Pratique du Shadowing: System Design: Why is Kafka Popular? - Apprendre l'anglais à l'oral avec la vidéo
Choisissez un mode d'entraînement
Chaque mode ouvre une salle d'entraînement en plein écran.
86 phrases
Si les phrases sont trop courtes ou trop longues, cliquez sur Edit pour les ajuster.
1
Why do LinkedIn, Netflix, and Uber all use Kafka to handle boolean messages per day?
2
It's not just about scale.
3
Kafka's distributed log design offers something unique, the ability to replay events, decouple services, and absorb traffic spikes.
4
In this video, we'll look at how Kafka achieves this and what trade-offs you are making when you use it.
5
The main reason companies use Kafka is to decouple their systems.
6
Instead of having services talk directly to each other, they communicate through Kafka.
7
This means producers and consumers can evolve independently, and Kafka absorbs traffic spikes that would otherwise overwhelm your systems.
8
It also enables replay for debugging and recovery when things go wrong.
9
So how does this distributed log actually work?
10
When you send a message to Kafka, it gets written to a partition, which is basically append-only log files sitting on disk.
11
These partitions live on servers called brokers, and when you put multiple brokers together, you get a Kafka cluster.
12
Partitions organize into topics, which are categories for your messages.
13
You might have a topic for payments, another for user clicks, and another for video uploads.
14
Producers write messages into topics, and consumers read them.
15
Every message contains a key, a value, a timestamp, and sometimes headers for metadata.
16
The key determines which partition your message lands in.
17
If you send multiple messages with the same key, they will always go to the same partition and stay in order.
18
When you don't provide a key, Kafka spreads messages around to balance the load across partitions.
19
A single broker on modern hardware can handle hundreds of thousands of messages per second
20
and store as much data as your disk can hold.
21
In practice though, the broker will usually hit network bandwidth limits before CPU or the disk becomes the bottleneck.
22
Today's video is sponsored by Warped, the best way to code with AI agents.
23
Too often, agents write code that's almost right, leaving developers stuck debugging instead of shipping.
24
Warped is different.
25
Ranked top of Terminal Bench and SWE Bench verified.
26
Warped's agent understands your context and writes production-ready code out of the box.
27
Prom, reveal, and refine all in one interface.
28
No context switching, no wasted time.
29
You stay in control.
30
And it pays off.
31
On average, users are saving over an hour a day with warp.
32
Download warp by clicking the link in the description.
33
Partitioning strategy is what determines whether your system scales gracefully or falls apart under low.
34
Pick the wrong partition key and you will end up with hard partitions, where one partition gets hammered while the others sit idle.
35
Imagine you're building a streaming service and you partition by movie ID.
36
Everything works fine until Friday night when a blockbuster drops and suddenly millions of users are streaming the same movie.
37
All these events hit the same partition and your system starts choking.
38
The solution is to use compound keys.
39
Combine the movie ID with a hash of the user ID
40
and now events for that blockbuster get spread across multiple partitions while each user sessions stay in order.
41
There are other partitioning schemes too, each with its own trade-offs.
42
For example, time-based partitions work great for log data because they make retention policies simple, but they complicate real-time aggregation.
43
Consumers track their progress through partitions using offsets, which are basically bookmarks to tell you which message you last processed.
44
They save these offsets back to Kafka periodically, so if they crash, they know exactly where to pick The timing of these commits matters.
45
Commit too early and you might lose messages if you crash.
46
Commit too late and you might process the same message twice.
47
Consumer groups let multiple consumers work together, with Kafka making sure each message gets processed by exactly one consumer in the group.
48
If a consumer fails, Kafka reassigns its partition to the surviving consumers through rebalancing.
49
It handles most failure scenarios without any manual intervention.
50
Kafka offers three delivery guarantees.
51
At most one is fast, but might lose messages.
52
At least once ensures no loss but might produce duplicates.
53
Exactly once is possible, but it is complicated to set up and run slower.
54
Durability comes with replication.
55
Every partition has one leader that handles all reads and writes, plus several followers that copy everything the leader does.
56
If the leader fails, one of the followers takes over.
57
Most production systems run with three replicas, which means you can lose a broker and still have backup.
58
You can configure Kafka to wait for all active replicas to acknowledge rights before considering them successful.
59
This gives you maximum safety but slows things down.
60
With three replicas, you can typically survive one broker failure without losing data.
61
These mechanics enable powerful patterns in production.
62
At Uber, location updates for millions of drivers reportedly flow through Kafka to calculate search pricing in real time.
63
They partition geographically so each region can scale independently.
64
Some companies use Kafka as their source of truth for data.
65
Instead of updating database records directly, they append every state change as an event to Kafka.
66
Want the current state?
