ByConity

ByConity is a data warehouse designed for changes in modern cloud architecture. It adopts a cloud-native architecture design to meet the requirements of data warehouse users for flexible scaling, separation of reads and writes, resource isolation, and strong data consistency. At the same time, it provides excellent query and write performance.

ByConity is using a large number of mature OLAP technologies, such as column storage engine, MPP execution, intelligent query optimization, vectorized execution, Codegen, indexing, and data compression; it also makes special technological innovations for the cloud scenarios and storage-computing separation architecture.

ByConity is built on top of ClickHouse. We appreciate the excellent work of the ClickHouse team.

Try ByConity

You can quickly bring up a ByConity playground by following this simple guide.

A minimal ByConity cluster include:

A FoundationDB database cluster to store meta data.
A HDFS cluster to store data.
A ByConity server to receive request from clients.
A ByConity read worker to carry execution of read requests forward from server.
A ByConity write worker to carry execution of write requests forward from server.
A ByConity TSO server to provide timestamp
A ByConity daemon manager to manage background jobs that run in server

Build ByConity

The easiest way to build ByConity is built in docker. ByConity executable file depend on Foundation DB library libfdb_c.so. So in order to run it, we need to install the FoundationDB client package. This link tells how to install. We can download client package from FoundationDB GitHub release pages, for example here. Another way to export the LD_LIBRARY_PATH so that the executable can find Foundation DB library libfdb_c.so. But this way is not recommended.

export LD_LIBRARY_PATH="{ByConity_source_path}/contrib/foundationdb/lib/"

It can also be built the following operating systems:

Linux

1. Prepare Prerequisites

The following packages are required:

Git
CMake 3.17 or newer
Ninja
C++ compiler: clang-11 or clang-12
Linker: lld

sudo apt-get update
sudo apt-get install git cmake ccache python3 ninja-build libssl-dev libsnappy-dev apt-transport-https

# install llvm 12
sudo apt install lsb-release wget software-properties-common gnupg # pre-requisites of llvm.sh
wget https://apt.llvm.org/llvm.sh
chmod +x llvm.sh
sudo ./llvm.sh 12

2. Checkout Source Code

git clone --recursive https://github.com/ByConity/ByConity.git

3. Build

cd ByConity
mkdir build && cd build
export CC=clang-12
export CXX=clang++-12
cmake ..
ninja

Then you can find the binary in the programs folder

clickhouse-client    # byconity client
clickhouse-server    # byconity server
clickhouse-worker    # byconity worker
tso_server           # byconity tso
daemon_manager       # byconity daemon manager
resource_manager     # byconity resource manager

Run ByConity Locally

The most convinience way for local development is to use docker-compose. You can use docker-compose to quickly create a byconity cluster from your local build binary. By using this approach, you do not need to worry about the setup of ByConity dependencies (FoundationDB and HDFS), it automatically launches them all. It is recommended to use this approach for ByConity development.

Alternatively, if you don't want to use docker, please follow the belowing guide to run ByConity in non-containerized environments. It assumes you have FoundationDB and HDFS set up and running locally:

Modify the template config
Run the local deployment script to run all the components

Modify the template config

The config templates can be found in deploy/template. You should replace the following in in byconity-server.xml and byconity-worker.xml:

Path_To_FDB with path to your FoundationDB fdb.cluster file path
HOST:PORT with the host and port of your name node in your HDFS cluster

    <catalog_service>
        <type>fdb</type>
        <fdb>
            <cluster_file>/Path_To_FDB/fdb.cluster</cluster_file>
        </fdb>
    </catalog_service>
    ...
    <tso_service>
        <port>49963</port>
        <type>fdb</type>
        <fdb>
            <cluster_file>/Path_To_FDB/fdb.cluster</cluster_file>
        </fdb>
        <tso_window_ms>3000</tso_window_ms>
        <tso_max_retry_count>3</tso_max_retry_count>
    </tso_service>
    ...
    <hdfs_nnproxy>hdfs://HOST:PORT</hdfs_nnproxy>

Run the local deployment script

Make sure you have python3.9 and tmux installed
Install missing libraries if any. For example:
1. pip3.9 install psutils
Run tmux in another terminal
Run the deploy script in a separate terminal. template_paths and program_dir args are compulsory
1. cd ByConity/deploy
2. python3.9 deploy.py --template_paths template/byconity-server.xml template/byconity-worker.xml --program_dir /home/ByConity/build/programs
3. There are other arguments for the script. For example, you can run 2 servers with argument -s 2

Deploy ByConity to physical machines

There are some way to deploy ByConity to physical machines:

Deploy via docker wrapper
Deploy using package manager

FluorineDog/ByConity