1
00:00:00,000 --> 00:00:01,110
‫So let's take some time

2
00:00:01,110 --> 00:00:03,901
‫to understand how caching works in CloudFront.

3
00:00:03,901 --> 00:00:05,135
‫So the cache lives

4
00:00:05,135 --> 00:00:07,387
‫at each CloudFront edge location.

5
00:00:07,387 --> 00:00:08,876
‫So you'll have as many caches

6
00:00:08,876 --> 00:00:10,374
‫as edge locations.

7
00:00:10,374 --> 00:00:12,504
‫And each object in the cache

8
00:00:12,504 --> 00:00:13,891
‫is going to be identified

9
00:00:13,891 --> 00:00:15,115
‫by a Cache Key.

10
00:00:15,115 --> 00:00:16,740
‫We'll see what the Cache Key is

11
00:00:16,740 --> 00:00:17,573
‫in the next slide.

12
00:00:17,573 --> 00:00:19,651
‫But the idea is that a request made

13
00:00:19,651 --> 00:00:22,150
‫through CloudFront edge location.

14
00:00:22,150 --> 00:00:23,501
‫The edge location is first going

15
00:00:23,501 --> 00:00:25,393
‫to check whether or not the object

16
00:00:25,393 --> 00:00:26,505
‫has been cached.

17
00:00:26,505 --> 00:00:27,930
‫And if it's in the cache,

18
00:00:27,930 --> 00:00:28,888
‫has it expired or not based

19
00:00:28,888 --> 00:00:30,513
‫on the time to live.

20
00:00:30,513 --> 00:00:31,737
‫And then if not,

21
00:00:31,737 --> 00:00:33,127
‫if it's not in the cache,

22
00:00:33,127 --> 00:00:34,493
‫then the request is forwarded

23
00:00:34,493 --> 00:00:36,524
‫and we'll see how it's forwarded

24
00:00:36,524 --> 00:00:37,866
‫to your origin.

25
00:00:37,866 --> 00:00:40,016
‫And then the response from the origin

26
00:00:40,016 --> 00:00:41,730
‫is cached into your edge location.

27
00:00:41,730 --> 00:00:45,259
‫So that future request return a cache result.

28
00:00:45,259 --> 00:00:47,874
‫So you want to maximize the Cache Hit ratio

29
00:00:47,874 --> 00:00:49,500
‫by minimizing request to the origin.

30
00:00:49,500 --> 00:00:51,142
‫That means that you want to cache

31
00:00:51,142 --> 00:00:52,244
‫as much stuff as possible

32
00:00:52,244 --> 00:00:54,250
‫in your edge locations.

33
00:00:54,250 --> 00:00:55,391
‫We'll see as well that it's possible

34
00:00:55,391 --> 00:00:57,002
‫for you to not wait

35
00:00:57,002 --> 00:00:59,749
‫until the item expires based on a TTL.

36
00:00:59,749 --> 00:01:00,745
‫If you wanna remove it from the cache,

37
00:01:00,745 --> 00:01:03,255
‫you can create an invalidation.

38
00:01:03,255 --> 00:01:04,761
‫So what is a CloudFront Cache Key

39
00:01:04,761 --> 00:01:06,235
‫and what do we cache on?

40
00:01:06,235 --> 00:01:07,872
‫Well, a Cache Key is a unique identifier

41
00:01:07,872 --> 00:01:09,783
‫for each object in the cache.

42
00:01:09,783 --> 00:01:11,986
‫And by default, if you don't do anything,

43
00:01:11,986 --> 00:01:13,394
‫it's made up of the host name

44
00:01:13,394 --> 00:01:17,130
‫and the resource portion of URL.

45
00:01:17,130 --> 00:01:17,963
‫So in this example,

46
00:01:17,963 --> 00:01:20,010
‫there is mywebsite.com.

47
00:01:20,010 --> 00:01:20,873
‫So that's the host name.

48
00:01:20,873 --> 00:01:21,706
‫And then there is the

49
00:01:21,706 --> 00:01:26,670
‫GET/content/stories/example-story.html.

50
00:01:26,670 --> 00:01:29,011
‫That is the resource portion of the URL.

51
00:01:29,011 --> 00:01:29,844
‫That means that anyone

52
00:01:29,844 --> 00:01:31,739
‫that makes a similar request

53
00:01:31,739 --> 00:01:33,500
‫in case there is a cache miss,

54
00:01:33,500 --> 00:01:34,400
‫well, first, we're going

55
00:01:34,400 --> 00:01:36,631
‫to get the object from the origin

56
00:01:36,631 --> 00:01:37,464
‫but then the object

57
00:01:37,464 --> 00:01:38,617
‫is going to be cached

58
00:01:38,617 --> 00:01:40,503
‫based on these two keys right here,

59
00:01:40,503 --> 00:01:42,255
‫the host name

60
00:01:42,255 --> 00:01:44,745
‫and the resource portion of the URL.

