ByteByteGo

Co-authored 4 books:
     - System Design Interview - Volume 1
     - System Design Interview - Volume 2
     - Coding Interview Patterns
     - Machine Learning System Design Interview

- Write weekly ByteByteGo newsletter on system design topics to 1,000,000+ subscribers.
- Grow the ByteByteGo YouTube channel to 1,000,000+ subscribers.

Author & Co-Founder

Twitter

Software Engineer

Apple

Zynga

Oracle

Software Developer

Modeln N

Associate Member of Technical Staff

Follow me for system design & book-writing tips.

Alex Xu is a software engineer and author. His book System Design Interview - An Insider’s Guide is one of Amazon's best-selling books in the “Distributed Systems & Computing” category.  As of December 2020, his book is being translated into 7 different languages (Traditional Chinese, Russian, Japanese, Spanish, Polish, Simplified Chinese, and Korean).

--------------
Newsletter: https://blog.bytebytego.com
Twitter: https://twitter.com/alexxubyte
YouTube: https://www.youtube.com/c/ByteByteGo

Alex Xu

Database Types You Should Know in 2025

There’s no such thing as a one-size-fits-all database anymore. Modern applications rely on multiple database types, from real-time analytics to vector search for AI. Knowing which type to use can make or break your system’s performance.

Relational: Traditional row-and-column databases, great for structured data and transactions.

Columnar: Optimized for analytics, storing data by columns for fast aggregations.

Key-Value: Stores data as simple key–value pairs, enabling fast lookups. 

In-memory: Stores data in RAM for ultra-low latency lookups, ideal for caching or session management.

Wide-Column: Handles massive amounts of semi-structured data across distributed nodes.

Time-series: Specialized for metrics, logs, and sensor data with time as a primary dimension.

Immutable Ledger: Ensures tamper-proof, cryptographically verifiable transaction logs.

Graph: Models complex relationships, perfect for social networks and fraud detection

Document: Flexible JSON-like storage, great for modern apps with evolving schemas.

Geospatial: Manages location-aware data such as maps, routes, and spatial queries.

Text-search: Full-text indexing and search with ranking, filters, and analytics.

Blob: Stores unstructured objects like images, videos, and files.

Vector: Powers AI/ML apps by enabling similarity search across embeddings.

Over to you: Which database type do you think will grow fastest in the next 5 years?

--
We just launched the all-in-one tech interview prep platform, covering coding, system design, OOD, and machine learning.

Launch sale: 50% off. Check it out: https://lnkd.in/euwKh6u8

#systemdesign #coding #interviewtips 
.

bytebytego

linkedin.com/company/the-rundown-ai

JSESSIONID

li_at

Talks about #systemdesign, #techinterview, and #bookpublishing

Author of 4 Bestselling Books | Co-Founder of ByteByteGo

Cloudflare is much more than just a CDN and DDoS protection service.

Let’s do a quick comparison of Cloudflare, AWS, and Azure.

Cloudflare has rapidly expanded beyond its traditional CDN roots, launching a suite of modern developer-first services like Workers, R2, D1, and so on. These offerings position it as a serious edge-native alternative to other cloud providers.

Here are the key cloud capabilities that Cloudflare supports:

1 - Edge Compute and Serverless
2 - Object and Blob Storage
3 - Relational Databases
4 - Containers
5 - Sandboxes
6 - Workflows
7 - AI Agents SDK
8 - Vector and AI search
9 - Data Connectivity
10 - AI Infrastructure
11 - Content Delivery Network
12 - DNS
13 - Load Balancing

Over to you: Have you used Cloudflare’s new offerings? What are your thoughts on them?

--
We just launched the all-in-one tech interview prep platform, covering coding, system design, OOD, and machine learning.

Launch sale: 50% off. Check it out: https://lnkd.in/euwKh6u8

#systemdesign #coding #interviewtips 
.

No alternative text description for this image

HTTP vs. HTTPS

When you open a website, the difference between HTTP and HTTPS decides whether your data travels safely or in plain sight. Here’s what actually happens under the hood:

HTTP:
- Sends data in plain text, anyone on the network can intercept it.
- The client and server perform a simple TCP handshake: SYN, SYN-ACK, ACK
- Fast but completely insecure. Passwords, tokens, and forms can all be read in transit.

HTTPS (SSL/TLS):
- Step 1: TCP Handshake: Standard connection setup.
- Step 2: Certificate Check: Client says hello. Server responds with hello and its SSL/TLS certificate. That certificate contains the server's public key and is signed by a trusted Certificate Authority. 

Your browser verifies this certificate is legitimate, not expired, and actually belongs to the domain you're trying to reach. This proves you're talking to the real server, not some attacker pretending to be it.

- Step 3: Key Exchange: Here's where asymmetric encryption happens. The server has a public key and a private key. Client generates a session key, encrypts it with the server's public key, and sends it over. Only the server can decrypt this with its private key.

Both sides now have the same session key that nobody else could have intercepted. This becomes the symmetric encryption key for the rest of the session.

- Step 4: Data Transmission: Now every request and response gets encrypted with that session key using symmetric encryption.

Over to you: What’s your go-to tool for debugging TLS issues, openssl, curl -v, or something else?

--
We just launched the all-in-one tech interview prep platform, covering coding, system design, OOD, and machine learning.

Launch sale: 50% off. Check it out: https://lnkd.in/euwKh6u8

#systemdesign #coding #interviewtips 
.

Karnav P.

Apache Kafka vs. RabbitMQ

Kafka and RabbitMQ both handle messages, but they solve fundamentally different problems. Understanding the difference matters when designing distributed systems.

Kafka is a distributed log. Producers append messages to partitions. Those messages stick around based on retention policy, not because someone consumed them. Consumers pull messages at their own pace using offsets. You can rewind, replay, reprocess everything. It is designed for high throughput event streaming where multiple consumers need the same data independently.

RabbitMQ is a message broker. Producers publish messages to exchanges. Those exchanges route to queues based on binding keys and patterns (direct, topic, fanout). Messages get pushed to consumers and then deleted once acknowledged. It is built for task distribution and traditional messaging workflows.

