Rust Distributed Service Simulation - Statistics Service

The system consists of a distributed International Statistics Service. The Application functionality is provided by a remote object residing at the server side. Client objects interact with the server through remote method invocations. The client can invoke the methods defined in the server’s remote interface specification.

⚠️ This is a simulation created for educational purposes. There is still unimplemented features, but the goal is to explore complex distributed systems development, and implement performance-enhancing features.

About the simulation

The simulation models a distributed system implemented in Rust, running on a local machine. To mimic real-world conditions, the code introduces artificial delays to simulate network latency. A "zone" represents a geographical area, and requests between different zones experience higher latency. The primary goal of this simulation is to evaluate the performance of a distributed system, with plans for future performance enhancements.

Key Features:

• 5 gRPC Servers: Each server operates in a distinct geographical zone.
• Client Distribution: 1000 clients are assigned to each server, totaling 5000 clients across the simulation.
• Request Flow: Clients first attempt to connect to the server within their own zone.
  • In-Zone Requests: Clients experience an 80 ms delay to simulate network latency within the same zone.
  • Cross-Zone Requests: If a client needs to connect to a server outside its zone, a 170 ms delay is applied to simulate the increased latency.

The simulation is finished when all 5000 clients have completed their gRPC requests.

Results

The following is box-plots of the most important statistics from the simulation:

Usage

To run a server with server_id 1:

cargo run --bin server 1

The client binary uses a file of requests to simulate different clients connecting and executing a request. To run the client with client_id 1:

cargo run --bin client request_files/client_1.txt 1

Resources

csv2sqlite - Python script to load CSV to SQLite:
https://github.com/rufuspollock/csv2sqlite

Dataset for the statistics:
https://public.opendatasoft.com/explore/dataset/geonames-all-cities-with-a-population-1000/table/?disjunctive.cou_name_en&sort=name

tonic (gRPC package) server client tutorial:
https://github.com/hyperium/tonic/blob/master/examples/helloworld-tutorial.md