61
00:01:44,745 --> 00:01:47,280
‫And then a request that looks similar

62
00:01:47,280 --> 00:01:48,113
‫with the same host name

63
00:01:48,113 --> 00:01:49,742
‫and the same resource portion

64
00:01:49,742 --> 00:01:50,874
‫will hit the cache

65
00:01:50,874 --> 00:01:52,640
‫and will get a cache hit.

66
00:01:52,640 --> 00:01:54,001
‫But sometimes you want your Cache Key

67
00:01:54,001 --> 00:01:55,133
‫to be a bit more complicated

68
00:01:55,133 --> 00:01:56,648
‫because sometimes, well,

69
00:01:56,648 --> 00:01:59,280
‫you have content that will vary

70
00:01:59,280 --> 00:02:00,113
‫based on the user

71
00:02:00,113 --> 00:02:02,940
‫or the device or the language

72
00:02:02,940 --> 00:02:03,875
‫or the location the user

73
00:02:03,875 --> 00:02:05,407
‫is from and so on.

74
00:02:05,407 --> 00:02:06,240
‫And so we want to do

75
00:02:06,240 --> 00:02:08,264
‫is to enhance the Cache Key

76
00:02:08,264 --> 00:02:10,004
‫and add more information.

77
00:02:10,004 --> 00:02:12,492
‫So we wanna add maybe HTTP headers

78
00:02:12,492 --> 00:02:14,997
‫or cookies or query strings in it.

79
00:02:14,997 --> 00:02:16,382
‫And to define how to create

80
00:02:16,382 --> 00:02:18,240
‫that Cache Key, we have to define

81
00:02:18,240 --> 00:02:22,020
‫what's called a CloudFront cache policy.

82
00:02:22,020 --> 00:02:23,752
‫So this cache policy can be caching

83
00:02:23,752 --> 00:02:26,160
‫based on the HTTP headers.

84
00:02:26,160 --> 00:02:27,245
‫So you can select none of them

85
00:02:27,245 --> 00:02:28,486
‫or a whitelist to say

86
00:02:28,486 --> 00:02:30,125
‫which one you want to include.

87
00:02:30,125 --> 00:02:31,650
‫You have the cookies.

88
00:02:31,650 --> 00:02:32,989
‫So you can select none of them,

89
00:02:32,989 --> 00:02:34,480
‫a whitelist or list you wanna include

90
00:02:34,480 --> 00:02:37,007
‫or all or all accept.

91
00:02:37,007 --> 00:02:38,249
‫And we have the query strings.

92
00:02:38,249 --> 00:02:39,880
‫Again, similarly, do you want

93
00:02:39,880 --> 00:02:41,923
‫to include none of them,

94
00:02:41,923 --> 00:02:44,393
‫a whitelist, all except, or all?

95
00:02:44,393 --> 00:02:45,626
‫And this is basically going

96
00:02:45,626 --> 00:02:47,236
‫to configure how your Cache Key

97
00:02:47,236 --> 00:02:48,856
‫is going to be created.

98
00:02:48,856 --> 00:02:49,689
‫In the cache policy,

99
00:02:49,689 --> 00:02:51,128
‫you can also control the TTL.

100
00:02:51,128 --> 00:02:52,755
‫So from zero seconds up

101
00:02:52,755 --> 00:02:54,501
‫to one year in the cache.

102
00:02:54,501 --> 00:02:55,875
‫And then you can also control

103
00:02:55,875 --> 00:02:57,498
‫that setting using specific headers

104
00:02:57,498 --> 00:02:59,520
‫called the cache control header

105
00:02:59,520 --> 00:03:00,738
‫or the expires header.

106
00:03:00,738 --> 00:03:03,398
‫So you can create your own cache policies

107
00:03:03,398 --> 00:03:04,624
‫or you can use predefined

108
00:03:04,624 --> 00:03:06,887
‫managed policies by AWS.

109
00:03:06,887 --> 00:03:08,010
‫And the very important thing

110
00:03:08,010 --> 00:03:09,510
‫I want you to remember

111
00:03:09,510 --> 00:03:11,249
‫is that all the HTTP headers,

112
00:03:11,249 --> 00:03:12,493
‫cookies and query strings

113
00:03:12,493 --> 00:03:14,631
‫that you include in the Cache Key

114
00:03:14,631 --> 00:03:16,864
‫will be automatically included

115
00:03:16,864 --> 00:03:20,430
‫and forwarded to your origin request.

116
00:03:20,430 --> 00:03:21,263
‫So in the case of HTTP header

117
00:03:21,263 --> 00:03:22,638
‫to give you an example,

118
00:03:22,638 --> 00:03:23,988
‫let's say we have this request

119
00:03:23,988 --> 00:03:25,988
‫and we have a language fr-fr,

120
00:03:25,988 --> 00:03:27,992
‫meaning that we request the blog

121
00:03:27,992 --> 00:03:28,975
‫in the French language.

