Discord engineering goes into details of how they migrated from Cassandra to ScyllaDB, improved the performance of their reads and writes and rearchitected their backend to support the new load. It is an interesting episode lets get into it
0:00 Intro 1:50 Relational vs Distributed 7:00 The Cassandra Troubles 11:00 SnowFlake vs UUID 14:30 B+Tree 19:20 B+Tree and SSDs 25:30 LSM Trees 31:00 Hot partitions 36:00 Cassandra Garbage Collector Pauses 40:00 Changing the Architecture 45:00 The Data Services 55:00 The Migration 1:02:00 Zoned Named Spaces 1:04:00 Summary
Article here How Discord Stores Trillions of Messages
How Discord Stores Trillions of Messages | Deep DiveHussein Nasser2023-03-11 | Fundamentals of Database Engineering udemy course (link redirects to udemy with coupon) database.husseinnasser.com
Discord engineering goes into details of how they migrated from Cassandra to ScyllaDB, improved the performance of their reads and writes and rearchitected their backend to support the new load. It is an interesting episode lets get into it
0:00 Intro 1:50 Relational vs Distributed 7:00 The Cassandra Troubles 11:00 SnowFlake vs UUID 14:30 B+Tree 19:20 B+Tree and SSDs 25:30 LSM Trees 31:00 Hot partitions 36:00 Cassandra Garbage Collector Pauses 40:00 Changing the Architecture 45:00 The Data Services 55:00 The Migration 1:02:00 Zoned Named Spaces 1:04:00 Summary
Article here How Discord Stores Trillions of Messages
Stay Awesome, HusseinMemcached brilliant solution to fragmentationHussein Nasser2024-01-31 | ...The Cost of Memory FragmentationHussein Nasser2024-01-30 | Fragmentation is a very interesting topic to me, especially when it comes to memory.
While virtually memory does solve external fragmentation (you can still allocate logically contiguous memory in non-contiguous physical memory) it does however introduce performance delays as we jump all over the physical memory to read what appears to us for example as contiguous array in virtual memory.
You see, DDR RAM consists of banks, rows and columns. Each row has around 1024 columns and each column has 64 bits which makes a row around 8kib. The cost of accessing the RAM is the cost of “opening” a row and all its columns (around 50-100 ns) once the row is opened all the columns are opened and the 8 kib is cached in the row buffer in the RAM.
The CPU can ask for an address and transfer 64 bytes at a time (called bursts) so if the CPU (or the MMU to be exact) asks for the next 64 bytes next to it, it comes at no cost because the entire row is cached in the RAM. However if the CPU sends a different address in a different row the old row must be closed and a new row should be opened taking an additional 50 ns hit. So spatial access of bytes ensures efficiency,
So fragmentation does hurt performance if the data you are accessing are not contiguous in physical memory (of course it doesn’t matter if it is contiguous in virtual memory). This kind of remind me of the old days of HDD and how the disk needle physically travels across the disk to read one file which prompted the need of “defragmentation” , although RAM access (and SSD NAND for that matter) isn’t as bad.
Moreover, virtual memory introduces internal fragmentation because of the use of fixed-size blocks (called pages and often 4kib in size), and those are mapped to frames in physical memory.
So if you want to allocate a 32bit integer (4 bytes) you get a 4 kib worth of memory, leaving a whopping 4092 allocated for the process but unused, which cannot be used by the OS. These little pockets of memory can add up as many processes. Another reason developers should take care when allocating memory for efficiency.
0:00 Intro 2:40 Memory Allocation 4:10 External fragmentation 9:00 Internal Fragmentation 11:30 Virtual Memory & Swap 15:20 Page Size 18:00 The genius of memcached 19:50 How CPU reads data from RAM 25:00 The cost of fragmentation in memory 29:00 MySQL 8.x regression 34:00 Summary
Stay Awesome, HusseinHow a Machine Becomes a Router | ip_forwardHussein Nasser2024-01-24 | Recently I learned how the Linux option net.ipv4.ip_forward can turn your machine into a router. I have been using this option in past when working with iptables but never paid attention to how it works. I explore this here.