The common mistake is using Kafka like a queue or RabbitMQ like an event log. They're different tools built for different use cases.

Over to you:  If you had to explain when NOT to use Kafka, what would you say?

--
We just launched the all-in-one tech interview prep platform, covering coding, system design, OOD, and machine learning.

Launch sale: 50% off. Check it out: https://lnkd.in/euwKh6u8

#systemdesign #coding #interviewtips 
.

Explaining 9 types of API testing. 
 
🔹 Smoke Testing 
This is done after API development is complete. Simply validate if the APIs are working and nothing breaks. 
 
🔹 Functional Testing 
This creates a test plan based on the functional requirements and compares the results with the expected results. 
 
🔹 Integration Testing 
This test combines several API calls to perform end-to-end tests. The intra-service communications and data transmissions are tested. 
 
🔹 Regression Testing 
This test ensures that bug fixes or new features shouldn’t break the existing behaviors of APIs. 
 
🔹 Load Testing 
This tests applications’ performance by simulating different loads. Then we can calculate the capacity of the application. 
 
🔹 Stress Testing 
We deliberately create high loads to the APIs and test if the APIs are able to function normally. 
 
🔹 Security Testing 
This tests the APIs against all possible external threats. 
 
🔹 UI Testing 
This tests the UI interactions with the APIs to make sure the data can be displayed properly. 
 
🔹 Fuzz Testing 
This injects invalid or unexpected input data into the API and tries to crash the API. In this way, it identifies the API vulnerabilities. 

--
We just launched Become an AI Engineer | Learn by Doing: Cohort 2. If you missed Cohort 1, now’s your chance to join us for Cohort 2.

Check it out here: https://bit.ly/3LdaWuw

#AI #AIEngineer #MachineLearning 
.

The AI Agent Tech Stack

1. Foundation Models: Large-scale pre-trained language models that serve as the “brains” of AI agents, enabling capabilities like reasoning, text generation, coding, and question answering.
2. Data Storage: This layer handles vector databases and memory storage systems used by AI agents to store and retrieve context, embeddings, or documents.
3. Agent Development Frameworks: These frameworks help developers build, orchestrate, and manage multi-step AI agents and their workflows.
4. Observability: This category enables monitoring, debugging, and logging of AI agent behavior and performance in real-time.
5. Tool Execution: These platforms allow AI agents to interface with real-world tools (for example, APIs, browsers, external systems) to complete complex tasks.
6. Memory Management: These systems manage long-term and short-term memory for agents, helping them retain useful context and learn from past interactions.

Over to you: What else will you add to the list?

--
We just launched the all-in-one tech interview prep platform, covering coding, system design, OOD, and machine learning.

Launch sale: 50% off. Check it out: https://lnkd.in/euwKh6u8

#systemdesign #coding #interviewtips 
.

Big Data Pipeline Cheatsheet for AWS, Azure, and Google Cloud

Each platform offers a comprehensive suite of services that cover the entire lifecycle:

1 - Ingestion: Collecting data from various sources
2 - Data Lake: Storing raw data
3 - Computation: Processing and analyzing data
4 - Data Warehouse: Storing structured data
5 - Presentation: Visualizing and reporting insights

AWS uses services like Kinesis for data streaming, S3 for storage, EMR for processing, RedShift for warehousing, and QuickSight for visualization.

Azure’s pipeline includes Event Hubs for ingestion, Data Lake Store for storage, Databricks for processing, Cosmos DB for warehousing, and Power BI for presentation.

GCP offers PubSub for data streaming, Cloud Storage for data lakes, DataProc and DataFlow for processing, BigQuery for warehousing, and Data Studio for visualization.

Over to you: What else would you add to the pipeline?

--
LAST DAY to enroll in Become an AI Engineer Learn by Doing, Cohort 2. If you missed Cohort 1, now is your chance to join us in Cohort 2.

Check it out here: https://bit.ly/3LdaWuw

#AI #AIEngineer #MachineLearning 
.

Things Every Developer Should Know: Concurrency is 𝐍𝐎𝐓 parallelism.
.
.
In system design, it is important to understand the difference between concurrency and parallelism. 

As Rob Pyke(one of the creators of GoLang) stated:“ Concurrency is about 𝐝𝐞𝐚𝐥𝐢𝐧𝐠 𝐰𝐢𝐭𝐡 lots of things at once. Parallelism is about 𝐝𝐨𝐢𝐧𝐠 lots of things at once." This distinction emphasizes that concurrency is more about the 𝐝𝐞𝐬𝐢𝐠𝐧 of a program, while parallelism is about the 𝐞𝐱𝐞𝐜𝐮𝐭𝐢𝐨𝐧. 

Concurrency is about dealing with multiple things at once. It involves structuring a program to handle multiple tasks simultaneously, where the tasks can start, run, and complete in overlapping time periods, but not necessarily at the same instant. 

Concurrency is about the composition of independently executing processes and describes a program's ability to manage multiple tasks by making progress on them without necessarily completing one before it starts another.

Parallelism, on the other hand, refers to the simultaneous execution of multiple computations. It is the technique of running two or more tasks or computations at the same time, utilizing multiple processors or cores within a computer to perform several operations concurrently. Parallelism requires hardware with multiple processing units, and its primary goal is to increase the throughput and computational speed of a system.

In practical terms, concurrency enables a program to remain responsive to input, perform background tasks, and handle multiple operations in a seemingly simultaneous manner, even on a single-core processor. It's particularly useful in I/O-bound and high-latency operations where programs need to wait for external events, such as file, network, or user interactions.

Parallelism, with its ability to perform multiple operations at the same time, is crucial in CPU-bound tasks where computational speed and throughput are the bottlenecks. Applications that require heavy mathematical computations, data analysis, image processing, and real-time processing can significantly benefit from parallel execution.

--
We just launched the all-in-one tech interview prep platform, covering coding, system design, OOD, and machine learning.

Launch sale: 50% off. Check it out: https://lnkd.in/euwKh6u8

#systemdesign #coding #interviewtips 
.

