AWS Solution Architect Interview Questions and Answers - Part 2 — Transcript

Full transcript

1. 0:00AWS
2. 0:04architectures how would you connect two
3. 0:08vpcs vpcs can be connected using VPC
4. 0:11peering it is a way to connect two vpcs
5. 0:14so that they can communicate privately
6. 0:16using their internal IP addresses
7. 0:18remember that it doesn't work like a
8. 0:21network Hub it's more like creating
9. 0:23point to point connections it cannot be
10. 0:26used to create a transitive connection
11. 0:28between other vpcs
12. 0:31how would you connect multiple
13. 0:34vpcs this can be done using a Transit
14. 0:36Gateway which acts like a central Hub
15. 0:39that simplifies connecting multiple vpcs
16. 0:42and on promise networks it centralizes
17. 0:45connections reducing complex
18. 0:46configurations compared to VPC peerings
19. 0:49point-to-point
20. 0:51approach how would you connect an on
21. 0:53promise network with AWS Cloud AWS
22. 0:57Direct Connect lets you establish a DED
23. 0:59dedicated high bandwidth network
24. 1:01connection between your on premise
25. 1:03Network and AWS bypassing the public
26. 1:07Internet site to site VPN creates a
27. 1:10secure encrypted tunnel using IPC to
28. 1:13connect your on premise Network to your
29. 1:15VPC in AWS it allows controlled private
30. 1:18communication between your resources
31. 1:20over a secure
32. 1:22tunnel how can you trace requests in
33. 1:25your microservices application deployed
34. 1:27on
35. 1:28AWS using AWS x-ray we can track
36. 1:32requests as they flow through the
37. 1:33application across Lambda functions and
38. 1:35other microservices this helps identify
39. 1:38bottlenecks visualize request flow and
40. 1:41debug
41. 1:42issues how would you migrate database
42. 1:45from Oracle to
43. 1:48RDS AWS DMS or database migration
44. 1:51service facilitates migrating data from
45. 1:54one database to another be it within AWS
46. 1:58or between your on premise environment
47. 2:00and
48. 2:02AWS how can you export a database to S3
49. 2:06both initially and then for incremental
50. 2:09changes AWS DMS handles both initial and
51. 2:13ongoing exports to S3 it can initially
52. 2:16do a full data export from database to
53. 2:19S3 and then continuously replicate
54. 2:21changes from database to S3 keeping it
55. 2:25synchronized how can you secure your web
56. 2:28application on AWS against common web
57. 2:32exploits using AWS wav or web
58. 2:35application firewall you can create a
59. 2:37security layer for your web applications
60. 2:40protecting them from malicious attacks
61. 2:42like SPL injection and cross-site
62. 2:44scripting WAP lets you define rules to
63. 2:47block suspicious traffic allow
64. 2:49legitimate requests and monitor web
65. 2:52traffic for security
66. 2:54threats how can you protect your web
67. 2:57application on AWS from DS attack
68. 3:01AWS Shield a managed dos Protection
69. 3:03Service safeguards your applications on
70. 3:06AWS from distributed denial of service
71. 3:08attacks that aim to overwhelm them it
72. 3:11offers two tires a standard free tire
73. 3:14for basic protection and advanced paid
74. 3:16tire for enhanced dos mitigation and
75. 3:19response
76. 3:22capabilities how can you deliver static
77. 3:24content faster to end users around the
78. 3:28world using using cloudfront a Content
79. 3:31delivery network service you can speed
80. 3:33up delivery of your website and app
81. 3:35content like images and videos by
82. 3:38caching them in a Global Network of edge
83. 3:40locations users then access the content
84. 3:43from the closest Edge location reducing
85. 3:45latency and improving load
86. 3:48times how can you enable internet access
87. 3:51for ec2 instances in a private Subnet in
88. 3:54a
89. 3:55VPC we can use NAD Gateway for this
90. 3:58which allows instant in a private subnet
91. 4:01to connect out to the internet but
92. 4:03prevents inbound connections from the
93. 4:05internet reaching those instances it
94. 4:08acts like a translator converting