You see, when your NIC receives a frame from the network with a MAC address the frame is copied to the kernel if the MAC address matches the NIC’s. If doesn’t match the NIC’s MAC address that frame is often dropped by the NIC.
Putting the NIC into promiscuous mode allows all frames in to the OS and the kernel does the filtering instead. This is useful if you have many virtual machines/containers with different MAC addresses exposed under the same physical NIC.
So similarly if the MAC address matches the NIC’s but the IP address in the IP packet doesn’t match machine’s IP, the packet is often dropped by the OS, unless the ip_forward option is enabled.
This essentially tells the OS, hey, you might receive packets that are not for you, please just forward them back through the network and apply any routing rules you might have. Here
This is exactly how a router works, it received tons of packets where the MAC matches it, but almost none of them are destined to the router (except if you want to go to the admin page on your router that is).
Coincidentally, this is also how a firewall works. You configure your hosts to forward all packets to machine F, such that the destinations remain as is but the frame becomes F, the firewall machine receives all packets, apply the filtering rules if those pass it forward the packet back to the network (thanks to the ip_forward) else if the firewall rules didn’t match, it swallows the packet.
This feels like a puzzle piece I have been holding on for a long time and finally fits perfectly.
0:00 Intro 2:00 Layer 2 and Layer 3 Address match 4:00 Layer 2 MAC doesn’t match 5:30 Promiscuous Mode 6:30 Layer 2 Match Layer 3 Address doesn’t 9:00 Acting Like a Router 11:30 Acting Like a Firewall
Stay Awesome, HusseinThe Real Hidden Cost of a RequestHussein Nasser2023-12-13 | In this video I explore the hidden costs of sending a request from the frontend to the backend
Stay Awesome, HusseinBefore you move to HTTP/3 ...Hussein Nasser2023-10-06 | HTTP/3 is getting popular in the cloud scene but before you migrate to HTTP/3 consider its cost. I explore it here.
0:00 Intro HTTP/3 is getting popular 3:40 HTTP/1.1 Cost 5:18 HTTP/2 Cost 6:30 HTTP/3 Cost
Stay Awesome, HusseinDeep dive on how static files are served with HTTP (kernel, sockets, file system, memory, zero copy)Hussein Nasser2023-10-05 | In this video I do a deep dive on how serving static files work in web servers.
0:00 Intro 2:00 Overview 3:00 Request handling and Receive Queue 8:50 Reading file from disk 13:50 Response and the Send Queue 24:00 Sending Response to the Client
Stay Awesome, HusseinThe Pros and Cons of Encrypted Client HelloHussein Nasser2023-09-30 | Discovering Backend Bottlenecks: Unlocking Peak Performance performance.husseinnasser.com
The Encrypted Client Hello or ECH is a new RFC that encrypts the TLS client hello to hide sensitive information like the SNI. In this video I go through pros and cons of this new rfc.
Stay Awesome, HusseinThe effect of Random UUID on database performanceHussein Nasser2023-08-26 | In this video I whiteboard how UUIDs hurt write (and read performance) when used on secondary and primary indexes. UUIDv4 are the most popular but they are truly random. Compared to snowflakes, ULID or UUIDv7 or even UUIDv1.
0:00 UUIDv4 2:30 B+Tree Indexes and UUID 5:30 Random UUIDv4 Insert Workload 12:40 Ordered Insert Workload (UUID7/ULID, Sequence) 14:00 Shared buffer pool flushes 15:00 Shopify ULID use case 17:00 URL shortner UUIDs?
Stay Awesome, Husseinhow greedy are your processes? | Linux nicenessHussein Nasser2023-08-15 | Discovering Backend Bottlenecks: Unlocking Peak Performance performance.husseinnasser.com
In this video I explain linux Nice processes and how to change the priority of a process.