122
00:03:28,975 --> 00:03:29,881
‫How does that work?

123
00:03:29,881 --> 00:03:31,235
‫Well, if we define a none

124
00:03:31,235 --> 00:03:32,504
‫HTTP header cache policy,

125
00:03:32,504 --> 00:03:34,251
‫then none of the headers

126
00:03:34,251 --> 00:03:36,146
‫are going to be cached

127
00:03:36,146 --> 00:03:38,142
‫and the headers will not be forwarded

128
00:03:38,142 --> 00:03:40,133
‫unless and I will show you what happens

129
00:03:40,133 --> 00:03:41,749
‫in the rest of this lecture.

130
00:03:41,749 --> 00:03:42,997
‫So by default, headers

131
00:03:42,997 --> 00:03:44,639
‫will not be forwarded.

132
00:03:44,639 --> 00:03:45,514
‫But this gives you

133
00:03:45,514 --> 00:03:46,494
‫the best caching performance

134
00:03:46,494 --> 00:03:48,749
‫because you don't have any headers in it.

135
00:03:48,749 --> 00:03:50,484
‫If you want to whitelist specific headers

136
00:03:50,484 --> 00:03:51,912
‫and that may mean necessary

137
00:03:51,912 --> 00:03:52,745
‫because while you want

138
00:03:52,745 --> 00:03:55,495
‫to have the language as a Cache Key,

139
00:03:55,495 --> 00:03:57,248
‫then you specify which headers you want

140
00:03:57,248 --> 00:03:58,252
‫to include in the Cache Key,

141
00:03:58,252 --> 00:04:00,113
‫for example, the language header.

142
00:04:00,113 --> 00:04:02,122
‫And then this header, the language

143
00:04:02,122 --> 00:04:04,883
‫will also be forwarded to your origin

144
00:04:04,883 --> 00:04:06,090
‫so that the origin

145
00:04:06,090 --> 00:04:07,388
‫can actually respond to the request

146
00:04:07,388 --> 00:04:08,388
‫and give you the blog

147
00:04:08,388 --> 00:04:10,765
‫in the correct language.

148
00:04:10,765 --> 00:04:12,244
‫So this is a very similar mechanism

149
00:04:12,244 --> 00:04:13,680
‫for query strings.

150
00:04:13,680 --> 00:04:15,240
‫So query strings are what happens

151
00:04:15,240 --> 00:04:16,985
‫in the URL after a question mark.

152
00:04:16,985 --> 00:04:18,999
‫So for example, border equals red

153
00:04:18,999 --> 00:04:20,259
‫and size equals large.

154
00:04:20,259 --> 00:04:21,994
‫So here, we want a cat image.

155
00:04:21,994 --> 00:04:23,504
‫But apparently, it's going to be customized

156
00:04:23,504 --> 00:04:25,250
‫a little bit by the origin.

157
00:04:25,250 --> 00:04:26,642
‫So again, if you have none,

158
00:04:26,642 --> 00:04:28,008
‫then none of the query strings

159
00:04:28,008 --> 00:04:29,543
‫are going to be used for the Cache Key

160
00:04:29,543 --> 00:04:30,376
‫and they're not going

161
00:04:30,376 --> 00:04:32,751
‫to be forwarded to your origin.

162
00:04:32,751 --> 00:04:33,750
‫Whitelist, you specify

163
00:04:33,750 --> 00:04:36,739
‫which query strings you want included.

164
00:04:36,739 --> 00:04:38,255
‫Include all-except, you specify

165
00:04:38,255 --> 00:04:39,754
‫which ones you don't want

166
00:04:39,754 --> 00:04:41,499
‫but the rest passes

167
00:04:41,499 --> 00:04:44,125
‫and all is including all the query strings

168
00:04:44,125 --> 00:04:45,383
‫in the Cache Keys

169
00:04:45,383 --> 00:04:46,680
‫and all the query strings

170
00:04:46,680 --> 00:04:47,513
‫are going to be forwarded.

171
00:04:47,513 --> 00:04:48,747
‫But of course, if you have many,

172
00:04:48,747 --> 00:04:51,540
‫this gives you the worst caching performance.

173
00:04:51,540 --> 00:04:52,373
‫So we've seen that

174
00:04:52,373 --> 00:04:54,392
‫when we define a cache policy,

175
00:04:54,392 --> 00:04:56,249
‫we can have query strings,

176
00:04:56,249 --> 00:04:58,644
‫cookies and headers as whitelist.

177
00:04:58,644 --> 00:05:00,495
‫For example, to choose (indistinct) want

178
00:05:00,495 --> 00:05:03,240
‫and they will be forwarded to the origin.