5 REST API Authentication Methods

1. Basic Authentication: Clients include a Base64-encoded username and password in every request header, which is simple but insecure since credentials are transmitted in plaintext. Useful in quick prototypes or internal services over secure networks.
2. Session Authentication: After login, the server creates a session record and issues a cookie. Subsequent requests send that cookie so the server can validate user state. Used in traditional web-apps.
3. Token Authentication: Clients authenticate once to receive a signed token, then present the token on each request for stateless authentication. Used in single-page applications and modern APIs that require scalable, stateless authentication.
4. OAuth-Based Authentication: Clients obtain an access token via an authorization grant from an OAuth provider, then use that token to call resource servers on the user’s behalf. Used in cases of third-party integrations or apps that need delegated access to user data.
5. API Key Authentication: Clients present a predefined key (often in headers or query strings) with each request. The server verifies the key to authorize access. Used in service-to-service or machine-to-machine APIs where simple credential checks are sufficient.

Over to you: Which other API Authentication method have you seen?

--
We just launched the all-in-one tech interview prep platform, covering coding, system design, OOD, and machine learning.

Launch sale: 50% off. Check it out: https://lnkd.in/euwKh6u8

#systemdesign #coding #interviewtips 
.

Git Merge vs. Rebase vs. Squash Commit! 
. 
. 
What are the differences? 
 
When we 𝐦𝐞𝐫𝐠𝐞 𝐜𝐡𝐚𝐧𝐠𝐞𝐬 from one Git branch to another, we can use ‘git merge’ or ‘git rebase’. The diagram below shows how the two commands work. 
 
𝐆𝐢𝐭 𝐌𝐞𝐫𝐠𝐞 
This creates a new commit G’ in the main branch. G’ ties the histories of both main and feature branches. 
 
Git merge is 𝐧𝐨𝐧-𝐝𝐞𝐬𝐭𝐫𝐮𝐜𝐭𝐢𝐯𝐞. Neither the main nor the feature branch is changed. 
 
𝐆𝐢𝐭 𝐑𝐞𝐛𝐚𝐬𝐞 
Git rebase moves the feature branch histories to the head of the main branch. It creates new commits E’, F’, and G’ for each commit in the feature branch. 
 
The benefit of rebase is that it has 𝐥𝐢𝐧𝐞𝐚𝐫 𝐜𝐨𝐦𝐦𝐢𝐭 𝐡𝐢𝐬𝐭𝐨𝐫𝐲. 
 
Rebase can be dangerous if “the golden rule of git rebase” is not followed. 
 
𝐓𝐡𝐞 𝐆𝐨𝐥𝐝𝐞𝐧 𝐑𝐮𝐥𝐞 𝐨𝐟 𝐆𝐢𝐭 𝐑𝐞𝐛𝐚𝐬𝐞 
Never use it on public branches!

--
We just launched the all-in-one tech interview prep platform, covering coding, system design, OOD, and machine learning.

Launch sale: 50% off. Check it out: https://lnkd.in/euwKh6u8

#systemdesign #coding #interviewtips 
.

Top 20 System Design Concepts You Should Know

1. Load Balancing: Distributes traffic across multiple servers for reliability and availability.
2. Caching: Stores frequently accessed data in memory for faster access.
3. Database Sharding: Splits databases to handle large-scale data growth.
4. Replication: Copies data across replicas for availability and fault tolerance.
5. CAP Theorem: Trade-off between consistency, availability, and partition tolerance.
6. Consistent Hashing: Distributes load evenly in dynamic server environments.
7. Message Queues: Decouples services using asynchronous event-driven architecture.
8. Rate Limiting: Controls request frequency to prevent system overload.
9. API Gateway: Centralized entry point for routing API requests.
10. Microservices: Breaks systems into independent, loosely coupled services.
11. Service Discovery: Locates services dynamically in distributed systems.
12. CDN: Delivers content from edge servers for speed.
13. Database Indexing: Speeds up queries by indexing important fields.
14. Data Partitioning: Divides data across nodes for scalability and performance.
15. Eventual Consistency: Guarantees consistency over time in distributed databases
16. WebSockets: Enables bi-directional communication for live updates.
17. Scalability: Increases capacity by upgrading or adding machines.
18. Fault Tolerance: Ensures system availability during hardware/software failures.
19. Monitoring: Tracks metrics and logs to understand system health.
20. Authentication & Authorization: Controls user access and verifies identity securely.

Over to you: Which other System Design concept will you add to the list?

--
We just launched the all-in-one tech interview prep platform, covering coding, system design, OOD, and machine learning.

Launch sale: 50% off. Check it out: https://lnkd.in/euwKh6u8

#systemdesign #coding #interviewtips 
.

Virtualization vs. Containerization

Before containers simplified deployment, virtualization changed how we used hardware. Both isolate workloads, but they do it differently.

- Virtualization (Hardware-level isolation): Each virtual machine runs a complete operating system, Windows, Fedora, or Ubuntu, with its own kernel, drivers, and libraries. The hypervisor (VMware ESXi, Hyper-V, KVM) sits directly on hardware and emulates physical machines for each guest OS. 

This makes VMs heavy but isolated. Need Windows and Linux on the same box? VMs handle it easily. Startup time for a typical VM is in minutes because you're booting an entire operating system from scratch.

- Containerization (OS-level isolation): Containers share the host operating system's kernel. No separate OS per container. Just isolated processes with their own filesystem and dependencies.

The container engine (Docker, containerd, CRI-O, Podman) manages lifecycle, networking, and isolation, but it all runs on top of a single shared kernel. Lightweight and fast. Containers start in milliseconds because you're not booting an OS, just launching a process.

But here's the catch: all containers on a host must be compatible with that host's kernel. Can't run Windows containers on a Linux host (without nested virtualization tricks).

Over to you: What's your go-to setup: containers in VMs, bare metal containers, or something else?

--
We just launched the all-in-one tech interview prep platform, covering coding, system design, OOD, and machine learning.

Launch sale: 50% off. Check it out: https://lnkd.in/euwKh6u8

#systemdesign #coding #interviewtips 
.

What is a REST API?

REST (Representational State Transfer) is an architectural style for building APIs that use HTTP for communication. To be considered RESTful, an API should follow six key constraints:

- Client-Server: Separates the user interface from data storage and processing. This allows each side to evolve independently.