In the Linux operating system, processes manage and execute applications and tasks. One of the key utilities for managing processes on Linux is the `nice` command, which adjusts the priority of a process. Every process on Linux is assigned a 'niceness' value, which ranges from -20 (highest priority) to 19 (lowest priority). The default niceness value for processes is 0. By adjusting this value, users can influence the scheduler's decisions, giving more CPU time to crucial tasks and less to background jobs. The related command, `renice`, allows users to modify the priority of an already running process. Both utilities, when used correctly, provide users with fine-grained control over system resources, ensuring that the system remains responsive and efficient in multitasking environments.
Fundamentals of Backend Engineering Design patterns udemy course (link redirects to udemy with coupon) backend.husseinnasser.com
Fundamentals of Networking for Effective Backends udemy course (link redirects to udemy with coupon) network.husseinnasser.com
Fundamentals of Database Engineering udemy course (link redirects to udemy with coupon) database.husseinnasser.com
Stay Awesome, HusseinHow CPU Efficient is your App?Hussein Nasser2023-08-12 | Identifying whether an application is CPU or IO bound is important for scaling, we explore this using the top command in Linux in this video.
0:00 Intro 2:30 us sy wa id 10:00 CPU Bound 12:20 Pressure Stall Information 14:45 IO Disk Bound
Stay Awesome, HusseinHow does the Kernel manage backend Connections?Hussein Nasser2023-08-07 | This is a question from my Discovering Backend Bottlenecks: Unlocking Peak Performance course, on how does the kernel manages TCP connections on the backend. We explain the ACCEPT and SYN queue.
Stay Awesome, HusseinWhat happens before the Backend gets the RequestHussein Nasser2023-08-01 | When we send a request to a backend most of us focus on the processing aspect of the request which is really just the last step. There is so much more happening before a request is ready to be processed, most of this step happens in the Kernel. I break this into 6 steps, each step can theoretically be executed by a dedicated thread or process. Pretty much all backends, web servers, proxies, frameworks and even databases have to do all these steps and they all do choose to do it differently.
Stay Awesome, HusseinThey Enabled Postgres Partitioning and their Backend fell apartHussein Nasser2023-06-24 | In a wonderful blog, Kyle explores the pains he faced managing a Postgres instance for a startup he works for and how enabling partitioning sigintfically created wait events causing the backend and subsequently NGINX to through 500 errors.
0:00 Intro 3:30 Background 6:30 April 1st The partitioning 8:30 May 6th things went south 10:30 April 11 The start of the waits 12:00 Locks 15:30 High volume of locks 17:30 The Problem with Postgres 13 20:00 Fastpath locking 24:00 Moving to Weekly Partitions 31:30 Summary
The performance of a backend application goes beyond its code and logic. Factors such as network latency, Kernel TCP/IP stack, security overhead and protocol serialization. All can impact the performance.
In my new course I explore how to identify bottlenecks in applications whether it is in the frontend, the backend, the intermediary proxy or even the database. I use real-world examples that I personally ran into in my 19 year career as a software engineer.
Stay Awesome, HusseinIs gRPC Web Native?Hussein Nasser2023-06-09 | ...WebSockets is not freeHussein Nasser2023-06-09 | ...WebTransport is a Game Changer ProtocolHussein Nasser2023-06-08 | backend.husseinnasser.com Fundamentals of Backend Engineering Design patterns udemy course (link redirects to udemy with coupon)
WebTransport is a cutting-edge protocol framework designed to support multiplexed and secure transport over HTTP/2 and HTTP/3. It brings together the best of web and transport technologies, providing an all-in-one solution for real-time, bidirectional communication on the web.
WebTransport can handle different types of data simultaneously over a single connection, thanks to its multiplexing capabilities. This includes both reliable and unreliable transmissions, which allows it to cater to a wide variety of applications, from those demanding utmost data reliability to those requiring ultra-fast, real-time interactions.
With its inherent support for low-latency connections, WebTransport proves a boon for real-time applications. It also offers improved client-to-server communication, breaking free from the traditional model where bidirectional streams are primarily client-initiated.
