Piccadilly
An Event-Driven High-Performance Key-Value Store

This application is a from scratch project to help me (KevinZonda) familiar with distributed systems.

Basic Concept

A ZooKeeper-like service, but aims to provide single instance service with High Performance KV store with Event-Driven Architecture.

PKV supports 2 write models:

Tested on MacBook Air M2 (2022) 8c CPU (4P+4E), 10c GPU, 24GB RAM, 2TB SSD.

PKV benched via direct function call (WAL, GC, Flush disabled). Redis benched via redis-benchmark -q -n 500000

Data Size	Linear	NoLinear	Redis
100,000	WR Time: 1.291293166s WR Perf: 774417.48 RPS RD Time: 162.476083ms RD Perf: 6154752.02 RPS	WR Time: 1.114430042s WR Perf: 897319.67 RPS RD Time: 161.21225ms RD Perf: 6203002.56 RPS	WR Perf (SET): 138159.72 RPS RD Perf (GET): 134716.42 RPS
500,000	WR Time: 7.05466275s WR Perf: 708751.10 RPS RD Time: 973.725708ms RD Perf: 5134916.29 RPS	WR Time: 6.241516166s WR Perf: 801087.41 RPS RD Time: 947.381083ms RD Perf: 5277707.24 RPS	WR Perf (SET): 137931.03 RPS RD Perf (GET): 137438.16 RPS

Name		Name	Last commit message	Last commit date
Latest commit History 77 Commits
KV		KV
client		client
cmd		cmd
doc		doc
example		example
pb		pb
serv		serv
tests		tests
types		types
utils		utils
.gitignore		.gitignore
README.md		README.md
build.sh		build.sh
config.example.json		config.example.json
go.mod		go.mod
go.sum		go.sum
install.sh		install.sh
pb.sh		pb.sh
sync.sh		sync.sh