- Stateless: Each request contains all the information needed to process it; the server does not store session state between requests.

- Uniform Interface: Consistent resource naming and formats across the API, e.g., “/products”, “/users”.

- Cacheable: Responses explicitly indicate if they can be cached (e.g., Cache-Control) to improve performance.

- Layered System: Requests may pass through multiple layers (load balancers, auth services, etc.) before reaching the API server, but to the client it appears as a single endpoint. 

- Code on Demand (Optional): The server can send executable code (like JavaScript) to the client to extend its functionality.


Over to you: Which REST constraint do you think is most often overlooked in real-world APIs?

--
We just launched the all-in-one tech interview prep platform, covering coding, system design, OOD, and machine learning.

Launch sale: 50% off. Check it out: https://lnkd.in/euwKh6u8

#systemdesign #coding #interviewtips 
.

The HTTP Mindmap

HTTP has evolved from HTTP/1.1 to HTTP/2, and now HTTP/3, which uses the QUIC protocol over UDP for improved performance. Today, it’s the backbone of almost everything on the internet, from browsers and APIs to streaming, cloud, and AI systems.

At the foundation, we have underlying protocols. TCP/IP for IPv4 and IPv6 traffic. Unix domain sockets for local communication. HTTP/3 running over UDP instead of TCP. These handle the actual data transport before HTTP even comes into play.

Security wraps around everything. HTTPS isn't optional anymore. WebSockets power real-time connections. Web servers manage workloads. CDNs distribute content globally. DNS resolves everything to IPs. Proxies (forward, reverse, and API gateways) route, filter, and secure traffic in between.

Web services exchange data in different formats, REST with JSON, SOAP for enterprise systems, RPC for direct calls, and GraphQL for flexible queries. Crawlers and bots index the web, guided by robots.txt files that set the boundaries.

The network world connects everything. LANs, WANs, and protocols like FTP for file transfers, IMAP/POP3 for email, and BitTorrent for peer-to-peer communication. For observability, packet capture tools like Wireshark, tcpdump, and OpenTelemetry let developers peek under the hood to understand performance, latency, and behavior across the stack.

Over to you: HTTP has been evolving for 30+ years, what do you think the next big shift will be?

--
We just launched the all-in-one tech interview prep platform, covering coding, system design, OOD, and machine learning.

Launch sale: 50% off. Check it out: https://lnkd.in/euwKh6u8

#systemdesign #coding #interviewtips 
.

How DNS Works

You type a domain name and hit enter, but what actually happens before that webpage loads is more complex than most people realize. DNS is the phonebook of the internet, and every request you make triggers a chain of lookups across multiple servers. 

Step 1: Someone types bytebytego. com into their browser and presses enter. 

Step 2: Before doing anything, the browser looks for a cached IP address. Operating system cache gets checked too. 

Step 3: Cache miss triggers a DNS query. The browser sends a query to the configured DNS resolver, usually provided by your ISP or a service like Google DNS or Cloudflare.

Step 4: Resolver checks its own cache. 

Step 5-6: If the resolver doesn't have the answer cached, it asks the root servers, “Where can I find .com?” For bytebytego. com, the root server responds with the address of the .com TLD name server. 

Step 7-8: Resolver queries the .com TLD server. TLD server returns the authoritative server address. 

Step 9-10: This server has the actual A/AAAA record mapping the domain to an IP address. The resolver finally gets the answer: 172. 67. 21. 11 for bytebytego. com.

Step 11-12: The IP gets cached at the resolver level for future lookups, and returned to the browser.

Step 13-14: The browser stores this for its own future use, and uses the IP to make the actual HTTP request. 

Step 15: The web server returns the requested content.

All this happens in milliseconds, before your first page even starts loading.

Over to you: Which DNS tools or commands do you rely on most, dig, nslookup, or something else?

--
We just launched the all-in-one tech interview prep platform, covering coding, system design, OOD, and machine learning.

Launch sale: 50% off. Check it out: https://lnkd.in/euwKh6u8

#systemdesign #coding #interviewtips 
.

Linux file permission illustrated. 

𝐎𝐰𝐧𝐞𝐫𝐬𝐡𝐢𝐩
Every file or directory is assigned 3 types of owner:

🔹Owner: the owner is the user who created the file or directory.
🔹Group: a group can have multiple users. All users in the group have the same permissions to access the file or directory.
🔹Other: other means those users who are not owners or members of the group.

𝐏𝐞𝐫𝐦𝐢𝐬𝐬𝐢𝐨𝐧
There are only three types of permissions for a file or directory.

🔹Read (r): the read permission allows the user to read a file.
🔹Write (w): the write permission allows the user to change the content of the file.
🔹Execute (x): the execute permission allows a file to be executed. 

Over to you: chmod 777, good idea?

--
We just launched the all-in-one tech interview prep platform, covering coding, system design, OOD, and machine learning.

Launch sale: 50% off. Check it out: https://lnkd.in/euwKh6u8

#systemdesign #coding #interviewtips 
.

Types of Virtualization

Virtualization didn’t just make servers efficient, it changed how we build, scale, and deploy everything. Here’s a quick breakdown of the four major types of virtualization you’ll find in modern systems:

1. Traditional (Bare Metal): Applications run directly on the operating system. No virtualization layer, no isolation between processes. All applications share the same OS kernel, libraries, and resources.

2. Virtualized (VM-based): Each VM runs its own complete operating system. The hypervisor sits on physical hardware and emulates entire machines for each guest OS. Each VM thinks it has dedicated hardware even though it's sharing the same physical server.

3. Containerized: Containers share the host operating system's kernel but get isolated runtime environments. Each container has its own filesystem, but they're all using the same underlying OS. The container engine (Docker, containerd, Podman) manages lifecycle, networking, and isolation without needing separate operating systems for each application.

Lightweight and fast. Containers start in milliseconds because you're not booting an OS. Resource usage is dramatically lower than VMs.

4. Containers on VMs: This is what actually runs in production cloud environments. Containers inside VMs, getting benefits from both. Each VM runs its own guest OS with a container engine inside. The hypervisor provides hardware-level isolation between VMs. The container engine provides lightweight application isolation within VMs.