Furthermore, every WebTransport connection is secured with TLS, offering a high level of data security that's in line with modern web standards.
Being compatible with HTTP/2 and HTTP/3, WebTransport is built to leverage the widespread deployment and proven capabilities of these protocols, while paving the way for a more dynamic and interactive web. It represents a significant step forward in the evolution of backend protocols, promising to revolutionize the way we handle web communications.
In this video I cover the WebTransport protocol
0:00 Intro 0:48 The Simple Web 2:30 WebSockets Limitations 5:00 HTTP/2 & HTTP/3 6:40 gRPC Limitations 8:40 WebTransport 9:30 Unreliable streaming 11:49 WebTransport over HTTP/2 and HTTP/3 12:55 Proxying WebTransport 13:30 Summary
Stay Awesome, HusseinYour SSD lies but thats ok .. I think | Postgres fsyncHussein Nasser2023-05-25 | fsync is a linux system call that flushes all pages and metadata for a given file to the disk. It is indeed an expensive operation but required for durability especially for database systems. Regular writes that make it to the disk controller are often placed in the SSD local cache to accumulate more writes before getting flushed to the NAND cells. However when the disk controller receives this flush command it is required to immediately persist all of the data to the NAND cells.
Some SSDs however don't do that because they don't trust the host and no-op the fsync, in some cases SSD buffers it for few microseconds to receive more writes and hope it doesn't crash in this microsecond. Others battery powered SSDs can completely ignore host flushes because it can survive a crash. In this video I explain this in details and go through details on how postgres provide so many options to fine tune fsync
0:00 Intro 1:00 A Write doesn’t write 2:00 File System Page Cache 6:00 Fsync 7:30 SSD Cache 9:20 SSD ignores the flush 9:30 15 Year old Firefox fsync bug 12:30 What happens if SSD loses power 15:00 What options does Postgres exposes? 15:30 open_sync (O_SYNC) 16:15 open_datasync (O_DSYNC) 17:10 O_DIRECT 19:00 fsync 20:50 fdatasync 21:13 fsync = off 23:30 Don’t make your API simple 26:00 Database on metal?
Fundamentals of Backend Engineering Design patterns udemy course (link redirects to udemy with coupon) backend.husseinnasser.com
Fundamentals of Networking for Effective Backends udemy course (link redirects to udemy with coupon) network.husseinnasser.com
Fundamentals of Database Engineering udemy course (link redirects to udemy with coupon) database.husseinnasser.com
Stay Awesome, HusseinThe problem with software engineeringHussein Nasser2023-05-20 | ego is the main problem to a defective software product. the ego of the engineer or the tech lead seeps into the quality of the product.
Fundamentals of Backend Engineering Design patterns udemy course (link redirects to udemy with coupon) backend.husseinnasser.com
Fundamentals of Networking for Effective Backends udemy course (link redirects to udemy with coupon) network.husseinnasser.com
Fundamentals of Database Engineering udemy course (link redirects to udemy with coupon) database.husseinnasser.com
Stay Awesome, HusseinClustered Collections makes Mongo faster but there is a costHussein Nasser2023-05-11 | Fundamentals of Database Engineering udemy course (link redirects to udemy with coupon) database.husseinnasser.com
In version 5.3, MongoDB introduced a feature called clustered collection which stores documents in the _id index as oppose to the hidden wiredTiger hidden index. This eliminates an entire b+tree seek for reads using the _id index and also removes the additional write to the hidden index speeding both reads and writes.