179
00:05:03,240 --> 00:05:04,073
‫But what if you want

180
00:05:04,073 --> 00:05:05,643
‫to include some stuff

181
00:05:05,643 --> 00:05:07,886
‫in the origin request, okay?

182
00:05:07,886 --> 00:05:09,391
‫But you don't want to include them

183
00:05:09,391 --> 00:05:10,761
‫in the Cache Key.

184
00:05:10,761 --> 00:05:11,747
‫In that case, you can define

185
00:05:11,747 --> 00:05:14,512
‫what's called an origin request policy.

186
00:05:14,512 --> 00:05:15,753
‫So the idea is that

187
00:05:15,753 --> 00:05:18,140
‫you can include extra HTTP headers

188
00:05:18,140 --> 00:05:20,398
‫or cookies or query strings

189
00:05:20,398 --> 00:05:22,495
‫but they will be forwarded to the origin

190
00:05:22,495 --> 00:05:23,621
‫but they're not going to be used

191
00:05:23,621 --> 00:05:25,015
‫in the Cache Key.

192
00:05:25,015 --> 00:05:26,372
‫So you can also,

193
00:05:26,372 --> 00:05:28,492
‫as part of the origin request policy,

194
00:05:28,492 --> 00:05:30,359
‫add custom HTTP headers

195
00:05:30,359 --> 00:05:32,749
‫or CloudFront HTTP headers to the origin

196
00:05:32,749 --> 00:05:34,385
‫even though they were not present

197
00:05:34,385 --> 00:05:35,894
‫in the viewer request.

198
00:05:35,894 --> 00:05:36,727
‫For example, if you wanted

199
00:05:36,727 --> 00:05:39,659
‫to pass an API key or a secret header.

200
00:05:39,659 --> 00:05:40,879
‫So you can create your own policy

201
00:05:40,879 --> 00:05:43,860
‫or you can use predefined managed policies.

202
00:05:43,860 --> 00:05:44,730
‫And so at this point,

203
00:05:44,730 --> 00:05:45,563
‫you may be confused like,

204
00:05:45,563 --> 00:05:46,396
‫what is really the difference

205
00:05:46,396 --> 00:05:47,652
‫between Cache Keys

206
00:05:47,652 --> 00:05:49,510
‫and origin request policy?

207
00:05:49,510 --> 00:05:50,501
‫Well, let me try to summarize

208
00:05:50,501 --> 00:05:52,009
‫as best as I can.

209
00:05:52,009 --> 00:05:54,381
‫So the request will come

210
00:05:54,381 --> 00:05:56,006
‫with some query strings,

211
00:05:56,006 --> 00:05:58,267
‫some cookies, some headers,

212
00:05:58,267 --> 00:06:00,245
‫and then we will cache based

213
00:06:00,245 --> 00:06:02,006
‫on the cache policy.

214
00:06:02,006 --> 00:06:03,371
‫For example, we want to cache here

215
00:06:03,371 --> 00:06:05,644
‫on the host name, the resource,

216
00:06:05,644 --> 00:06:08,392
‫and a header called authorization.

217
00:06:08,392 --> 00:06:10,732
‫But then, your origin may need more

218
00:06:10,732 --> 00:06:12,863
‫than these three things to actually work

219
00:06:12,863 --> 00:06:15,491
‫and to actually serve properly the request.

220
00:06:15,491 --> 00:06:18,510
‫So you may want to add the user agents,

221
00:06:18,510 --> 00:06:21,993
‫the session ID and the ref query string

222
00:06:21,993 --> 00:06:24,662
‫as part of your request to the origin.

223
00:06:24,662 --> 00:06:25,738
‫So in this case,

224
00:06:25,738 --> 00:06:26,641
‫the request to the origin

225
00:06:26,641 --> 00:06:28,395
‫is going to be enhanced

226
00:06:28,395 --> 00:06:30,650
‫but then the caching will not happen

227
00:06:30,650 --> 00:06:32,247
‫based on what we forwarded

228
00:06:32,247 --> 00:06:34,050
‫to the origin request policy.

229
00:06:34,050 --> 00:06:34,889
‫It's only going to be based

230
00:06:34,889 --> 00:06:36,244
‫on the cache policy.

231
00:06:36,244 --> 00:06:37,258
‫So I hope you see the synergy

232
00:06:37,258 --> 00:06:38,491
‫between these two

233
00:06:38,491 --> 00:06:39,751
‫and that should be enough

234
00:06:39,751 --> 00:06:41,758
‫to answer questions at the exam.

235
00:06:41,758 --> 00:06:43,020
‫All right, that's it.

236
00:06:43,020 --> 00:06:44,160
‫I hope you liked it

237
00:06:44,160 --> 00:06:46,203
‫and I will see you in the next lecture.