This is the architecture behind Kubernetes clusters on AWS, Azure, and GCP. Your pods are containers, but they're running inside VMs you never directly see or manage.

Over to you: In your experience, which setup strikes the best balance between performance and flexibility?

--
We just launched the all-in-one tech interview prep platform, covering coding, system design, OOD, and machine learning.

Launch sale: 50% off. Check it out: https://lnkd.in/euwKh6u8

#systemdesign #coding #interviewtips 
.

Vijaya Kalyani K.

Can a web server provide real-time updates? 
. 
. 
An HTTP server cannot automatically initiate a connection to a browser. As a result, the web browser is the initiator. What should we do next to get real-time updates from the HTTP server? 
 
Both the web browser and the HTTP server could be responsible for this task. 
 
🔹Web browsers do the heavy lifting: short polling or long polling. With short polling, the browser will retry until it gets the latest data. With long polling, the HTTP server doesn’t return results until new data has arrived. 
🔹HTTP server and web browser cooperate: WebSocket or SSE (server-sent event). In both cases, the HTTP server could directly send the latest data to the browser after the connection is established. The difference is that SSE is uni-directional, so the browser cannot send a new request to the server, while WebSocket is fully-duplex, so the browser can keep sending new requests. 
 
Over to you: of the four solutions (long polling, short polling, SSE, WebSocket), which ones are commonly used, and for what use cases? 

--
We just launched the all-in-one tech interview prep platform, covering coding, system design, OOD, and machine learning.

Launch sale: 50% off. Check it out: https://lnkd.in/euwKh6u8

#systemdesign #coding #interviewtips 
.

How do AirTags work?

AirTags work by leveraging a combination of Bluetooth technology and the vast network of Apple devices to help you locate your lost items. 
Here's a breakdown of how they function: 

1. Bluetooth Signal: Each AirTag emits a secure Bluetooth signal that can be detected by nearby Apple devices (iPhones, iPads, etc.) within the Find My network. 

2. Find My Network: When an AirTag comes within range of an Apple device in the Find My network, that device anonymously and securely relays the AirTag's location information to iCloud. 

3. Location Tracking: You can then use the Find My app on your own Apple device to see the approximate location of your AirTag on a map. 

Limitations: 
Please note that AirTags rely on Bluetooth technology and the presence of Apple devices within the Find My network. If your AirTag is in an area with few Apple devices, its location may not be updated as frequently or accurately. 

--
We just launched the all-in-one tech interview prep platform, covering coding, system design, OOD, and machine learning.

Launch sale: 50% off. Check it out: https://lnkd.in/euwKh6u8

#systemdesign #coding #interviewtips 
.

Joshua Debayo -

Nobody cares, just win | Stack up your wins & be so good they can’t ignore you.

Company Owner

Joshua Debayo

I get messages like this all the time now:

“Josh, what are you saying in your DMs to get these replies?”
 “Can I see the exact script you use?”

Today, you can have my framework for FREE.

Not the script.
Not a word-for-word template.

A framework you can plug into your own:

LinkedIn outreach
Cold email
DM follow ups

So you can actually think for yourself.

Because here’s the harsh reality:

You don’t learn much from copying & pasting someone else’s message.

You learn from understanding the thinking behind it:

Why this opener instead of that one
Why this question here
Why this CTA at the end
Why it works for outbound and getting leads to reply

That’s what I’m giving you.

A breakdown of:

How I structure my first message
How I transition into “I’m interested, tell me more” replies
How I write follow ups that don’t feel needy
How I get real conversations started with people who can actually sign as clients

No magic line.

Just a simple process you can adapt to your niche, offer, and voice.

Want the framework? 👇🏾

Like this post
Comment “framework”
I’ll send it over.

PS –  LinkedIn has 1 billion users. I'm only sending this to 497 people.

After that, it’s gone.

Keep working,
You’ve got this,
I’m rooting for you.

In the last 12 days, I’ve added an extra $18,000 in MRR.

I’m giving away my exact word-for-word script to get qualified high intent prospects (rich people) for FREE.

All from LinkedIn.

This is the only platform I use to generate revenue & sign clients.

❌No content on other socials.
❌No hundreds of sales calls.

I’m averaging around 7-10 calls/week. 

(Check the dates on the picture if you think it’d 🧢) 

Today, I’m doing something I probably shouldn’t do.

I’m giving away my exact word-for-word script for FREE.

This isn’t fake scarcity.
I’m only giving this to 97 people.

I don’t want it to get saturated because I still use it myself 🤣

It’s the exact message I use to book sales meetings with recruiting & wealth management firms.

We only partner with companies doing $1M+ in revenue,
so this message works well for higher-level businesses.

Inside, you’ll get:

➡️ The exact word for word script
➡️ The simple 3 step follow up

That’s it.

As a bonus, I’ll add a short video walking through my thought process & explain why it does so well.

Again – this is only being sent to 97 people‼️

⸻

Want the guide + video? 

You must do the following:

👍 Like this post
👤 Follow me
💬 Comment “96”

I’ll send it over.

PS – LinkedIn  has 1 billion users. I’m genuinely only sending this to 97 people.

After that, it’s gone & I’ll probably delete this post.

It took me approx 2 years to get good at signing clients on LinkedIn.

It took me 60 days to get half decent at cold email.

All thanks too Saad Belcaid. 

Being half decent was enough to help me 10x my business in 60 days.

These kinda stuff don't need to take you years to get good at.
It literally took me 60 days to get half decent at cold email.

Both channels are more similar than people think.

You create a list of ideal clients.
You craft a message that turns their head.
You book them into your calendar.

When you break it down like that, it’s simple.
Not easy — but simple.

The same rules apply:

- Stay consistent.
- Test more often.
- Refine what’s working.
- Cut what isn’t.

LinkedIn or cold email…
It’s all the same game with a different wrapper.

If you learn the fundamentals,
You can win on any platform.

Keep working.
You’ve got this.
I’m rooting for you.

PS - I'll give 5 lucky agency owners my script for my campaign like this post comment '7.7' and its yours

I've generated 27 positive replies so far this week on LinkedIn.

Steal my framework for FREE.

