Lecture 137: DX Failure detection with BFD
- 课程：[NEW] AWS Certified Advanced Networking Specialty 2026
- 视频时长：02:58
- 来源：Udemy Lecture 27898836
- Slide 数量：8
- 画面切换：00:41、01:01、01:08、01:17、01:42、01:47、02:28
扫描说明
逐条检查了 Udemy 页面中的 37 条 Transcript cue，并扫描了全长约 2:58 的视频。依据 BFD 基础回顾、Direct Connect 故障转移要点逐条渐显以及 Cisco 路由器配置示例，共划分为 8 个小节。候选截图质检时发现部分图片残留播放器进度线，因此重新连接浏览器并逐张重截；最终 8 张截图均在真实媒体边界内完整呈现，无黑边、无顶部课程栏、无底部或中心播放器控件，也无右侧或下方裁切，并统一输出为 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. 使用 BFD 实现 Direct Connect 快速故障转移
时间范围：00:00–00:41
总结 / Summary
中文
双向转发检测（Bidirectional Forwarding Detection，BFD）可快速发现 Direct Connect（DX）链路故障，从而缩短切换到冗余连接的时间。
English
Bidirectional Forwarding Detection quickly detects Direct Connect link failures and shortens the time required to fail over to a redundant connection.
Transcript
00:00 Hello and welcome to this lecture.
00:01 Now that we have seen some of the architecture when it comes to the high availability of direct connect
00:06 connections.
00:07 Now let's understand how to failover fast, which means if you detect the failure in one of the link,
00:13 how fast you can failover to another link.
00:15 And that is important because then ultimately it would reduce your network downtime or a network outage.
00:22 So for direct connect, there is something called BFD.
00:25 If you enable the BFD that is bidirectional forwarding detection, then you can minimize this failover
00:32 time to a secondary link.
00:35 So in the earlier lecture, if you remember, we already saw what PhD means in generic sense, and it
2. BFD 工作原理回顾
时间范围：00:41–01:01
总结 / Summary
中文
BFD 是轻量的 Hello 协议（Hello Protocol），邻居之间周期性交换控制报文；连续未收到响应时，它能在约一秒内宣布链路失效，并可配合 BGP、OSPF 或静态路由。
English
BFD is a lightweight hello protocol that exchanges control packets between peers; after consecutive missed responses, it can declare a link failed in about one second and work with BGP, OSPF, or static routes.
Transcript
00:41 helps you to take the failure in the network links and it detects that real quick so that you can fail
00:47 over to the secondary link.
00:49 And now in this lecture, let's understand how BFT is applied for the direct Connect connection and
00:54 then how much failover time you would have if you enable the PhD and if you don't enable the BFT.
3. Direct Connect 中的 BFD 概览
时间范围：01:01–01:08
总结 / Summary
中文
在多条 Direct Connect（DX）连接或 VPN 备份架构中，BFD 用于更快地把流量切换到冗余连接。
English
In designs with multiple Direct Connect links or VPN backup, BFD provides faster failover to the redundant connection.
Transcript
01:01 So while using the multiple connection or using VPN backup, it's always important to have fast failure
4. 冗余架构需要快速故障转移
时间范围：01:08–01:17
总结 / Summary
中文
仅有冗余链路并不足够；还必须快速检测主路径故障，才能减少网络停机时间（Network Downtime）。
English
Redundant links alone are not enough; the primary-path failure must be detected quickly to reduce network downtime.
Transcript
01:08 to a redundant connection.
01:09 So if you don't apply the BFD, so by default BGP waits for three keeps alive to fail at a hold down
5. 默认 BGP 检测最长约 90 秒
时间范围：01:17–01:42
总结 / Summary
中文
未使用 BFD 时，BGP 通常等待三个 Keepalive 报文丢失，并按约 90 秒的保持时间（Hold Time）判断故障，因此切换可能较慢。
English
Without BFD, BGP typically waits for three missed keepalives and a hold time of about 90 seconds before declaring failure, so failover can be slow.
Transcript
01:17 time of 90 seconds, so it takes up to 90 seconds to detect the failure if BFT is not applied.
01:25 However, as you know, with BFD, that time can be reduced to as low as one second.
01:30 Now, BFD is not something which is, you know, enabled at your end.
01:36 Automatically, however, will enable the BFD by default on DCS connections.
6. AWS 侧默认启用 BFD
时间范围：01:42–01:47
总结 / Summary
中文
AWS 在 Direct Connect（DX）连接的 AWS 侧默认启用 BFD，但客户还必须在自己的路由器上配置 BFD，检测才能端到端生效。
English
AWS enables BFD by default on the AWS side of Direct Connect connections, but customers must also configure it on their routers for end-to-end detection.
Transcript
01:42 But as I said, you need to make sure that it is also enabled at your end.
7. 300 毫秒间隔与三次检测
时间范围：01:47–02:28
总结 / Summary
中文
AWS 使用 300 毫秒的 BFD 存活检测间隔（Liveness Detection Interval）和 3 次检测计数（Detection Multiplier）；连续三次未响应约为 900 毫秒，因此故障转移可低于一秒。
English
AWS uses a 300 ms BFD liveness-detection interval and a detection multiplier of three; three consecutive missed responses take about 900 ms, enabling failover in under one second.
Transcript
01:47 And if you talk about the calculation of one second typically sets the BFD liveliness detection interval
01:54 to 300 milliseconds and then BFD liveliness detection count is three, which means if there are three
02:01 consecutive failures, that means the peer does not send the response back for three consecutive packets
02:07 in the interval of 300 milliseconds.
02:10 Overall, around in 900 milliseconds, you will detect that failure.
02:14 So you can think around within one second.
02:18 And these settings would differ as per your router type.
02:21 Now, if you are using the Cisco router, then you have to fire this command to set your BFD interval
8. Cisco 路由器 BFD 配置示例
时间范围：02:28–02:58
总结 / Summary
中文
Cisco 路由器示例命令 bfd interval 300 min_rx 300 multiplier 3 把发送间隔、最小接收间隔都设为 300 毫秒，并在三次未响应后判定故障；具体语法取决于路由器型号。
English
The Cisco example bfd interval 300 min_rx 300 multiplier 3 sets both transmit and minimum receive intervals to 300 ms and declares failure after three misses; exact syntax depends on the router platform.
Transcript
02:28 to be 300 millisecond and then three is the count of the failure detections.
02:35 So this is what the BFD means.
02:38 And as I said for your exam, it is good to know that with BFD you can lower the network failure detection
02:45 to as low as one second.
02:48 So that's it about BFD.
02:50 And with that now let's get into the security side of the direct connect connection in the next lecture.
02:55 So thanks for watching.
整讲总结 / Overall Summary
中文
双向转发检测（Bidirectional Forwarding Detection，BFD）为 Direct Connect（DX）冗余架构提供快速故障检测。未启用 BFD 时，BGP 可能依赖约 90 秒的 Hold Time；启用后，AWS 侧采用 300 毫秒检测间隔和 3 次计数，可在约 900 毫秒内发现故障。AWS 侧默认启用 BFD，但客户路由器也必须配置相匹配的参数，才能实现低于一秒的故障转移。
English
Bidirectional Forwarding Detection provides rapid failure detection for redundant Direct Connect architectures. Without BFD, BGP may rely on a hold time of roughly 90 seconds; with BFD, AWS uses a 300 ms interval and a multiplier of three, detecting failure in about 900 ms. AWS enables BFD on its side by default, but the customer router must use matching settings to achieve sub-second failover.
返回顶部 ↑