Web-scale Data Management Project - Group 20

Deployment types:

Kubernetes cluster

Run the deploy-cloud.sh script. If you get the following error error when creating "k8s/ingress-service.yaml", run kubectl apply -f k8s/. To delete, run the delete-cloud.sh script.

Requirements: You need to have access to kubectl of a k8s cluster.

Minikube (local k8s cluster)

Run deploy-charts-minikube-local.sh.

Requirements: You need to have minikube (with ingress enabled) and helm installed on your machine.

Docker-compose (local development)

Run docker-compose up --build.

Requirements: You need to have docker and docker-compose installed on your machine.

Project structure

• env Folder containing the Redis env variables for the docker-compose deployment

• helm-config Helm chart values for Redis and ingress-nginx

• k8s Folder containing the kubernetes deployments, apps and services for the ingress, order, payment and stock services.

• k8s-local Folder containing the kubernetes deployments, except they don't pull from the cloud

• order Folder containing the order application logic and dockerfile.

• payment Folder containing the payment application logic and dockerfile.

• stock Folder containing the stock application logic and dockerfile.

• test Folder containing some basic correctness tests for the entire system. (Feel free to enhance them)

Extra:

After implementing our synchronous SAGA based architecture we tried implementing a partly asynchronous Event-Driven architecture with a message queue. This was however not entirely consisten but did seem to scale very well. In the 7-implement-event-driven-architecture branch, you can run docker-compose up --build and see a working version with 6 (per microservice) proxies and 10 services (per microservice) running together. We did however not manage to deploy it with k8s (yet)

Declaration of Honour:

• Joey: Stock implementation + start of proxies for event-driven architecture
• Gaspar: Sagas implementation + Cloud deployment/testing :( --> GCP's quotas inhibit horizontal scaling
• Safouane: Payment implementation + CRDT Investigation with antidoteDB :( --> Practical implementation of CRDT's with lowerbounds is lacking
• Merlijn: Order Implementation + Event Driven Architecture investigation :( --> Async programming is hard, k8s deployment was hard