(If you think I’m chatting 💩 check my previous post – different faces 🥸)

This system will definitely get saturated.
I won’t be sharing these resources as much in 2026.

From next year, I’ll focus my content on:

➜ Recruitment agencies
➜ Wealth management firms

I can confidently say I know how to consistently sign high-ticket clients on LinkedIn.

The other day I reached out to a complete stranger…
Two days later I closed a $12k deal.

You’ll get my full framework for FREE today.

❌ It’s not luck.
❌ It’s not being at the right place at the right time.

It’s doing the work & making it unreasonable for you not to win.

LinkedIn has made it super easy to reach high-net-worth people and get them interested in your service/product.

Want the framework?

You MUST:

➜ Like this post
➜ Follow me
➜ Comment “20”

I’ll send it over.

PS – I’ll be deleting this post in 12 hours. ⏰

Yesterday we hit a record.

15 positive replies in 6 hours from cold outreach on LinkedIn.

Complete strangers.
Didn’t know anything about me…
except my initial message.

A few of them even wanted me to call them there & then.

Here’s 3 simple reasons you’re not getting positive replies on LinkedIn 👇🏾

1. You don’t speak their language

You’re writing like a marketer, not like someone in their industry.
Use industry lingo.
Use their terminology.

Make them think:
“Okay, this person gets my world.”

2. Your volume is too low

You’re not getting replies…
because you’re not sending enough messages.

Outreach is a numbers game.

More DM’s = more conversations = more Leads and sales meetings from LinkedIn.

3. You’re messaging at the wrong time

Don’t expect replies if you’re reaching out at crazy hours
or when your ICP is slammed.

Pay attention to:

Time zones
Work patterns
When they’re actually online

Right message + right person + right time = replies.

Dial in your language.
Increase your volume.
Respect timing.

Do that consistently and LinkedIn cold Outreach becomes a lot less “cold”.

Keep working,
You’ve got this,
I’m rooting for you

I consistently book 8–10 calls a week on LinkedIn.

But a year ago, I couldn’t buy a meeting.

To anyone who’s struggling to book meetings on LinkedIn don't stress.

I still get more messages like this than positive ones.

It’s just part of the game.
I get messages like this every single day.

 It doesn’t affect me because I know it’s a numbers game.

I started treating outbound on LinkedIn & Instantly.ai like a video game.

Every day I try to get 1% better:

→ My initial message
→ My follow ups
→ My positioning
→ My offer

I’m always polishing something.

Keep working,
You’ve got this,
I’m rooting for you.

PS - Click here if you need help with your outreach on LinkedIn : https://lnkd.in/esFSh9wh

I cold DM’ed an IT recruiter on Monday.

By Wednesday he wired me $2,000.

Here are 3 simple B2B companies can do the same : 

(Not guaranteed — but this is working insanely well on LinkedIn.)

⸻

1. Volume x Luck = Money
You never know who’s ready to buy today.
This recruiter had a problem he needed fixed now.

But I wouldn’t have found him without volume.
More Outreach → more shots → more Sales Meetings → more $$.

Just do more DM’s.

⸻

2. Sales Nav + InMail = low-key cheat code
If you’re pitching directly, use InMail.

They get seen.
They get replies.
They get Leads.

InMails just convert better. Don’t overthink it.

⸻

3. Follow up like you actually care
My follow-up closed this deal.
Without it? Ghosted.

Follow up > First message.
Every. Single. Time.

⸻

If you want more clients, stop looking for hacks.

Do more Outreach, use the tools that get seen, and follow up like it matters.

Keep working, 
You’ve got this.
I’m rooting for you.

Joshua Debayo - 

The last 2 weeks I’ve had so many DMs asking:

“What are you doing bro, how do you get so many calls?”

I’m a small fish in this thing.

Spoke to Josiah .A  .A the other week…
He booked 30 calls in 24 hours 😭

I’m here thinking I’m doing decent booking 5–7 calls a week.

Here’s 3 simple ways to start booking more calls with real intent prospects on LinkedIn:

(I don’t give any guarantee claims, but it’s working for me personally)

⸻

1. Think outside the box 📦

Everyone’s doing the same sh*t.
Messaging the same way.
Trying to find “the winning script”.

I remember listening to William Brown & he mentioned doing the complete opposite of what everyone’s doing.

You’ll be in your own lane.

⸻

2. Speak their language

I got a response from a cold outreach saying:

“Let’s book a call & would like to know the immediate next steps.”

You only get messages like this if you know their world and can speak their terminology.

⸻

3. Follow ups

I think I mention this every day.
98% of people don’t follow up.
They forget. Can’t be bothered, etc.

All the revenue I’ve generated this year…
90% has come through follow ups.

Whether that’s a follow up on LinkedIn,
or a follow up call.

The real ones know — the $$ is in the follow ups.

⸻

Keep working,
You’ve got this.
I’m rooting for you. 

PS - Might relaunch my community in January to help some people out, DM me if you might be interested.

graphical user interface, text, application

If your LinkedIn posts are flopping, read this.

Last week my impressions jumped by 255.5%.

It generated me 4 inbound leads and 2 new clients — all from content that people want to see.

And now after posting consistently on LinkedIn for over 2 years, I'm slowing understanding how the algorithm actually works.

Here are 3 stupid-simple tweaks I used to turn dead LinkedIn posts into leads, sales meetings and most importantly high paying clients.

(More impressions = more eyeballs = more leads, more sales meetings, more DM’s — without being chained to LinkedIn or outreach 24/7).

1. Don’t sell in your posts.

Is quite obvious LinkedIn doesn’t like you selling in your posts.

If your post looks salesy, they ain't going to push it to people in your ICP.

When I sell, I do it through case studies.
Whether its my case studies or my clients.

Be strategic.

2. Educate your audience

Every post should teach something.
Give them 1 thing they can take today and apply to their business.

Most people will read your post

 if it’s genuinely useful.

3. Post at the same time

I didn’t believe this at first.
But posting at the same time consistently has helped my posts a lot.

Do these 3 things and watch your impressions go crazy 📈

Keep working,
You’ve got this.
I’m rooting for you.

PS – Drop in the comments anything else that helps boost impressions.

