This project utilizes Cassandra database to run theater booking system. In this system a client books some place at a given play of his/her interests. Theater consists of 6 rooms (That's a BIG theater) so there can be 6 plays played simultanously. Each room consists of 12 rows and 10 places at each row, so each room can contain 120 clients. In total there can be 6*120 places booked. (720)

Database consists of 2 tables:

1. plays (storing data about currently available spectacles)

• room: room identifer
• play: title of spectacle

2. rooms (storing data about the state of each room (occupancy))

• room: room identifier
• row: row number
• place: place/column number
• occupied: no/yes -> occupation identifier
• client: client identifier (who booked given place)

As I believe the problem is read-heavy (not all clients will buy the ticket, but all would like to see available places) I used LeveledCompactionStrategy which is recommended in such case. Also having in mind the rule:

• [read-consistency-level] + [write-consistency-level] > [replication-factor] I decided to set read-consistency-level to high and write-consistency-level to low value (assuming 3 nodes that would be 3 and 1 respectively)

To run the project:

1. Copy the content of cluster_start and run it in bash terminal to start the cluster and its nodes (using docker)
2. Run pip install -r requirements.txt to install required packages
3. Initialize the database with python initial.py. You can also use it to roll back database state to initial one (empty)
4. To run a client write python main.py "Unique_client_name" eg: python main.py Jaroslaw
5. Enjoy

• To run some stress tests see files: stress_1.py, stress_2.py, stress_3.py

Problems:

1. Running more than 2 nodes on my equipment caused running out of memory problems (My laptop has 16GB RAM and 2 nodes run in docker containers + additionall system services consume 15 GB already (around 5GB for each node)) -> Any tries to limit memory consumption failed:

• limiting docker container max RAM -> node crashes

2. Stress tests I intended to run with cassandra-stress didn't succeded -> problem with "Failed to connect over JMX; not collecting these stats" Trying to install JMX and configuring it didn't bring any improvement and the Internet stays silent in this matter.
3. Another try, I considered nosqlbench but when trying to run tests after some time the tool crashes.

Great thanks to authors of tutorials:

• https://www.freecodecamp.org/news/the-apache-cassandra-beginner-tutorial/#how-to-set-up-a-cassandra-cluster
• https://docs.datastax.com/en/developer/python-driver/3.25/getting_started/