1
00:00:00,420 --> 00:00:06,450
Now you're not expected to know this for the CCMA exam but I'm going to demonstrate an example of quality

2
00:00:06,450 --> 00:00:10,580
of service on two Cisco routers in this example.

3
00:00:10,580 --> 00:00:18,740
I've got two routers running in genius three show run interface serial to slash zero shows us the configuration

4
00:00:19,490 --> 00:00:26,780
these routers simply have IP addresses configured and I have no shut to the interfaces no other configuration

5
00:00:26,780 --> 00:00:28,070
has been done.

6
00:00:28,250 --> 00:00:35,150
Show interface several to slash zero shows us the speed of the interface which is using the default

7
00:00:35,240 --> 00:00:45,560
of one point five four megabits per second so on this interface of the router we can use the command

8
00:00:45,680 --> 00:00:49,290
auto cores.

9
00:00:49,490 --> 00:00:57,090
Now notice we do have the discovery option where we can configure the Rada to discover traffic that

10
00:00:57,090 --> 00:00:59,070
traverses that interface.

11
00:00:59,250 --> 00:01:05,790
But in this example I'm simply going to enable the interface for VoIP and I'm not going to specify the

12
00:01:05,790 --> 00:01:13,220
option Trust which means it's not going to trust the markings that it receives from Roorda to now typically

13
00:01:13,220 --> 00:01:21,080
in the real world you would do that you would trust traffic from one writer to another unless rather

14
00:01:21,080 --> 00:01:25,820
1 is a service provider and wrote it to as a customer as an example.

15
00:01:25,820 --> 00:01:31,850
Now here I'm going to use order cause VoIP trust so that you can see the difference between trusting

16
00:01:32,330 --> 00:01:36,730
and not trusting a neighboring router.

17
00:01:36,800 --> 00:01:44,390
So rather one doesn't trust it to write it to trusts rather one based on the order cost configuration

18
00:01:45,730 --> 00:01:55,090
now order quotes creates a large amount of configuration so rather than you manually typing configuration

19
00:01:56,270 --> 00:02:05,250
it's done for you because write it to trusts rather one it's created a quality of service class map

20
00:02:05,910 --> 00:02:14,850
that matches DCF so essentially if it to received traffic from write a one marked as E.F. it's going

21
00:02:14,850 --> 00:02:24,560
to go into this order cause VoIP RTP trust CLOs that's referenced by this policy map the configuration

22
00:02:24,560 --> 00:02:31,550
that you see here is an example of in QC or the modular quality of service command line interface it

23
00:02:31,550 --> 00:02:33,260
consists of two main parts.

24
00:02:33,260 --> 00:02:43,000
We've got class maps and we've got policy maps and policy maps then bound onto interfaces so in this

25
00:02:43,000 --> 00:02:52,780
example we've got the service policy which references the policy map bound to the serial to slash the

26
00:02:52,780 --> 00:03:05,890
zero interface so going back to our class maps we've got two class maps one matching DCP E.F. and one

27
00:03:05,890 --> 00:03:07,940
matching either.

28
00:03:07,960 --> 00:03:16,690
Notice the match any keyword class selector 3 or IP precedence 3 and assured forwarding Class 3 1 or

29
00:03:16,690 --> 00:03:23,240
a 3 1 rather tooth trusting the markings that it receives from reader 1.

30
00:03:23,320 --> 00:03:32,280
So it's looking for traffic that matches E.F. and that class is then matched within this policy map.

31
00:03:32,380 --> 00:03:40,240
So notice the name or the quoted voice RTP trust is matched here and it's given 70 percent of the interface

32
00:03:40,240 --> 00:03:40,800
bandwidth.

33
00:03:40,840 --> 00:03:42,520
When there's congestion.

34
00:03:42,850 --> 00:03:50,770
So in other words voice traffic match by E.F. will be given a priority bandwidth of 70 percent.

35
00:03:50,770 --> 00:03:52,210
That's a priority queue.

36
00:03:53,200 --> 00:03:56,660
This is an example of low latency queuing.

37
00:03:56,820 --> 00:04:02,550
We've got classes of traffic but we've also got a priority key.

38
00:04:02,550 --> 00:04:08,910
So this is the priority key in low latency killing of voice will be given 70 percent of the bandwidth

39
00:04:09,750 --> 00:04:10,980
other traffic.

40
00:04:10,980 --> 00:04:17,790
So in this case we've got a VoIP control is given 5 percent of the bandwidth notice.

41
00:04:17,790 --> 00:04:25,920
This matches the class configured over here which is matching C S three and a half 31.

42
00:04:26,250 --> 00:04:33,070
That's essentially called signalling protocols such as sip H3 3G 3 and skinny.

43
00:04:33,120 --> 00:04:39,890
In other words those are protocols you use to set up telephone calls so notice what order quizzes done.

44
00:04:39,900 --> 00:04:50,030
It's allocated 75 percent of the interface bandwidth to voice calls so the actual voice traffic the

45
00:04:50,030 --> 00:04:59,300
RTP traffic gets 70 percent call signaling gets 5 percent and the remaining traffic is shared using

46
00:04:59,300 --> 00:05:00,950
fake queuing.

47
00:05:00,980 --> 00:05:06,730
So this is an example of waited for queuing on the default class default.

48
00:05:06,730 --> 00:05:15,550
Class is going to match anything not explicitly matched so a low latency queue is actually pretty key.

49
00:05:15,640 --> 00:05:20,160
Class based weighted fake queuing but that's too much of a mouthful.

50
00:05:20,470 --> 00:05:22,620
So we call it low latency queuing.

51
00:05:22,810 --> 00:05:24,600
There's the pretty key.

52
00:05:24,590 --> 00:05:32,660
Here's an example of a class and this is an example of waited for queuing within a class low latency

53
00:05:32,660 --> 00:05:38,420
queuing is pretty queuing class based waited for queuing.

54
00:05:38,780 --> 00:05:44,660
And in this example we approach rising VoIP traffic over other traffic types.

55
00:05:44,660 --> 00:05:48,930
70 percent of the bandwidth can be taken by voice calls.

56
00:05:49,010 --> 00:05:51,900
It's a priority queue so it will be serviced first.

57
00:05:52,010 --> 00:05:55,900
5 percent of the interface bandwidth could be taken by call signalling.

58
00:05:56,330 --> 00:06:00,230
But this is a guaranteed minimum bandwidth not a priority.

59
00:06:00,230 --> 00:06:04,770
This is a guaranteed minimum bandwidth of 70 percent of the interface bandwidth.

60
00:06:05,150 --> 00:06:12,230
But it's a priority queue so voice traffic will be poor terrorized over all other traffic if voice traffic

61
00:06:12,230 --> 00:06:20,780
is only using 20 percent of the interface bandwidth other traffic can use whatever is not reserved for

62
00:06:20,780 --> 00:06:21,900
voice.

63
00:06:21,980 --> 00:06:28,480
So this is a maximum amount of bandwidth that voice can take and voice will be policed at that level.

64
00:06:29,360 --> 00:06:34,850
But in the worst case scenario other traffic will only get 25 percent of the bandwidth.

65
00:06:34,850 --> 00:06:43,740
This policy is applied on the serial to slash a zero interface outbound so any traffic leaving the Rada

66
00:06:44,130 --> 00:06:46,110
will be prior terrorised or cute.

67
00:06:46,110 --> 00:06:47,610
Based on this configuration.