Last month my inbox was full of:
“OOO” and “not interested” and “leave me alone”.

Now we’re averaging 2 positive replies per day from cold outreach on Instantly.ai  

Here are 3 tweaks that’ll help you get more positive replies from your outreach (works for both email & LinkedIn ):

⸻

1. Stop asking for a call in the first message

I’m testing 2 campaigns right now:

Campaign A: “Would an intro call makes sense?”
Campaign B: “Want a bit more info first?”

Both are performing…
But WAY more people say “yes” to more info than to a call.

Most people don’t want to hop on a random call with a stranger.
Give value first. Earn the right to ask for a call later.

⸻

2. Follow up like a pro

This is the easiest way to get more positive replies.
Most of our good responses are from follow ups.

And because 90% of agencies never follow up properly, it’s easier for you to win.

Simple structure:

- Day 0: Initial message
- Day 2: Follow up 1
- Day 4: Follow up 2

If they’re not interested after that, move on.
You’re after leads, not validation.

⸻

3. Speak their language

You want them reading your message thinking:
“Okay… this person actually gets it.”

Use industry-specific words.
Mention the exact problems they’re dealing with.
Show you understand their world better than a generic “We help businesses grow” line.

I’ll give 14 agencies owners my simple follow up scripts for FREE.

Like this post
Comment “follow up” 
I’ll send it over. 

Keep working,
You’ve got this.
I’m rooting for you.

I’m giving away a 22-page “State of Recruitment 2026” playbook for FREE.

If you run a recruiting firm, this might change the game for you.

If you’re not building an AI-powered outbound system yet you’re already behind for 2026.

I’ve partnered with Joseph Tiareh, Ed Albazii & David Tiareh and we’ve done something wild:

We spent 6 days creating this guide.

We’re giving the entire playbook away for FREE —
but only to recruitment & staffing firms.

No coaches.
No generic B2B.

⸻

What’s inside the guide 👇🏾

This isn’t a bs ebook.
It’s a breakdown of where the North American recruitment market is heading and how agencies can actually win:

1️⃣ Market reality for 2026
→ U.S. staffing market heading to $183.3B
→ Why growth will come from systems & operations
→ What this means for small / mid-size agencies

2️⃣ Talent shortages & candidate behaviour
→ 69% of employers still struggling to hire
→ 86% of workers open to new roles, but more selective
→ How to message flexibility, progression & culture so candidates respond

3️⃣ The AI shift you can’t ignore
→ 96% of talent teams already using AI
→ Where AI gives 30–50% speed gains
→ How to plug AI into outbound without losing the human touch

4️⃣ Job-signal outreach (the big one)
→ Using live job posts as intent signals
→ Turning “we’re hiring” into “we’re now your client”
→ Simple framework to spot signals & hit decision-makers first

5️⃣ Strategic checklist for agency owners
→ Metrics that actually matter in 2026
→ Positioning your firm as AI-literate, human-led
→ What to focus on first if you run a lean team

⸻

Who this is for

→ Recruitment / staffing firm owners
→ Based in North America
→ Already placing candidates & want more predictable BD + outbound

If you want to:

→ Build an AI-assisted outbound engine
→ Use job signals to find warm accounts
→ Stop relying on random referrals & inbound

This guide will save you months of guessing.

⸻

Want the guide?

Do the following 👇🏾

👍 Like this post
💬 Comment “STATE”
👤 Follow me

I’ll send you the State of Recruitment 2026 PDF + a short video on how I’d use it as an agency owner.

I’m only sending this to 265 recruiting firms.
After that, it’s gone.

Keep working,
You’ve got this,
I’m rooting for you.

Most people blamed Thanksgiving for a slow week. 

We still ended up booking 7 calls on LinkedIn. 

Steal my video guide for FREE‼️

✔️ 6 wealth management firms.
✔️ 2 recruiting firms.

All doing $1M+ in ARR. 

Last night I recorded a free training showing exactly how we do this.

My brother Jeremiah Debayo  is the genius behind the outbound.

I haven’t touched outreach since June. 
He manages everything. 

I do the content.
He fills my calendar.

I probably shouldn’t…
but I’m giving the whole thing away for FREE.

Inside the training you’ll get:

→ How to find qualified leads on SalesNav.ai  
→ How to leverage InMails to message prospects
→ How to never run out of leads to reach out to

All for FREE.

Want the training tutorial?

You must do the following:

👍 Like this post
👤 Follow me
💬 Comment “video”

I’ll send it over.

PS – Only sending this to 78 people.

It took me 1,453 messages (including follow ups) to sign one client last week on LinkedIn  

1453 messages sent.
31 positive replies. 
5 booked calls. 
1 client closed. 

You’re not doing enough.
Do more. 

The only metric that really matters: how many reach-outs it takes to sign a client.

Here are 3 simple ways to land a LinkedIn client this week:

1. My profile is set up to do the initial selling

Most people do Outreach with a CV-looking profile.
Then wonder why nobody answers their DM’s.

My LinkedIn profile is built like a landing page:

- Who I help
- The problem I solve
- The system I use
- Proof I can do it

Mostly importantly, a video in my featured section breaking down my whole process from A-Z.

By the time they reply, they already know why we’re talking.

2. I send stupidly simple messages

No essays.
No “jump on a quick call?” in the first line.

I focus on:

One clear problem they have
One clear way I fix it
One tiny next step ( “Want the breakdown?” )

That’s it.
Short, casual, specific.

That’s how I turn Leads into sales meetings on LinkedIn every week.

3. I follow up like a machine

Most of my positive replies come from follow ups.
People are busy.

They miss messages.
They don’t hate you.

I follow up 2–3 times with:

A quick bump
A piece of Content
A relevant result

Then I move on.

Keep working.
You’ve got this.
 I’m rooting for you

I used to think posting on LinkedIn was cringe.

Now it funds for my lifestyle.

At first they find it cringe.
1 year later they'll ask how you've done it.

I tend to get messages like this more often now.

The other day one of my clients recommended me to his friend.

You know what he sent over first?

Not my website.
Not my number.
Not my email.

He sent… my LinkedIn profile.

