AWS · Advanced Networking Specialty

网络课程笔记

Lecture 176 · 11 张幻灯片

幻灯片、双语总结及 Transcript · 图片已内嵌,可离线阅读。

原文默认收起
lecture-176.md

Lecture 176: Hands On: AWS Global Accelerator

  • 课程:[NEW] AWS Certified Advanced Networking Specialty 2026
  • 视频时长:08:49
  • 来源:Udemy Lecture 28556213
  • Slide 数量:11
  • 画面切换:00:28、01:12、01:38、02:01、04:11、04:52、05:07、05:49、07:21、08:14

扫描说明

逐条读取了 Udemy 页面中的 194 条英文 Transcript cue,并按时间轴扫描全长约 08:49 的视频。依据 Accelerator、Listener、Endpoint Group、空 Endpoint、双区域 EC2、源站验证、Region 修正、Endpoint 配置、Health Check 与 Static IP、VPN 路由验证及清理划分为 11 个小节。Udemy 对部分连续字幕使用共用跳转点,时间轴保留页面返回的 cue 时间;截图按实际视频画面复核,均在侧栏关闭、播放器控件隐藏后截取,统一为 1364×770。

Transcript 来自 Udemy 页面自带的英文自动字幕,没有使用 ASR。下面保留页面原文,包括其中的语法或识别错误。

1. 创建 Standard IPv4 Accelerator

时间范围:00:00–00:28

Creating a Standard IPv4 AWS Global Accelerator

总结 / Summary

中文

进入 Global Accelerator 创建 Demo Accelerator,类型选择 Standard,IP Address Type 选择 IPv4。Standard 会把流量自动路由到距用户最近的健康 Endpoint。

English

Create a demo accelerator, choose the Standard accelerator type, and select IPv4 as the IP address type. A standard accelerator automatically routes traffic to the healthy endpoint nearest to the user.

Transcript

00:00 All right, so let's go ahead

00:01 and practice using Global Accelerator.

00:04 So, for this, we're gonna have to create an infrastructure

00:06 in two regions to make the Global Accelerator useful.

00:06 But let's go ahead and create it

00:10 and see what the options are.

00:13 So, first of all, we need to enter an accelerator name,

00:13 so Demo, and then an accelerator type.

00:19 It's going to be Standard.

00:20 Then, the routing we want is going to be IPv4

00:24 because right now we use IPv4 addresses.

00:26 Let's click on Next.

2. 添加 TCP 80 Listener

时间范围:00:28–01:12

Adding a TCP port 80 listener to the accelerator

总结 / Summary

中文

Listener 使用 TCP Protocol 与 Port 80,因为实验中的 EC2 Web Server 接收 HTTP Traffic。Client Affinity 保持 None;如需让同一 Client 固定到同一 Backend,可改为 Source IP。

English

The listener uses TCP on port 80 because the EC2 web servers receive HTTP traffic. Client affinity remains None; Source IP affinity could instead keep the same client mapped to the same backend.

Transcript

00:28 And now we have listeners,

00:29 so what do we want our accelerator to listen on?

00:29 So the protocol is gonna be TCP,

00:33 and the port is going to be 80

00:33 because we're going to send HTTP traffic to our instances.

00:33 Client affinity allows you to handle

00:33 whether or not you want the same client

00:33 to go to the same backend instance.

00:33 So you can have none or have it by source IP,

00:33 but for now we'll leave it as none.

00:33 Let's go Next, and now we have listeners.

00:33 So this is where you create endpoint groups

00:33 for multiple listeners.

00:33 So we can configure regions,

00:33 and then we can configure instances behind the scenes.

00:33 So, for example, here we are going to choose

00:33 US-East-1, very nice.

00:33 And then we add a second endpoint group,

3. 规划两个 Regional Endpoint Group

时间范围:01:12–01:38

Selecting Regions for Global Accelerator endpoint groups

总结 / Summary

中文

为 Listener 添加两个 Endpoint Group,目标是 us-east-1 与 eu-west-1。Global Accelerator 本身显示为 Global Service,但每个 Endpoint Group 对应一个 Region;Traffic Dial 默认 100%,Port Override 和 Health Check 稍后再配置。

English

Add two endpoint groups to the listener, targeting us-east-1 and eu-west-1. Global Accelerator itself is a global service, but each endpoint group belongs to one Region. Traffic dial stays at 100%, with port overrides and health checks configured later.

Transcript

01:12 which is going to be EU-West-1,

01:14 which is the region that I want to use right now.