However like we know in software engineering, everything has a cost. This feature does come with a few. In this video I discuss the following
0:00 Intro 1:00 How Original Collections Work 6:50 How Clustered Collections Work 10:00 Benefits 17:00 Limitations
Fundamentals of Backend Engineering Design patterns udemy course (link redirects to udemy with coupon) backend.husseinnasser.com
Fundamentals of Networking for Effective Backends udemy course (link redirects to udemy with coupon) network.husseinnasser.com
Stay Awesome, HusseinPrime Video Swaps Microservices for Monolith: 90% Cost ReductionHussein Nasser2023-05-06 | Prime video engineering team has posted a blog detailing how they moved their live stream monitoring service from microservices to a monolith reducing their cost by 90%, let us discuss this
0:00 Intro 2:00 Overview 10:35 Distributed System Overhead 21:30 From Microservices to Monolith 29:00 Scaling the Monolith 32:30 Takeaways
Stay Awesome, HusseinPostgres, JSON and TOAST #databaseHussein Nasser2023-05-02 | ...A Deep Dive in How Slow SELECT * isHussein Nasser2023-05-02 | Fundamentals of Database Engineering udemy course (link redirects to udemy with coupon) database.husseinnasser.com
In a row-store database engine, rows are stored in units called pages. Each page has a fixed header and contains multiple rows, with each row having a record header followed by its respective columns. When the database fetches a page and places it in the shared buffer pool, we gain access to all rows and columns within that page. So, the question arises: if we have all the columns readily available in memory, why would SELECT * be slow and costly? Is it really as slow as people claim it to be? And if so why is it so? In this post, we will explore these questions and more.
0:00 Intro 1:49 Database Page Layout 5:00 How SELECT Works 10:49 No Index-Only Scans 18:00 Deserialization Cost 21:00 Not All Columns are Inline 28:00 Network Cost 36:00 Client Deserialization
Stay Awesome, HusseinAWS Serverless Lambda Supports Response StreamingHussein Nasser2023-04-08 | Lambda now supports Response payload streaming, now you can flush changes to the network socket as soon as it is available and it will be written to the client socket. I think this is a game changing feature
0:00 Intro 1:00 Traditional Lambda 3:00 Server Sent Events & Chunk-Encoding 5:00 What happens to clients? 6:00 Supported Regions 7:00 My thoughts
Stay Awesome, HusseinThe Cloudflare mTLS vulnerability - A Deep Dive AnalysisHussein Nasser2023-04-06 | Cloudflare released a blog detailing a vulnerability that has been in their system for nearly two years. it is related to mTLS or mutual TLS and specifically client certificate revocation. I explore this in details
0:00 Intro 3:00 The Vulnerability 7:00 What happened? 8:50 Certificate Revocation 12:30 Rejecting certain endpoints 17:00 Certificate Authentication 20:30 Certificate serial number 24:00 Session Resumption (PSK) 35:00 The bug 37:00 How they addressed the problem
Stay Awesome, HusseinThe Virgin Media ISP outage - What happened?Hussein Nasser2023-04-06 | BGP (Border gateway protocol) withdrawals caused the Virgin media ISP customers to lose their Internet connection. I go into details on this video.
0:00 Intro 2:00 What happened? 4:11 How BGP works? 11:50 Version media withdrawals 15:00 Deep dive
Fundamentals of Backend Engineering Design patterns udemy course (link redirects to udemy with coupon) backend.husseinnasser.com
Fundamentals of Networking for Effective Backends udemy course (link redirects to udemy with coupon) network.husseinnasser.com
Fundamentals of Database Engineering udemy course (link redirects to udemy with coupon) database.husseinnasser.com
Stay Awesome, HusseinGitHub SSH key is Leaked - How bad is this?Hussein Nasser2023-03-30 | GitHub Accidentally Exposed their SSH RSA Private key, this is the message you will get .
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ @ WARNING: REMOTE HOST IDENTIFICATION HAS CHANGED! @ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ IT IS POSSIBLE THAT SOMEONE IS DOING SOMETHING NASTY! Someone could be eavesdropping on you right now (man-in-the-middle attack)! It is also possible that a host key has just been changed. The fingerprint for the RSA key sent by the remote host is SHA256:uNiVztksCsDhcc0u9e8BujQXVUpKZIDTMczCvj3tD2s. Please contact your system administrator. Add correct host key in ~/.ssh/known_hosts to get rid of this message. Host key for github.com has changed and you have requested strict checking. Host key verification failed.
In this video I discuss how bad is this,.