That’s when it clicked for me:

If it looks “cringe” to some people, who cares?

They’ll forget your post in 10 seconds.

But that one piece of content could bring you leads & more $$.

If you're in B2B, LinkedIn is your storefront.
(Make sure you go all in before 2026 FYI). 

Your outreach lands better when your profile already tells your story.

Your DM’s feel warmer when people have seen your face and your journey.

AI is taking over a lot of things.
But it can’t take over you and your personal story.

Keep working,
You've got this.
I'm rooting for you.

PS - Tag someone you know that needs to hear this ‼️

My previous post has gone semi-viral so let me explain what I do if you're brand new here:

I run a middleman agency.

I connect recruitment firms and wealth management firms with decision-makers.

For recruiters I connect them with HR managers who are actively looking for new candidates.

I recorded a video recently explaining how we helped a healthcare agency win a contract with a hospital.

Click here to watch -https://lnkd.in/eJXSAW-5

From that one contract:

→ He’ll place 4 candidates
→ Probably make around $30k from it

All from simple outbound messaging on Instantly.ai.

For wealth management firms →

I connect them with high-net-worth individuals/businesses (basically people with a lot of money) who need an financial advisor to help them invest it properly.

A few things I don’t do:

→ I don’t book calls for my clients
→ I don’t offer any guarantees
→ I don't make false promises

We just:

 → Get attention
 → Start conversations
 → Connect the right people

They handle the rest.

I also have a 1–1 consulting offer.

This is far from my main priority, but I know it can help some people.

Keep working,
You’ve got this,
I’m rooting for you.

➔ Click here to apply to partner with us - https://lnkd.in/eS6E_-wS

➔ If you want my help scaling your agency on LinkedIn, book a call here → https://lnkd.in/esFSh9wh

We don’t book calls for our clients. 

We just make leads raise their hand 🙋🏾‍♂️

Let me break this down. 

Our job is very simple but effective:

Get a lead to reply with messages like👇🏾

“Yup, I’m interested. Tell me more.”
“Send me more information.”

Getting an initial response from a DM is pretty simple once you dial in your outreach.

The harder part is this:

Turning that reply into an actual meeting.

We don’t try to be the hero.
We’ve given that responsibility back to our clients.

Here’s why: 

✔️ They know their offer inside out
✔️ They can qualify way better than any script
✔️ They can articulate their services with nuance
✔️ Their close rate is always higher when they’re in the convo

A lead comes through from email.

We warm it up.
They handle the rest.

It makes our life easier.

But more importantly:

Clients are happier, more involved, and way more willing to follow up because they’ve spoken to the prospect themselves.

Sometimes the best “system” is:

We get attention.
They do the rest.

Keep working,
You’ve got this,
I’m rooting for you.

“Yes! I’d love to talk to candidates.”

Another interested lead for our healthcare recruitment client.

A simple cold message we rinse & repeat every single day.

No AI.
No personalisation.

I try to be as transparent as possible with my posts without sharing client details, of course.

We got this response yesterday from the email campaigns we’ve been running on Instantly.ai

You’d be amazed at how many replies we get on Sundays.

People turn their campaigns off over the weekend.
We do the complete opposite & double down.

Campaign stats so far for this particular client 👇🏾

✔️ 2,021 messages sent
✔️ 57 replies
✔️ 13 interested leads
✔️ 6 booked calls

I only started doubling down on email outreach in October.

Before that, my plan was simple 👇🏾

Go all in on LinkedIn  for 2 years…
Then layer in email to book more sales meetings for clients.

Here’s what I’ve learned — and it’s just 1 very simple thing:

→ Email works.
→ LinkedIn works.
→ Every platform works.

The reason it’s not working for you is because you don’t give it enough time to actually see results.

Simple as that.

Give it time.
Keep testing.

You’ll eventually craft a message that gets a stranger’s attention.

Once you do, double down. 📈

“You’re literally one good message away from getting in contact with anyone in the world”.

Remember that🫴🏾

Keep working,
You’ve got this,
I’m rooting for you.

PS - If you run a recruiting or wealth management firm, click here → https://lnkd.in/eCPAgwAB 👈🏾

PSS - I’ll be in my comments section for about 20 minutes, hit me up if you have any questions.

Big Annoucment coming your way‼️

We’re taking this to another level. 

I’ve partnered with Ed Albazii Joseph Tiareh David Tiareh & we’re handing the biggest ever recruitment playbook.

➡️ Monday 1st December 
➡️ 9am EST
➡️ 2pm UK.

Only 298 agencies will get access to this. 

This is going to probably the biggest playbook we’ve ever given away for free.

If you thought my other playbooks were good, this is some next level stuffs 

It’s a complete cheat code to scale before the end of the year.

This will specifically be for recruitment firms.

Whether you need candidates.
Or you’ll looking to win new contracts.

We’ve got you. 

Stay tuned. 
Put it on your calendar.
This is ganna go crazy. 

Keep working
You’ve got this.
I’m rooting for you.

Men lie.
Women lie.

Numbers don’t.

Not everyone is transparent with their clients numbers & metrics online.

This is an outbound campaign I’ve been running on Instantly.ai for the past 2 weeks for our recruiting clients.

I always try to be as transparent as possible without sharing client information.

As you can see, this isn’t a perfect campaign.

The first campaign is solid.
The second & third can definitely improve.
There’s no such thing as perfect in outbound.

With outreach, there’s so much testing & volume you need to do to make it work.

Most agencies quit after the first campaign.

They never get to the part where they’re actually signing clients consistently.

I’ve seen winning campaigns only start hitting after campaign 12.

Test. Learn. Improve. Execute. 

Keep working,
You’ve got this.
I’m rooting for you.

PS – If you run a wealth management or recruiting firm, apply here to partner with us → https://lnkd.in/eS6E_-wS

taplio

The Battle

Alex Xu

Joshua Debayo

Followers

Posts

Engagements

Likes

Performance ScoreAlex Xu's LinkedIn Performance

Performance ScoreJoshua Debayo's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Alex Xu

Joshua Debayo

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Alex Xu's LinkedIn Performance

Performance ScoreJoshua Debayo's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreAlex Xu's LinkedIn Performance