UCU Software Architecture Project

Features

• Functionality: Deposit money Send money List transactions Login/Sign Up with JWT Token User/General Bank Analytics
• Technologies: Kafka AWS React Python Docker Grafana Prometheus Databricks
• Frameworks: FastAPI Faust
• Databases: AWS Keyspaces MongoDB
• AWS Resources: EKS ELB CloudWatch API Gateway S3 Amplify IAM KMS VPC
• Orchestration: Kubernetes Cortex

Description

unobank is a web banking project created by the students of Ukrainian Catholic University. It allows user to register and top up their accounts, send money based on the card_id of recipient, and see a list of transactions. The project's architecture is heavily decoupled by using seven microservices that interact with each other. We leverage the AWS cloud platform to enable secure, fast, and robust infrastructure.

Project documentation -- https://proximal-bladder-a8d.notion.site/Online-Web-Banking-Project-74a56734638b44c184b3505ad26338d4

The high-level diagram of our services from the infrastructure side looks like this:

The high-level diagram of our services from the interaction side looks like this:

Cassandra Interaction

Microservice	Cassandra Tables
CardManager	[cards, unique_users_daily]
OrchestratorService	[cards, transactions_by_card, reserved_transactions]
CardService	[cards, reserved_transactions]
TransactionService	[transactions, transactions_by_card, successful_transactions_daily, transactions_preaggregated_daily, transactions_preaggregated_monthly]
AnalyticsService	[bank_statistics_daily, transactions_preaggregated_daily, transactions_preaggregated_monthly]

Transaction processing in more detail:

Big Data Functionality in the project

• Use Cassandra for transaction logic
• Enable Ggneral bank statistics and pre-aggregated analytics for each user
• Create data generator for transactions to test Big Data functionality
• Use Kafka for transaction logic defined

Code related to above points located in ddl, analytics_service, card_manager, card_service and transaction_service. For more details read our documentation -- https://proximal-bladder-a8d.notion.site/Online-Web-Banking-Project-74a56734638b44c184b3505ad26338d4

Run the Project

Start Kafka

Firstly, run Kafka to enable communication between microservices. Download Kafka, unzip the archive and run the below scripts:

• Start zookeeper server: ./bin/zookeeper-server-start.sh config/zookeeper.properties
• Start kafka bootstrap server: ./bin/kafka-server-start.sh config/server.properties
• Create topic: ./bin/kafka-topics.sh --zookeeper 127.0.0.1:2181 --topic TransactionService --create --partitions 3 --replication-factor 1

You can verify everything works correctly using the command below to read messages from topics with Kafka consumer CLI:
./bin/kafka-console-consumer.sh --bootstrap-server 127.0.0.1:9092 --topic TransactionService --from-beginning

Prepare Configurations

For each of the microservices, create a virtual environment. More info: https://www.linuxcapable.com/how-to-setup-python-3-virtual-environment-on-ubuntu-20-04/

# Prepare the Auth Service
python3.8 -m venv auth_service_venv
source auth_service_venv/bin/activate
pip install -r requirements.txt

# Prepare the Registration Service
python3.8 -m venv registration_service_venv
source registration_service_venv/bin/activate
pip install -r requirements.txt

# Prepare the Transaction Service
python3.8 -m venv transaction_service_venv
source transaction_service_venv/bin/activate
pip install -r requirements.txt

# Prepare the Card Service
python3.8 -m venv card_service_venv
source card_service_venv/bin/activate
pip install -r requirements.txt

# Prepare the Card Manager Service
python3.8 -m venv card_manager_venv
source card_manager_venv/bin/activate
pip install -r requirements.txt

# Prepare the Orchestration Service
python3.8 -m venv orchestrator_service_venv
source orchestrator_service_venv/bin/activate
pip install -r requirements.txt

# Prepare the Analytics Service
python3.8 -m venv analytics_service_venv
source analytics_service_venv/bin/activate
pip install -r requirements.txt

Start the Microservices

NOTE: In kubernetes communication between microservices is not hardcoded and works via load balancer. However, if you want to start microservices LOCALLY specify the same ports as in the examples below, since links to microservices are temporary hardcoded in the .env files.