0:00 Intro 1:10 What happened? 3:00 SSH vs TLS Authentication 6:00 SSH Connect 7:45 How bad is the github leak? 15:00 What should you do? 18:50 Is ECDSA immune?
Stay Awesome, HusseinHow Linus Tech Tips got hackedHussein Nasser2023-03-24 | In this short video we explain how was it possible for Linux to get hacked with cookies hijacking.
0:00 Intro 0:47 TLDR what happened 5:10 Cookies in Chrome 7:30 Cookies Hijacking 8:46 Session Tokens (Access/Refresh) 10:00 Remedies
Stay Awesome, HusseinAll Postgres Locks Explained | A Deep DiveHussein Nasser2023-03-19 | Fundamentals of Database Engineering udemy course (link redirects to udemy with coupon) database.husseinnasser.com
In this video I do a deep dive in all locks obtained by postgres, I learned a lot while making this video and hope you enjoy it.
0:00 Intro 2:30 What are Locks? 5:30 Overview of Postgres Locks 9:10 Table-Level Locks 11:40 ACCESS EXCLUSIVE 17:40 ACCESS SHARE 19:00 ROW SHARE20:15 ROW EXCLUSIVE 21:15 SHARE UPDATE EXCLUSIVE 23:30 SHARE 24:50 SHARE ROW EXCLUSIVE 25:18 EXCLUSIVE 25:30 Table Lock Conflict Matrix 28:30 Row-Level Locks 30:00 FOR UPDATE 33:00 FOR NO KEY UPDATE 34:00 FOR SHARE 34:40 FOR KEY SHARE35:10 Row Lock Conflict Matrix 39:25 Page-Level Locks 42:00 Deadlocks 46:00 Advisory Locks 47:20 Summary
Stay Awesome, HusseinPinterest moves to HTTP/3Hussein Nasser2023-03-16 | Pinterest moves to HTTP/3 on all their clients and edge CDNs this year. They witnessed interesting gains but not without good lesson learned. The main one was the mismatch of alt-svc vs DNS ttls.
I cover this on the next episode of the backend engineering course. 0:00 Intro 2:00 Moving h2 to h3 through alt-svc 5:00 Why HTTP/3 6:00 HTTP/1 vs HTTP/2 9:00 TCP Head of Line blocking in HTTP/2 11:00 How HTTP/3 addresses HOL 12:15 Connection Migration 13:30 Stream level congestion control 14:10 1-RTT - 0-RTT 15:41 Pinterest challenges moving HTTP/3 19:00 Migration 21:15 Future work 22:30 Summary
Stay Awesome, HusseinNodeJS Architecture - I/OHussein Nasser2023-03-14 | Node is a non-blocking Javascript runtime. You can concurrently run an HTTP Server, read files from disk, send UDP datagrams, accept TCP connections from clients and still have room to execute Javascript code operations without blocking. Most of these operations are known as I/O, you send an input to a device, file or a socket and it replies back with an output. Node achieves non-blocking I/O with mostly a single thread executed asynchronously using a library called lib_uv. The exception being DNS queries, which use a thread pool. This means when you use fetch or axios to make an HTTP request to a domain, the DNS resolution for that domain will most probably go through the thread pool, while the actual request itself will be sent asynchronously on the main thread.
In this episode of the backend engineering show I go through an article I wrote discussing NodeJS Asynchronous I/O.
0:00 Intro 3:00 Part 1 Socket/IO 9:48 Part 2 File I/O 12:42 Part 3 DNS 16:22 Part 4 NodeJS Single Threaded 19:10 Part 5 NodeJS Thread Pool 21:23 Part 6 DNS lookup bottleneck in Node
Fundamentals of Networking for Effective Backends udemy course (link redirects to udemy with coupon) network.husseinnasser.com
Fundamentals of Database Engineering udemy course (link redirects to udemy with coupon) database.husseinnasser.com
Stay Awesome, HusseinHow Loom Leaked Session CookiesHussein Nasser2023-03-14 | On March 7 2023, Loom users started seeing each others data as a result of cookies getting leaked from the CDN. This loom security breach is really critical. Let us discuss
0:00 Intro 1:00 Why Cookies 2:00 How this happens 5:50 What caused it? 7:30 How Loom solved it? 8:20 Reading the RCA 10:30 Remedies
Stay Awesome, HusseinA Postgres Page can fit 291 rows on a single column tableHussein Nasser2023-02-21 | ...Postgres System Columns Explained (ctid, xmin,xmax)Hussein Nasser2023-02-20 | In this video I explain the difference between ctid,xmin and xmax columns in postgres with repeatable read and read committed.