01:14 As you can see, for Global Accelerator,

01:14 because it's a global service,

01:14 on the top-hand side it says "Global".

01:14 So that means you cannot choose any region,

01:14 it's just a global service.

01:14 So, here we say, "Hey, we want to send

01:14 some traffic to US-East-1 and some traffic to US-West-1."

01:14 You could configure port overrides and health checks,

01:14 but no need to right now.

01:14 Let's click on Next.

4. 先创建 Accelerator,再补充 Endpoint

时间范围:01:38–02:01

Endpoint groups showing no available endpoints during accelerator creation

总结 / Summary

中文

创建向导显示两个 Region 尚无可用 Endpoint,因此先完成 Accelerator,再去各 Region 创建 EC2 Instance 或 Load Balancer,最后返回绑定资源。

English

The creation wizard shows that neither Region has an available endpoint. Finish creating the accelerator first, create EC2 instances or load balancers in the Regions, and then return to associate those resources.

Transcript

01:38 And now we need to have endpoints in these regions.

01:38 So right now we don't have any endpoints,

01:38 so we cannot finish the configuration

01:38 but let's create this accelerator.

01:38 And while it's created,

01:38 we're going to add endpoints afterward.

01:38 So now endpoints are going to be,

01:57 for example, EC2 instances or load balancers and so on.

5. 在 us-east-1 与 eu-west-1 各启动一台 EC2

时间范围:02:01–04:11

Launching regional EC2 web servers with user data

总结 / Summary

中文

在两个 Region 分别启动 Amazon Linux t3.micro,不使用 Key Pair,并让 Security Group 允许来自 Internet 的 HTTP。User Data 安装并启动 httpd,页面输出 Hostname 和 Region,便于识别 Global Accelerator 最终选择的 Endpoint。

English

Launch an Amazon Linux t3.micro instance in each Region without a key pair and allow internet HTTP in the security group. User data installs and starts httpd and returns the hostname and Region so the selected Global Accelerator endpoint is easy to identify.

Transcript

02:01 So, I'm going to open EC2 instances

02:03 and first I'm going to choose US-East-1,

02:06 which is what I want to have one of my endpoint groups in.

02:09 And I will add an instance, and I will launch the instance.

02:13 We're gonna go through it very quickly,

02:14 so I'll call this one Demo Accelerator.

02:14 We'll use the latest Amazon Linux, this is fine.

02:14 T3 Micro, this is fine.

02:14 We don't need a key pair.

02:14 We can select a security group.

02:14 Yes, we're going to create one,

02:14 and we're going to allow HTTP traffic from the internet.

02:14 This is going to allow our HTTP instance to be exposed.

02:14 And then under Advanced Details

02:14 I'm going to, at the very bottom, select my user data,

02:14 and I'm going to choose what we had before, so this,

02:14 but we're going to paste it in

02:14 and we're gonna say, "Hello World

02:14 from hostname (indistinct) in US-East-1."

02:53 This way we know which region

02:53 and which instance belongs to which region.

02:53 So this is good, I'm going to just launch the instance

02:53 and copy this again.

02:53 So now our instance is being launched.

02:53 Ao I'm going to do the exact same thing,

02:53 so let's go to the instance, refresh this,

02:53 and we're going to copy this

02:53 IPv4 address in here and enter it.

02:53 Right now, it's not going to work, of course,

02:53 because it's too early,

02:53 but at least we have the IP address ready.

02:53 And we're going to do the exact same thing,

02:53 but in EU-West-1, so I click on EU-West-1.

02:53 I'm going to launch an instance

02:53 and then call it one Demoinstance.

02:53 Then, the same setup as before, no key pair required.

02:53 We'll allow HTTP traffic from the internet.

02:53 Scroll down, and we're going to paste the user data

02:53 but this time, for the user data, which is right here,

02:53 I'm going to have it as EU-West-1.

02:53 All right, so now we'll have two instances.

02:53 I'm going to launch this one as well.

02:53 So we have two instances now that have been launched,

6. 直接验证两个 Regional Web Server

时间范围:04:11–04:52

Verifying the us-east-1 EC2 web server directly

总结 / Summary

中文

分别通过两台 EC2 的 Public IPv4 Address 访问页面,确认 us-east-1 与 eu-west-1 都能返回对应的 Hello World。先验证 Origin 可以把后续问题限定在 Accelerator 配置。

English

Open each EC2 public IPv4 address and confirm that the us-east-1 and eu-west-1 pages return their respective Hello World responses. Verifying the origins first isolates later issues to accelerator configuration.

