Skip to content

Author: Emmanuel Goossaert

Implementing a Key-Value Store – Part 10: High-Performance Networking: KingServer vs. Nginx

This is Part 10 of the IKVS series, “Implementing a Key-Value Store”. You can also check the Table of Contents for other parts. In this series of articles, I describe the research and process through which I am implementing a key-value database, which I have named “KingDB”. The source code is available at http://kingdb.org. Please note that you do not need to read the previous parts to be able to follow what is going on here. The previous parts were mostly exploratory, and starting with Part 8 is perfectly fine.

In this article, I explain the model and inner workings of KingServer, the network server for KingDB. In order to put things into perspective I also cover Nginx, the high-performance HTTP server well-known for its network stack, and how it differs from KingDB.

How to get started with infrastructure and distributed systems

Most of us developers have had experience with web or native applications that run on a single computer, but things are a lot different when you need to build a distributed system to synchronize dozens, sometimes hundreds of computers to work together. I recently received an email from someone asking…

Implementing a Key-Value Store – Part 9: Data Format and Memory Management in KingDB

This is Part 9 of the IKVS series, “Implementing a Key-Value Store”. You can also check the Table of Contents for other parts. In this series of articles, I describe the research and process through which I am implementing a key-value database, which I have named “KingDB”. The source code is available at http://kingdb.org. Please note that you do not need to read the previous parts to be able to follow. The previous parts were mostly exploratory, and starting with Part 8 is perfectly fine.

In this article, I explain how the storage engine of KingDB works, including details about the data format. I also cover how memory management is done through the use of a compaction process.

Implementing a Key-Value Store – Part 8: Architecture of KingDB

This is Part 8 of the IKVS series, “Implementing a Key-Value Store”. You can also check the Table of Contents for other parts. In this series of articles, I describe the research and process through which I am implementing a key-value database, which I have named “KingDB”. The source code is available at http://kingdb.org. Please note that you do not need to read the previous parts to be able to follow. The previous parts were mostly exploratory, and starting with Part 8 is perfectly fine.

In the previous articles, I have laid out the research and discussion around what needs to be considered when implementing a new key-value store. In this article, I will present the architecture of KingDB, the key-value store of this article series that I have finally finished implementing.

An afternoon in 1983 with Hector

ad-small

Back in 1998, I was 13 years old and had no money to buy a computer. But that wouldn’t stop me. I didn’t care if I had a brand new computer or a used one, all I wanted was a computer. From the conversations of adults around me, I heard it was common for businesses to renew their hardware and throw their old computers into the trash. So I figured, all I had to do was to be in the right trash at the right time, or even better, make the trash come to me. In these years, there were tons of computer paper magazines, so I decided to send a letter to one of them, of which I have now forgotten the name, to publish an ad in their classified section: “13 year-old student in the Paris area, will come to your home or office to get any computers you are about to throw in the bin”. From this classified, I got a 486DX-80 PC, a printer, and a Hector HR2+ computer.

The 486DX-80 and the printer have served me well, and are long gone. Last summer, while visiting my family in France I decided to take a look into the attic. As I was making my way through rusty nails and spider webs, I noticed a bag covered with dust in a dark corner of the room. I had the feeling that I was about to make a very good discovery, and I was not disappointed. I opened the bag with excitement, and there it was, the almighty Hector 2HR+ computer!

In the bag came all the booklets, cables, and cassettes, so I decided that I would spend the afternoon writing code on that machine, and that I would run this code before nightfall. But before I go any further on that, a bit of history about the Hector 2HR+…