0:00 Intro 2:00 ctid tuple id 7:00 xmin (minimum transaction id) with read committed 8:00 xmin with read committed 13:00 xmin with repeatable read 19:00 xmax (maximum transaction id)
Fundamentals of Backend Engineering Design patterns udemy course (link redirects to udemy with coupon) backend.husseinnasser.com
Fundamentals of Networking for Effective Backends udemy course (link redirects to udemy with coupon) network.husseinnasser.com
Fundamentals of Database Engineering udemy course (link redirects to udemy with coupon) database.husseinnasser.com
Stay Awesome, HusseinPostgres Internal Architecture ExplainedHussein Nasser2023-02-16 | Creating a listener on the backend application that accepts connections is simple. You listen on an address-port pair, connection attempts to that address and port will get added to an accept queue; The application accepts connections from the queue and start reading the data stream sent on the connection.
However, what part of your application does the accepting and what part does the reading and what part does the execution? You can architect your application in many ways based on your use cases. I have a medium post just exploring the different options. In this video I explore the PostgreSQL process architecture in details. Please note that the information here is derived from both the Postgres doc and code. Discussions about scalability and performance are solely based on my opinions.
Stay Awesome, HusseinHow Alt-Svc switches HTTP/2 clients to use HTTP/3Hussein Nasser2023-02-13 | The Alt-Svc header/frame is a capability that allows the server to adverse alternative services to the connected application available in protocols, ports or domains. It is available as a response header alt-svc and also as an HTTP/2 frame. Let us discuss this capability.
0:00 Intro 1:38 what is alt-svc? 5:30 uses of h3 in alt-svc 8:00 alt-svc header 10:00 Alt-svc header with 103 early hints 14:48 h2 altsvc frame 18:30 SVCB DNS record 21:20 Summary
Fundamentals of Backend Engineering Design patterns udemy course (link redirects to udemy with coupon) backend.husseinnasser.com
Fundamentals of Networking for Effective Backends udemy course (link redirects to udemy with coupon) network.husseinnasser.com
Fundamentals of Database Engineering udemy course (link redirects to udemy with coupon) database.husseinnasser.com
Stay Awesome, HusseinLower and Upper ASCII differs by a single bit (0x20)Hussein Nasser2023-02-02 | ...yOuR DNS qUeRies wILl eVenTualLy lOok lIKe ThIs (0x20 DNS encoding)Hussein Nasser2023-01-28 | Correction: Google is implementing the proposal originally submitted by researchers from Georgia institute of tech. I incorrectly said in the video that google is proposing this .
Google is finally implementing a proposal from 2008 by researchers from Georgia institute of technology to make DNS cache poisoning .
0:00 Intro 2:00 How DNS Work 5:00 DNS Cache Poisoning 14:00 gOoGLe dot CoM 16:20 ASCII 0x20 casing 18:30 Randomizing the casing with encryption 22:30 limitations of this proposal 24:00 Credits
Fundamentals of Backend Engineering Design patterns udemy course (link redirects to udemy with coupon) backend.husseinnasser.com
Fundamentals of Networking for Effective Backends udemy course (link redirects to udemy with coupon) network.husseinnasser.com
Fundamentals of Database Engineering udemy course (link redirects to udemy with coupon) database.husseinnasser.com
Stay Awesome, HusseinDropBox Removed their SSDs, got 20% faster writesHussein Nasser2023-01-24 | In this episode of the backend engineering show I’ll discuss how Dropbox improved their write through put by 20% by removing all their SSDs (yes I was surprised too). DropBox uses an SSD layer as a write-back cache with SMR drives as their backend persistent storage. They changed their model to write directly to the hard drives.