Transcript

04:11 one in each region, and we have two IPs

04:11 so I'm going to open this IP address in here as well.

04:11 So let's see for this instance if it's ready or not.

04:11 So let's switch regions again to get back the IP

04:11 because I lost it.

04:33 So here's my instance.

04:36 And, as you can see, it says, "Hello World

04:38 from my EC2 instance in US-East-1."

04:41 So this one has been working correctly,

04:44 and this one is probably just a matter of time.

04:46 Here we go, here it's, "Hello World from EU-West-1."

04:49 So now our two instances are ready.

7. 检查 Listener 并发现 Region 选择错误

时间范围:04:52–05:07

Global Accelerator listener showing two endpoint groups

总结 / Summary

中文

打开 TCP 80 Listener 后发现第二个 Endpoint Group 误建成 us-west-1,而目标 EC2 位于 eu-west-1。Endpoint 必须与 Endpoint Group 位于同一个 Region,因此需要删除错误 Group 并重新添加。

English

Opening the TCP port 80 listener reveals that the second endpoint group was mistakenly created in us-west-1 while the EC2 instance is in eu-west-1. Because an endpoint must be in the same Region as its endpoint group, the incorrect group must be removed and replaced.

Transcript

04:52 And in Global Accelerator, in our accelerator,

04:55 we're going to now configure these instances.

04:58 So let's click on this listener,

04:58 and we have two endpoint groups,

04:58 one for US-East-1 and one for US-West...

8. 添加正确的 Endpoint Group 与 EC2 Endpoint

时间范围:05:07–05:49

Selecting an EC2 endpoint type and endpoint weight

总结 / Summary

中文

新增 eu-west-1 Endpoint Group,并在 us-east-1 与 eu-west-1 各选择对应 EC2 Instance。可选 Endpoint Type 包括 ALB、NLB、EC2 Instance 和 Elastic IP;Weight 用于同一 Group 内有多个 Endpoint 时分配流量,本实验保持默认值。

English

Add the eu-west-1 endpoint group and select the matching EC2 instance in both us-east-1 and eu-west-1. Supported endpoint types include ALB, NLB, EC2 instance, and Elastic IP. Weight controls distribution among multiple endpoints in one group and remains at its default here.

Transcript

05:07 And actually, I did US-West-1 and not EU-West-1,

05:07 so let's configure one for EU-West-1,

05:07 which is what I wanted to.

05:07 And now we're going to add endpoints for this EU-West-1

05:07 and, as you can see in here, we have several options.

05:07 So we have Application Load Balancer, Network Load Balancer,

05:07 EC2 instance, and Elastic IP address.

05:07 Right now, we've configured an EC2 instance,

05:28 and we can find this instance right here.

05:31 And the weight allows you to

05:35 see how you want to distribute traffic

05:35 if you had multiple endpoints.

05:35 But this is good, we have one, we'll save it.

05:39 Now, for this one right here, US-West-1,

05:39 we actually don't need it so I'm going to remove it.

05:39 This was an oversight from me.

9. 配置 Health Check 并等待两个 Endpoint Healthy

时间范围:05:49–07:21

Deployed accelerator with two static IPs and healthy endpoint groups

总结 / Summary

中文

两个 Endpoint Group 都使用 HTTP、Path /、Port 80、10 秒 Interval 与 Threshold 2 的 Health Check。等待 Accelerator Provisioning Status 变为 Deployed 后,检查通过,Listener 显示 All Healthy,并可看到两枚 Static IPv4 Address。

English

Both endpoint groups use HTTP health checks on path / and port 80 with a 10-second interval and threshold of 2. After the accelerator reaches Deployed status, the checks pass, the listener shows All healthy, and two static IPv4 addresses are available.

Transcript

05:49 And for US-East-1, I'm going to click on it, add endpoints.

05:50 And here, again, we can add our EC2 instance.

05:50 Here we go, and click on Save.

05:50 So now we're having issues

05:50 with the health check setup for our EC2 instances,

05:50 so let's go into the configuration of US-East-1.

05:50 We're going to configure the health checks,

05:50 and here we'll say, "Okay, check the protocol HTTP.

05:50 The path is /, the port is 80,

05:50 and you can have a 10-second health check interval,

05:50 and we'll have it as two threshold count to make it faster."

05:50 So perfect for this one.

05:50 And we're going to configure

05:50 the second endpoint group as well, the one for EU-West-1,

05:50 and we'll apply the exact same thing.

05:50 So we're going to do health checks on HTTP,

05:50 the health check being 10 seconds,

05:50 and the threshold being two.

