --> 00:05:18,420
on how VPLS works, how it acts
as a switch for the customer,

98
00:05:18,420 --> 00:05:21,930
and how it actually handles
the forwarding of traffic,

99
00:05:21,930 --> 00:05:24,720
both if we have a
single emulated LAN

100
00:05:24,720 --> 00:05:26,910
or if we have more than one.

101
00:05:26,910 --> 00:05:30,780
Now, of course, it's time to
actually implement this and get

102
00:05:30,780 --> 00:05:31,530
it working.

103
00:05:31,530 --> 00:05:34,240
And that's exactly
what we get to do next.

104
00:05:34,240 --> 00:05:37,440
Until then, I hope this has
been informative for you.

105
00:05:37,440 --> 00:05:40,460
And I'd like to thank
you for viewing.

106
00:05:40,460 --> 00:05:44,000