0:00 Intro 2:00 Article Summary 3:00 SMR Drives 6:00 SSD Cache & WriteBack 8:00 Replacing Cache 9:30 Storage Engine Background 14:30 Why did they do it 15:00 The limitation of SSDs & Zoned Namespaces 19:30 Updating the Storage Engine 22:30 Tradeoffs 26:00 Rollout 28:00 Summary
Stay Awesome, HusseinAnalyzing ChatGPT BackendHussein Nasser2023-01-21 | ChatGPT is a chatbot launched by OpenAI in November 2022. It is built on top of OpenAI's GPT-3 family of large language models. In this video I use devtool to analyze backend api of chatGPT, and go through the following topics:
0:00 Intro 1:00 Initial Connection 4:45 Protocols (HTTP/2, HTTP/3) 8:50 Authentication 10:00 Conversations, Paging & UUID 14:50 AI Models (text-davinci-002-render) 16:45 Messages 19:00 Server Sent Events 23:00 Moderating Messages & UUID Discussion 26:40 Generating Conversation Titles 28:00 SSE without EventSource 33:00 Feedback API 34:00 Limitations of UUID Messages on Databases 37:00 How ULID Improves reads/writes 38:00 Two ChatGPT talking to Each other
Stay Awesome, HusseinHow the first version of MongoDBHussein Nasser2023-01-10 | ...How MongoDB stores documentsHussein Nasser2023-01-09 | ...How Fast is MySQL on HTTP/3?Hussein Nasser2023-01-05 | The communication between backend applications and database systems always fascinated me. The protocols keep evolving and we are in constant search for an efficient protocol that best fit the workload of Backend-DB communication.
In this episode of the backend engineering show I go through a blog written by @PlanetScale doing an experimentation of using HTTP/3 and HTTP/2 comparing it with MySQL Binary protocol.
Stay Awesome, HusseinHow Shopify’s engineering improved database writes by 50% with ULIDHussein Nasser2022-12-23 | Fundamentals of Database Engineering udemy course (link redirects to udemy with coupon) database.husseinnasser.com
Shopify posted a blog on tips to for scalable payment system, one tip peeked my interest related to switching from UUID to ULID. I explore the reasoning behind this in this video.
0:00 Intro 1:30 idempotency 6:30 UUID vs ULID 9:50 Clustered Index 13:30 Why UUID4 Inserts are slow 17:15 How ULID helps Shopify 22:00 Problem with tail pages 25:00 Does ULID help in all cases?
Fundamentals of Backend Engineering Design patterns udemy course (link redirects to udemy with coupon) backend.husseinnasser.com
Fundamentals of Networking for Effective Backends udemy course (link redirects to udemy with coupon) network.husseinnasser.com
Fundamentals of Database Engineering udemy course (link redirects to udemy with coupon) database.husseinnasser.com
Stay Awesome, HusseinThe Frontend of a DatabaseHussein Nasser2022-12-19 | ...MongoDB Internal ArchitectureHussein Nasser2022-12-17 | I’m a big believer that database systems share similar core fundamentals at their storage layer and understanding them allows one to compare different DBMS objectively. For example, How documents are stored in MongoDB is no different from how MySQL or PostgreSQL store rows. Everything goes to disk, the trick is to fetch what you need from disk efficiently with as fewer I/Os as possible, the rest is API.
In this video I discuss the evolution of MongoDB internal architecture on how documents are stored and retrieved focusing on the index storage representation. I assume the reader is well versed with fundamentals of database engineering such as indexes, B+Trees, data files, WAL etc, you may pick up my database course to learn the skills. Let us get started.
0:00 Intro 2:00 SQL vs NOSQL 18:00 MongoDB first version MMAPV1 26:30 MongoDB Wired Tiger 38:00 Clustered Collections