05:50 So that means we need to wait about 20 seconds

05:50 to make sure that our endpoints are going to be healthy,

06:53 so let's wait and I'll get back to you.

06:56 All right, so I'm back in my Global Accelerator,

06:59 you can get back here,

07:00 and I had to wait for the provisioning status to be deployed

07:04 for my health checks to pass.

07:05 So this is why the health checks were unhealthy.

07:07 But now that it's been provisioned

07:10 and my health checks are passing,

07:11 we can see that we have two static IP addresses

07:16 available for our Global Accelerator.

07:18 We have this one and that one.

07:20 And so what I can do now

10. 用 Static IP 与 VPN 验证 Anycast Routing

时间范围:07:21–08:14

Testing Global Accelerator from a United States VPN location

总结 / Summary

中文

直接访问任一 Static IP 时,流量被送到距离当前用户较近的 eu-west-1。连接 United States VPN 后刷新,两个 IP 都返回 us-east-1;断开 VPN 后又回到 eu-west-1,证明 Anycast 与最近 Edge Routing 正常工作。

English

Opening either static IP normally reaches the closer eu-west-1 endpoint. After connecting through a United States VPN, both IPs return us-east-1; disconnecting the VPN returns traffic to eu-west-1, confirming anycast and nearest-edge routing.

Transcript

07:21 is just try to use the Global Accelerator.

07:21 So I'm going to open a URL,

07:21 and, as you can see, by opening this IP address

07:21 I'm redirected to my instance in EU-West-1.

07:21 And I keep on refreshing

07:21 and I keep on being redirected to the same instance.

07:21 An dif I go to this other IP address,

07:21 I also have only EU-West-1.

07:21 So the reason is because I'm redirected

07:21 to the one that's closest to me,

07:47 but I have a VPN just to show you.

07:47 So I can connect into the United States

07:48 to be closer to the US itself,

07:48 and then go here and refresh

07:48 and now I'm connected to US-East-1.

07:48 So, here, both of them are US-East-1.

07:48 So this shows you that the Global Accelerator

07:48 is working as intended.

07:48 For this part of the demo you would need to have a VPN,

08:07 but, again, if you don't have one

08:07 that's completely fine, just see as I do.

08:07 And as you can see, again, if I refresh now

08:07 we're back in EU-West-1.

08:07 So this is pretty cool.

11. Disable、Delete 并终止双区域 EC2

时间范围:08:14–08:49

Selecting the accelerator for cleanup

总结 / Summary

中文

实验结束后先 Disable Accelerator,等待状态更新后再 Delete。随后切换到两个 Region,分别 Terminate 两台 EC2 Instance,避免 Global Accelerator 和计算资源继续计费。

English

After the exercise, disable the accelerator and delete it once the status updates. Then visit both Regions and terminate the two EC2 instances so neither the accelerator nor compute resources continue generating charges.

Transcript

08:14 And now to ensure you don't have any costs,

08:14 please remove the accelerator, so delete it.

08:14 You need to disable it,

08:23 and once it's disabled you can delete it.

08:23 And then remember that you have instances in two regions,

08:23 so terminate it.

08:23 Terminate one and then terminate two.

08:23 So they're both being deleted, that's perfect.

08:23 My accelerator is now disabled so I can just delete it,

08:23 and we're good to go.

08:44 So, that's it for this lecture.

08:44 I hope you liked it, and I will see you in the next lecture.

整讲总结 / Overall Summary

中文

本实验创建 Standard IPv4 Global Accelerator、TCP 80 Listener 和两个 Regional Endpoint Group,在 us-east-1 与 eu-west-1 部署 EC2 Web Server,并绑定为 Endpoint。HTTP Health Check 通过后,Accelerator 提供两枚 Static Anycast IP;本地与美国 VPN 测试证明流量会进入最近 Edge 并转向合适 Region。重点注意 Endpoint 必须与 Group 同 Region、Weight 控制组内分流,以及清理时必须先 Disable Accelerator 再 Delete,同时终止两个 Region 的 EC2。

English

This exercise creates a Standard IPv4 Global Accelerator, a TCP port 80 listener, and two regional endpoint groups, then deploys EC2 web servers in us-east-1 and eu-west-1 and associates them as endpoints. After HTTP health checks pass, the accelerator exposes two static anycast IPs; local and US VPN tests show traffic entering the nearest edge and reaching the appropriate Region. Key details are that endpoints must match their endpoint-group Region, weights control distribution within a group, and cleanup requires disabling before deleting the accelerator and terminating EC2 instances in both Regions.

返回顶部 ↑