1
00:00:00,000 --> 00:00:00,060


2
00:00:00,060 --> 00:00:00,726
[VIDEO PLAYBACK]

3
00:00:00,726 --> 00:00:03,110
- Dot com and DNS does that conversion.

4
00:00:03,110 --> 00:00:04,860
Even more dynamically
than the old school.

5
00:00:04,860 --> 00:00:05,050
[END PLAYBACK]

6
00:00:05,050 --> 00:00:06,060
DAVID MALAN: So these demos I love.

7
00:00:06,060 --> 00:00:08,976
But they are admittedly a little
hard to do sometimes technologically.

8
00:00:08,976 --> 00:00:12,090
Using traceroute, NSLookup,
Kang, or a few other command

9
00:00:12,090 --> 00:00:15,710
line tools on Linux, Mac OS, or Windows.

10
00:00:15,710 --> 00:00:19,230
Traceroute especially, I think it's
just been so eye opening for myself,

11
00:00:19,230 --> 00:00:21,120
back in the day and so
many students since.

12
00:00:21,120 --> 00:00:23,161
When you actually trace
the route between a place

13
00:00:23,161 --> 00:00:29,400
like Harvard and maybe MIT, Stanford,
or cnn.co.jp across the Pacific Ocean.

14
00:00:29,400 --> 00:00:32,670
Because you really see the
increased latency in the hops

15
00:00:32,670 --> 00:00:36,670
that are actually
traveling through an ocean.

16
00:00:36,670 --> 00:00:38,460
Or just down the road in this case.

17
00:00:38,460 --> 00:00:42,840
DOUG LLOYD: Yeah, it really is eye
opening to actually see those hops

18
00:00:42,840 --> 00:00:43,636
and take away.

19
00:00:43,636 --> 00:00:45,510
It's not just one machine
talking to another.

20
00:00:45,510 --> 00:00:48,135
It's got to go through a lot of
different things along the way.

21
00:00:48,135 --> 00:00:50,910
The unfortunate thing, with
an example like traceroute,

22
00:00:50,910 --> 00:00:53,450
is that it's often blocked.

23
00:00:53,450 --> 00:00:54,705
It gets shut down.

24
00:00:54,705 --> 00:00:57,330
DAVID MALAN: You either see a
whole bunch of stars or asterisks

25
00:00:57,330 --> 00:00:58,140
on the screen.

26
00:00:58,140 --> 00:01:00,690
Or it just doesn't give you
back correct information.

27
00:01:00,690 --> 00:01:04,260
In fact on CS50 IDE because of the
cloud infrastructure is running on,

28
00:01:04,260 --> 00:01:05,190
it's not enlightening.

29
00:01:05,190 --> 00:01:08,960
So I'm doing a little bit
of cooking show magic here.

30
00:01:08,960 --> 00:01:11,115
Where usually the
terminal window I'm using

31
00:01:11,115 --> 00:01:13,740
is actually connected to another
server where it's not blocked.

32
00:01:13,740 --> 00:01:16,022
So I can simulate what's going on.

33
00:01:16,022 --> 00:01:17,730
Just because it's so
visually compelling.

34
00:01:17,730 --> 00:01:21,120
But it's a lot harder for students
or teachers in their own schools

35
00:01:21,120 --> 00:01:23,890
to replicate that demo without
a bit of advance prep work.

36
00:01:23,890 --> 00:01:26,854
DOUG LLOYD: But with teachers
who maybe have Mac computers,

37
00:01:26,854 --> 00:01:29,020
it is something that you
can do in the Mac terminal.

38
00:01:29,020 --> 00:01:30,478
DAVID MALAN: It is and Windows too.

39
00:01:30,478 --> 00:01:32,820
It's just traceRT instead of route.

40
00:01:32,820 --> 00:01:35,300
DOUG LLOYD: So it's
definitely possible to do it.

41
00:01:35,300 --> 00:01:39,270
It would be cool if students could also
do it pretty easily from their IDEs.

42
00:01:39,270 --> 00:01:40,170
DAVID MALAN: I agree.

43
00:01:40,170 --> 00:01:42,510
This is a minor enhancement
I've done in recent years.

44
00:01:42,510 --> 00:01:45,060
Dash Q1 sends just one
query instead of three.

45
00:01:45,060 --> 00:01:47,430
Because it was an FAQ as
to why there are three

46
00:01:47,430 --> 00:01:49,140
numbers for every row on the screen.