NOTE: Inside each microservice folder add a .env file and secrets folder with .pem files for RSA2048. Otherwise, the launch will be unsuccessful.

# Start the Auth Service
source auth_service_venv/bin/activate
bash start_service.sh

# Start the Registration Service
source registration_service_venv/bin/activate
bash start_service.sh

# Start the Transaction Service
source transaction_service_venv/bin/activate
bash start_service.sh

# Start the Card Service
source card_service_venv/bin/activate
bash start_service.sh

# Start the Card Manager Service
source card_manager_venv/bin/activate
bash start_service.sh

# Start the Orchestration Service
source orchestrator_service_venv/bin/activate
bash start_service.sh

# Start Kafka Streams workers
faust -A kafka_streams worker -l info
# or 
bash start_service.sh

Start the FrontEnd

FrontEnd is implemented using React. To launch the app, install npm on Ubuntu:
https://www.digitalocean.com/community/tutorials/how-to-install-node-js-on-ubuntu-20-04

In case of errors on linux check this link: https://stackoverflow.com/questions/43494794/webpack-html-webpack-plugin-error-child-compilation-failed

If above step is performed, run the below commands:

npm install
npm start

Deploy the Project

# Main reference -- https://docs.cortex.dev/workloads/async/example
aws ecr create-repository --repository-name web_banking_auth_service

# Current value
AWS_ACCOUNT_ID=218145147595

aws ecr get-login-password --region eu-central-1 | docker login --username AWS --password-stdin 218145147595.dkr.ecr.eu-central-1.amazonaws.com

# Connect to Kubernetes with kubectl
# https://aws.amazon.com/premiumsupport/knowledge-center/eks-cluster-connection/
aws eks --region eu-central-1 update-kubeconfig --name web-banking

# Change links on microservices in env files based on a new load balancer

# Deploy Kafka 
# See "Configure Confluent for Kubernetes" section

# Create topics in control-center: TransactionService, CardService, ResultsTopic.
# To connect to it use the next command
kubectl port-forward controlcenter-0 9021:9021

# Change log level from DEBUG to INFO in all microservices

# [If needed] Deploy services interacted with kafka
docker build . -t transaction_service:0.1
docker tag transaction_service:0.1 denys8herasymuk/web-banking-transaction-service:0.1
docker push denys8herasymuk/web-banking-transaction-service:0.1

# Deploy transaction service
bash deploy_service.sh

# Deploy card service
bash deploy_service.sh

# Deploy all API microservices
cortex deploy

# [If needed] Create microservice docker image
docker tag auth_service:0.1 218145147595.dkr.ecr.eu-central-1.amazonaws.com/web_banking_auth_service
docker tag registration_service:0.1 218145147595.dkr.ecr.eu-central-1.amazonaws.com/web_banking_registration_service
docker push 218145147595.dkr.ecr.eu-central-1.amazonaws.com/web_banking_auth_service

# [If needed] Deploy all other microservices, for each use the next command
bash deploy_service.sh

# Set up API Gateway

# Just useful command
cortex delete auth-service

Kubernetes Cluster Info

# Get a cluster info
cortex cluster info

# Open dashboards
# http://<operator_url>/dashboard

# Connect to Kubernetes with kubectl
# https://aws.amazon.com/premiumsupport/knowledge-center/eks-cluster-connection/
aws eks --region eu-central-1 update-kubeconfig --name web-banking

# To find logs, go to your account Cloudwatch

# ECR login
aws ecr get-login-password --region eu-central-1 | docker login --username AWS --password-stdin 218145147595.dkr.ecr.eu-central-1.amazonaws.com

Configure Confluent for Kubernetes

helm repo add confluentinc https://packages.confluent.io/helm
helm repo update
helm upgrade --install confluent-operator confluentinc/confluent-for-kubernetes
kubectl apply -f ./confluent-platform.yaml

# View Control Center
kubectl port-forward controlcenter-0 9021:9021

# And go to http://localhost:9021