95. 4:10private IP addresses to a public IP for
96. 4:13outbound
97. 4:15traffic how can you generate a trigger
98. 4:18each time a tracked vehicle enters a
99. 4:20specific geographic
100. 4:22area we could use Amazon location
101. 4:24service which lets us add location
102. 4:26functionality to Applications it has
103. 4:28features like geens ing which Define
104. 4:30virtual boundaries or geofences so any
105. 4:33track vehicles that enters or leaves
106. 4:36these geofences triggers geofence events
107. 4:39which we can use to send alerts like
108. 4:41notify the delivery Personnel when
109. 4:43nearing a stop or initiate other
110. 4:46actions a large company has deployed iot
111. 4:50devices around the world that captures
112. 4:52realtime sensor data like temperature
113. 4:55humidity Etc how can they efficiently
114. 4:58ingest and St store this high volume
115. 5:00data on cloud which must be available
116. 5:03for time based
117. 5:05analysis the iot devices can send sensor
118. 5:08data to AWS iot core service or mqt
119. 5:12which is then filtered and routed using
120. 5:14iot core rules and stored in a Time
121. 5:16stream database time stream is a fast
122. 5:19scalable time series database which can
123. 5:21efficiently store large streams of data
124. 5:24points like sensor readings or
125. 5:26application metrics with timestamps this
126. 5:29facilit
127. 5:30data analysis of recent and historical
128. 5:32data over
129. 5:34time an e-commerce company hosts a web
130. 5:37application in a US region how can they
131. 5:40ensure that customers in Europe and Asia
132. 5:43are able to access this web application
133. 5:45without experiencing any
134. 5:48lag here we could use AWS Global
135. 5:51accelerator service which comes with
136. 5:53static anycast IP addresses you
137. 5:55configure your DNS service to point to
138. 5:57these IP addresses the global
139. 6:00accelerator routes user traffic over the
140. 6:03fast AWS Network to your application
141. 6:05endpoint it's ideal for applications
142. 6:08accessed by a geographically distributed
143. 6:11audience normally when a user accesses
144. 6:14an application on AWS it goes over
145. 6:16public internet until it reaches the
146. 6:19destination Regional endpoint this can
147. 6:21be slow depending on network hops
148. 6:23available bandwidth and so on however if
149. 6:26a end users's request can travel through
150. 6:29fast AWS Global Network it can reach the
151. 6:31application's Endo in AWS faster thereby
152. 6:36delivering better performance AWS Global
153. 6:38accelerator allows you to do exactly
154. 6:41that therefore it is a service to
155. 6:43improve availability and performance of
156. 6:45Internet
157. 6:47applications how can you Aggregate and
158. 6:50view security status of your AWS
159. 6:52resources and Trigger actions based on
160. 6:56that we would use aw security Hub
161. 6:58service here it provides a central
162. 7:01dashboard for security findings it
163. 7:03collects findings from built-in AWS
164. 7:05Security Services like guard Duty
165. 7:08inspector Etc partner security tools and
166. 7:11your own custom Integrations the
167. 7:13findings can be sent to Cloud watch
168. 7:15events where event rules could be
169. 7:18configured to trigger
170. 7:22actions how can you establish
171. 7:24asynchronous bidirectional messaging
172. 7:26connection between clients and server
173. 7:30websockets with API Gateway provides
174. 7:32realtime two-way communication between
175. 7:34web clients and backend services on AWS
176. 7:37this establishes a full duplex
177. 7:39persistent connection allowing
178. 7:41bidirectional data flow unlike
179. 7:43traditional HTTP requests API Gateway
180. 7:45routes incoming websockets messages to
181. 7:48appropriate backend services like Lambda
182. 7:50functions this setup is ideal for
183. 7:53scenarios requiring live updates like
184. 7:55chat apps or collaborative
185. 7:58editing what makes Amazon Kinesis highly
186. 8:01scalable and fast Kinesis has a
187. 8:04distributed architecture which allows it