67
We play the events.
68
This pattern, called event sourcing, gives you a complete audit trail of everything that happened in your system.
69
But Kafka isn't the right choice for every use case.
70
It optimizes for throughput, not latency.
71
The batching and buffering that enables high throughput adds some delay, making it unsuitable for request response patterns.
72
Kafka only guarantees order within a single partition, not across an entire topic.
73
If you absolutely need global ordering, you are stuck with a single partition, which kills your ability to parallelize.
74
Most systems work around this by accepting partial ordering.
75
Exactly once processing requires careful setup on both producer and consumer size,
76
but when you need it for financial transactions or critical data pipelines, the complexity is worth it.
77
Kafka works because it decouples producers from consumers, letting them evolve independently without breaking each other.
78
Traffic spies that would overwhelm a direct connection get absorbed by the log.
79
When something goes wrong in production, you can replay events to see exactly what happened.
80
But this power comes with a cost.
81
Kafka adds significant operational complexity to your stack.
82
Ready to age your next technical interview?
83
Join our community where we offer comprehensive courses on system design, coding, behavioral questions, machine learning, and object-oriented design.
84
Learn more at bytebytego.com.
85
Thank you.
86
Thank you.
Télécharger l'application
Tout pour parler couramment
Prononciation IAChaque phrase notée
Phonétique (IPA)Maîtrisez chaque son
VocabulaireEnrichissez votre lexique
Jeu de vocabulaireApprenez en jouant
Même chaîne
C13
Proxy vs Reverse Proxy (Real-world Examples)
Daily ConversationUniversity StudentsWorking Adults
C12
Cache Systems Every Developer Should Know
Daily ConversationUniversity StudentsWorking Adults
C12
Top 5 Redis Use Cases
Daily ConversationUniversity Students
C13
Everything You Need to Know About DNS: Crash Course System Design #4
Daily ConversationUniversity StudentsWorking Adults
C12
CAP Theorem Simplified
Daily ConversationUniversity StudentsWorking Adults
C12
What Is REST API? Examples And How To Use It: Crash Course System Design #3
Daily ConversationUniversity StudentsWorking Adults
C18
8 Most Important System Design Concepts You Should Know
Daily ConversationUniversity StudentsWorking Adults
C11
The Secret Sauce Behind NoSQL: LSM Tree
Daily ConversationUniversity StudentsWorking Adults
C11
System Design: Why is Kafka fast?
Business EnglishUniversity StudentsWorking Adults
C16
What Are Microservices Really All About? (And When Not To Use It)
Business EnglishUniversity StudentsWorking Adults
C13
System Design Was HARD - Until You Knew the Trade-Offs
Daily ConversationUniversity StudentsWorking Adults
C13
How Does a URL Shortener Work?
Business EnglishInterview EnglishUniversity Students
C151
Kubernetes Explained in 6 Minutes | k8s Architecture
Business EnglishUniversity StudentsWorking Adults
C14
What is API Gateway?
Business EnglishUniversity StudentsWorking Adults
C12
10+ Key Memory & Storage Systems: Crash Course System Design #5
University StudentsWorking Adults
C24
What Is Redis Really About? Why Is It So Popular?
Business EnglishUniversity StudentsWorking Adults
C1FANG Interview Question | Process vs Thread
Daily ConversationUniversity Students
B2How Git Works: Explained in 4 Minutes
Business EnglishUniversity StudentsWorking Adults
FOR YOU
Vidéo recommandée
C15
222 phrases
The Multi-Agent Architecture That Actually Ships — Luke Alvoeiro, Factory
Business EnglishUniversity StudentsWorking Adults
C14
164 phrases
System Design Mock Interview: Design Facebook Messenger
Business EnglishUniversity StudentsWorking Adults
C11
186 phrases
Why Shopify Moved Inventory Reservations from Redis to MySQL
Business EnglishUniversity StudentsWorking Adults
C12
339 phrases
Easy Migration from Postgres to Databricks Lakebase
Business EnglishUniversity StudentsWorking Adults
C12
18 phrases
SD91: Reliability vs Scalability
Business EnglishUniversity StudentsWorking Adults
C12
204 phrases
Kafka Tutorial for Beginners | Everything you need to get started
Business EnglishUniversity StudentsWorking Adults
C12
38 phrases
Opportunity Cost and Tradeoffs
Business EnglishUniversity StudentsWorking Adults
C12
11 phrases
What are Design Patterns? | Introduction to Design Patterns and Principles | Geekific
Business EnglishUniversity StudentsWorking Adults
C12
42 phrases
System Design: How to Build an API Rate Limiter
Business EnglishUniversity StudentsWorking Adults
C13
101 phrases
3.3k
Business EnglishInterview EnglishUniversity Students
Vocabulaire et conseils d’expression pour cette leçon
Cette leçon d’expression orale de niveau C1 s’appuie sur la vidéo « System Design: Why is Kafka Popular? ». Les mots qui reviennent le plus souvent : Kafka, partition, message, consumer, event. Cette vidéo contient 86 phrases et 1080 mots à répéter en shadowing. La partie parlée dure 7:38. Le locuteur parle à un rythme régulier d’environ 142 mots par minute, confortable pour le shadowing. Seuls 75 % des mots font partie des 3 000 mots les plus courants en anglais, le vocabulaire est donc exigeant.