47
00:01:49,140 --> 00:01:49,980
DOUG LLOYD: Over the times?

48
00:01:49,980 --> 00:01:53,220
DAVID MALAN: Yeah and it's just meant
to give you a visual average of things.

49
00:01:53,220 --> 00:01:55,650
Or just a few numbers you can
lump together and average.

50
00:01:55,650 --> 00:01:56,910
But it was just a distraction.

51
00:01:56,910 --> 00:01:58,680
And it also frankly,
just makes the text wrap,

52
00:01:58,680 --> 00:02:00,596
especially when the fonts
are a little bigger.

53
00:02:00,596 --> 00:02:02,790
So I tried to simplify it
a little bit like this.

54
00:02:02,790 --> 00:02:05,240
But it's great fun because
this is Harvard specific now.

55
00:02:05,240 --> 00:02:07,440
But having students
at Harvard, at least,

56
00:02:07,440 --> 00:02:12,180
see familiar phrases like SC for
Science Center or GW for gateway.

57
00:02:12,180 --> 00:02:13,320
Which we introduce here.

58
00:02:13,320 --> 00:02:14,970
Or NOCS for Northern crossroads.

59
00:02:14,970 --> 00:02:19,276
Or better yet, is when
you start to see things.

60
00:02:19,276 --> 00:02:21,150
In this case, of course,
we're not seeing it.

61
00:02:21,150 --> 00:02:24,750
Sometimes you see airport codes
in the name of the domains.

62
00:02:24,750 --> 00:02:25,560
DOUG LLOYD: Yep.

63
00:02:25,560 --> 00:02:28,129
DAVID MALAN: Which hint at
what city-state they're in.

64
00:02:28,129 --> 00:02:30,420
DOUG LLOYD: In fact, we have
had an example in the past

65
00:02:30,420 --> 00:02:32,860
where we have spoken to Yale.

66
00:02:32,860 --> 00:02:34,130
We've tapped into Yale.

67
00:02:34,130 --> 00:02:37,410
Which I think somehow gets
routed through JFK airport.

68
00:02:37,410 --> 00:02:38,430
Even though it's past--

69
00:02:38,430 --> 00:02:40,138
DAVID MALAN: It might
not be the airport.

70
00:02:40,138 --> 00:02:42,390
DOUG LLOYD: The airport for that area.

71
00:02:42,390 --> 00:02:45,012
That overshoots New Haven a
little bit from Cambridge.

72
00:02:45,012 --> 00:02:47,220
DAVID MALAN: But there's
good opportunities too here.

73
00:02:47,220 --> 00:02:49,553
They're pretty non-deterministic
every time we run this.

74
00:02:49,553 --> 00:02:53,490
Whereby that last hop took far
fewer milliseconds than the second

75
00:02:53,490 --> 00:02:54,420
to last hop.

76
00:02:54,420 --> 00:03:00,450
Even though each of these is supposed
to be from point A to point I plus one,

77
00:03:00,450 --> 00:03:01,070
so to speak.

78
00:03:01,070 --> 00:03:02,370
They're not cumulative.

79
00:03:02,370 --> 00:03:05,460
But that just hints at the
variability in routers, buffers,

80
00:03:05,460 --> 00:03:09,060
in any congestion there might be,
or need to retransmit or so forth.

81
00:03:09,060 --> 00:03:12,676
So it captures some
real world realities.

82
00:03:12,676 --> 00:03:15,300
DOUG LLOYD: Particularly dramatic
when you're crossing an ocean

83
00:03:15,300 --> 00:03:17,400
and you only have a couple
of cables that you can use.

84
00:03:17,400 --> 00:03:19,320
And all the traffic is
trying to get through them.

85
00:03:19,320 --> 00:03:20,653
DAVID MALAN: That's my favorite.

86
00:03:20,653 --> 00:03:23,180
When you actually see a jump
of 100 more milliseconds.

87
00:03:23,180 --> 00:03:25,222
Suggestive that there's
something big in the way.

88
00:03:25,222 --> 00:03:26,929
DOUG LLOYD: But it's
still pretty amazing

89
00:03:26,929 --> 00:03:29,280
that you can get to the other
side of the world so fast.

90
00:03:29,280 --> 00:03:32,580
DAVID MALAN: I tell people that we
have trans Atlantic, trans Pacific

91
00:03:32,580 --> 00:03:33,800
cables underneath the ocean.

92
00:03:33,800 --> 00:03:35,491
It still blows my mind.

93
00:03:35,491 --> 00:03:35,991