1
00:00:06,630 --> 00:00:09,070
- Now let's talk about edge locations.

2
00:00:09,070 --> 00:00:11,583
So edge locations within AWS

3
00:00:13,190 --> 00:00:17,510
power several different services,
namely Amazon Cloud Front,

4
00:00:17,510 --> 00:00:19,673
which gives us a content delivery network,

5
00:00:20,710 --> 00:00:24,740
Route 53, which gives
us a global DNS service,

6
00:00:24,740 --> 00:00:29,740
and API Gateway, which gives
us serverless REST APIs.

7
00:00:30,160 --> 00:00:34,260
And so with edge locations,
you can see here on the map,

8
00:00:34,260 --> 00:00:36,570
we have a number of
different edge locations,

9
00:00:36,570 --> 00:00:39,020
not all of them are actually drawn,

10
00:00:39,020 --> 00:00:41,720
but we do have, currently,
at the time of this video,

11
00:00:41,720 --> 00:00:45,230
more than 100 edge locations worldwide.

12
00:00:45,230 --> 00:00:50,230
Many of them actually
exist outside of a region.

13
00:00:50,260 --> 00:00:52,233
They could be just like a colocation,

14
00:00:53,388 --> 00:00:56,240
or a small data center that AWS has

15
00:00:56,240 --> 00:00:59,700
and runs in that area
specifically for an edge location.

16
00:00:59,700 --> 00:01:03,160
Now, we don't necessarily
have direct access

17
00:01:03,160 --> 00:01:06,480
to the edge location itself,
they're more passive.

18
00:01:06,480 --> 00:01:09,320
So if we were to use
Cloud Front, for example,

19
00:01:09,320 --> 00:01:12,170
and leverage the content delivery network,

20
00:01:12,170 --> 00:01:15,520
then we could have our application run in,

21
00:01:15,520 --> 00:01:18,670
say, U.S. East in Virginia.

22
00:01:18,670 --> 00:01:23,040
Instead of making our European
audience cross the Atlantic

23
00:01:23,040 --> 00:01:26,810
for every single request, we can use DNS

24
00:01:26,810 --> 00:01:30,600
to forward our European audience

25
00:01:30,600 --> 00:01:33,340
to an edge location in Europe,

26
00:01:33,340 --> 00:01:36,400
so that they can download content

27
00:01:36,400 --> 00:01:40,780
that is cached on a
server closest to them.

28
00:01:40,780 --> 00:01:42,990
So the first person who asks for a piece

29
00:01:42,990 --> 00:01:47,770
of content would have the
latency of crossing the Atlantic,

30
00:01:47,770 --> 00:01:50,750
but then everyone else,
perhaps the next million people

31
00:01:50,750 --> 00:01:53,390
who ask for that same URL,

32
00:01:53,390 --> 00:01:57,960
would benefit from having
it cached closest to them.

33
00:01:57,960 --> 00:02:01,820
And then of course, the
same thing works with DNS.

34
00:02:01,820 --> 00:02:06,030
When we need to fulfill a DNS query,

35
00:02:06,030 --> 00:02:08,550
we want to fulfill that
as quickly as possible

36
00:02:08,550 --> 00:02:09,880
with the lowest latency.

37
00:02:09,880 --> 00:02:12,920
And so, by having DNS servers

38
00:02:13,860 --> 00:02:16,460
running in edge locations
all over the world,

39
00:02:16,460 --> 00:02:20,070
this allows us to respond to DNS queries

40
00:02:20,070 --> 00:02:21,990
with the lowest latency as well.

41
00:02:21,990 --> 00:02:26,480
And all of these things,
regions, availability zones,

42
00:02:26,480 --> 00:02:29,940
and edge locations, we
will be referring back to

43
00:02:29,940 --> 00:02:31,513
as the course progresses.