Vocabulaire clé de cette vidéo
Les 15 mots les plus avancés de la vidéo, avec leur prononciation et leur sens :
| Mot | Prononciation | Sens |
|---|---|---|
| broker nom | /ˈbɹoʊkɚ/ | courtier, représentant |
| decouple verbe | [diːˈkʌpəɫ] | découpler |
| replica nom | /ˈɹɛplɪkə/ | réplique, copie exacte |
| absorb verbe | /æbˈsɔɹb/ | absorber |
| disk nom | /dɪsk/ | disque |
| independently adverbe | /ˌɪndɪˈpɛndəntli/ | indépendamment |
| append verbe | /ʌˈpɛnd/ | appendre, suspendre |
| blockbuster nom | /ˈblɑkˌbʌs.tɚ/ | blockbuster |
| debug verbe | /diːˈbʌɡ/ | débugger, déboguer |
| overwhelm verbe | /ˌoʊ.vɚˈwɛlm/ | abreuver |
| throughput nom | /ˈθɹuː.pʊt/ | débit |
| warp nom | /ˈwɔːp/ | déformation |
| distribute verbe | /dɪˈstɹɪbjuːt/ | distribuer |
| evolve verbe | /ɪˈvɑlv/ | élaborer |
| spike nom | /spaɪk/ | clou, pointe |
Les verbes à particule que vous entendrez
| Mot | Prononciation | Sens |
|---|---|---|
| fall apart verbe | tomber en morceaux | |
| pay off verbe | soudoyer | |
| set up verbe | /ˌsɛt ˈʌp/ | mettre en place, installer |
| take over verbe | reprendre |
Prononciation à surveiller
Le locuteur utilise 4 contractions et formes réduites, comme don't, isn't, we'll. Prononcez-les sous leur forme courte, telles que vous les entendez.
- Les sons « th »: throughput /ˈθɹuː.pʊt/, bandwidth /ˈbændwɪdθ/
- Les sons « sh » et « zh »: partition /pɑɹˈtɪ.ʃən/, aggregation /æɡɹəˈɡeɪʃən/, replication /ɹɛplɪˈkeɪʃən/, hash /ˈhæʃ/, retention /ɹɪˈtɛnʃən/
- Mots longs — placez bien l’accent: independently /ˌɪndɪˈpɛndəntli/, complexity /kəmˈplɛk.sɪ.ti/, aggregation /æɡɹəˈɡeɪʃən/, durability /dəɹəˈbɪləti/, metadata /ˈmɛt.əˌdeɪ.tə/
Les sons difficiles pour les francophones :
- /h/ — il se souffle, il n’est pas muet: hash /ˈhæʃ/, behavioral /bɪˈheɪvjəɹəl/, header /ˈhɛdɚ/
- /tʃ/ et /dʒ/ — à ne pas adoucir en « ch » et « j »: choke /t͡ʃoʊk/, batch /bæt͡ʃ/
- /r/ anglais — langue recourbée, sans frotter la gorge: partition /pɑɹˈtɪ.ʃən/, broker /ˈbɹoʊkɚ/, replica /ˈɹɛplɪkə/, absorb /æbˈsɔɹb/, replay /ˈɹiːpleɪ/
Comment s’entraîner avec cette vidéo
- Écoutez la vidéo en entier une fois sans parler et notez les mots que vous ne connaissez pas.
- Répétez phrase par phrase à vitesse normale, en reprenant chacune jusqu’à ce que votre rythme corresponde à celui du locuteur.
- Enregistrez-vous et comparez avec l’original, en faisant attention à des mots comme broker, decouple, replica.
Qu'est-ce que la technique du Shadowing ?
Le Shadowing est une technique d'apprentissage des langues fondée sur la science, développée à l'origine pour la formation des interprètes professionnels. Le principe est simple mais puissant : vous écoutez de l'anglais natif et le répétez immédiatement à voix haute — comme une ombre suivant le locuteur avec un décalage de 1 à 2 secondes. Les recherches montrent une amélioration significative de la précision de la prononciation, de l'intonation, du rythme, des liaisons, de la compréhension orale et de la fluidité.