188. 8:07to distribute data processing workload
189. 8:09across multiple resources the data
190. 8:11stream is spread across multiple charts
191. 8:13and each chart can be processed
192. 8:15independently in parall this capability
193. 8:18enables kineses to handle High data
194. 8:20injection rates and process large
195. 8:22volumes of data in real time let's
196. 8:24examine this more closely partition keys
197. 8:26are used to assign data records to
198. 8:28shards they therefore with an
199. 8:30appropriate partition key strategy data
200. 8:32can be distributed smartly across shards
201. 8:34Kinesis also provides ordering of data
202. 8:37records Within A Shard data capacity of
203. 8:40your stream is a function of number of
204. 8:45shards Kinesis has the ability to scale
205. 8:47the number of shards dynamically each
206. 8:49Shard can support a certain level of
207. 8:51throughput and by increasing the number
208. 8:53of shards Kinesis can scale up its
209. 8:55processing capacity and therefore handle
210. 8:58higher data injection rates this elastic
211. 9:01scaling ensures that kineses can
212. 9:03maintain high speed processing with
213. 9:05increased
214. 9:07workloads what is the typical structure
215. 9:10of a cicd pipeline in
216. 9:13AWS a typical cicd pipeline uses a
217. 9:17combination of services to streamline
218. 9:19the development and release process
219. 9:21developer commits code to code commit
220. 9:23code build automatically triggers a
221. 9:24build process compiles your code runs
222. 9:27tests and produces deployment ready
223. 9:29packages code deployer takes over
224. 9:32deploying the built packages to your
225. 9:34chosen Target for running your
226. 9:37application how would you implement a
227. 9:40microservices architecture in
228. 9:43AWS here's an example the account
229. 9:45service inventory service and Order
230. 9:47service are implemented as microservices
231. 9:50via Lambda functions each service has
232. 9:53its own data storage API Gateway routes
233. 9:56incoming requests to these services
234. 10:01and here's a variation of the same where
235. 10:03microservices are deployed in containers
236. 10:06managed by AWS
237. 10:10fargate how can you handle
238. 10:12multi-protocol traffic using load
239. 10:15balances in
240. 10:17AWS this is an example of handling
241. 10:19multiprotocol traffic using network and
242. 10:22application load balancers the network
243. 10:24load balancer has two listeners TCP on
244. 10:26Port 80 and UDP on Port 53
245. 10:30the TCP listener is configured to
246. 10:32forward traffic to an application load
247. 10:33balancer Target group while the UDP
248. 10:36listener is configured to forward UDP
249. 10:38traffic to another Target Group which
250. 10:40has instances that will accept UDP
251. 10:42traffic this configuration works well
252. 10:44for applications that use multi-protocol
253. 10:47connections such as media services using
254. 10:50HTTP for signaling and RTP for streaming
255. 10:55content an e-commerce company receives
256. 10:58massive amount ounts of raw data in the
257. 11:01form of CSV files from its warehouses
258. 11:04around the world via FTP every day the
259. 11:07company wants to provide its business
260. 11:09analyst an easy way to analyze this data
261. 11:11using ad hoc SQL queries how would you
262. 11:14design
263. 11:16this let's look at the architecture for
264. 11:18this use case we have a data lake with
265. 11:21two buckets raw and processed incoming
266. 11:23raw CSV files are saved in S3 bucket raw
267. 11:27while cleaned files are stored in
268. 11:29processed bucket a glue crawler crawls
269. 11:31the raw and process data buckets to
270. 11:33create metadata tables in glue data
271. 11:35catalog a glue job cleans and transforms
272. 11:38the raw data to a storage and query
273. 11:40efficient format like Park and saves it
274. 11:43in the processed bucket analysts can now
275. 11:46run ad hoc queries using ethena against
276. 11:49the data in the processed bucket