Welcome to Treebeard, a fault-tolerant and highly scalable Oblivious RAM (ORAM) data store designed to provide strong security guarantees by hiding access patterns from the storage server and randomizing client access.
- Scalability: Treebeard is designed with horizontal scalability in mind, enabling scaling the system without leaking security information.
- Fault Tolerance: Using the Raft consensus algorithm, Treebeard ensures the resilience of your data store even in the face of node crashes or failures, enhancing the overall reliability of your system.
- High Throughput: Treebeard achieves high-throughput performance, ensuring that your applications can handle large volumes of data seamlessly and efficiently.
- Configurable: Treebeard recognizes the diverse needs of different projects and offers a high level of configurability to adapt to your specific requirements.
- Easy to Deploy: We've designed the setup of Treebeard to be straightforward, ensuring that you can integrate our secure and high-performance data store effortlessly into your projects.
- Easy to Extend: Treebeard's modular architecture makes it easy to extend functionality. You can add new storage layers to Treebeard to support different use cases.
The following parts are generated, but if you want to generate them again yourself, you can run the provided script:
./scripts/generate_protos.shYou will need Go 1.20 to run the project.
Each of the directories in the experiments directory contains several experiments. For example, dist_experiments has four subdirectories (uniform, zipf0.2, zipf0.6, zipf0.8, zipf0.99). The run_scripts.sh file in the experiments directory runs all of the experiments.
To run all the experiments:
./experiments/run_experiments.shThis script uses ansible/deploy.yaml and ansible/experiment.yaml to deploy the experiment and run the experiment. It then gathers the results on the local folder.
Each experiment should have a trace.txt file, which is the generated YCSB trace file. These are not provided since they are very large. You can generate them using the YCSB client.
Every experiment has its own configuration files. The following are the current configurations:
- jaeger_endpoints.yaml: configures the jaeger backend for collecting the distributed tracing telemetries.
- oramnode_endpoints.yaml: endpoints for the ORAM services.
- shardnode_endpoints.yaml: endpoints for the Shard node services.
- router_endpoints.yaml: endpoints for the router services.
- redis_endpoints.yaml: endpoints for the redis services.
- parameters.yaml: configurable parameters for each experiment. The comments explain what each configurable variable does.
Feel free to change the files to add a new experiment.