Lineage Stash: Fault tolerance (almost) for free Privacy ...

1
Lineage Stash: Fault tolerance (almost) for free Stephanie Wang, Philipp Moritz, Robert Nishihara, Alexey Tumanov, Ion Stoica Privacy Preserving Data Muling System Project 2: Alvin Tan, Tess Despres, Jean-Luc Watson, Shishir Patil Solution Results Problem Current data muling systems utilize metadata that can reconstruct mobility traces or social networks. Background Use a state-of-the-art anonymous communication protocol, Express [1] (depicted below), to implement a privacy preserving data muling system. What is data muling? What are the privacy concerns? We demonstrate mobility tracking through simulation using only sensor/mule IDs & time of transmission. Simulation: System implementation: We evaluate Express against a baseline plaintext implementation on hardware. Next Steps Energy evaluation of our system on embedded HW Bidirectional communication Batch writes to reduce latency Protection against malicious mules MQTT Publish MQTT Subscribe Write to memcache Express Client Server A Write to in-memory tables Server B Private write (hide data source) Public read (can charge $) hello! We also use a simulation to evaluate scalability with: 100 mobile mules (10-20m BLE range) 1000 stationary sensors (1-15m/s move speed) 128 byte data payloads [1] Eskandarian, S., Corrigan-Gibbs, H., Zaharia, M., & Boneh, D. (2021). Express: Lowering the cost of metadata-hiding communication with cryptographic privacy. In 30th USENIX Security Symposium (USENIX Security 21) (pp. 1775-1792). <data> Sensor A, 9:17am - Mule B

Transcript of Lineage Stash: Fault tolerance (almost) for free Privacy ...

Page 1: Lineage Stash: Fault tolerance (almost) for free Privacy ...

Lineage Stash: Fault tolerance (almost) for freeStephanie Wang, Philipp Moritz, Robert Nishihara, Alexey Tumanov, Ion Stoica

Privacy Preserving Data Muling SystemProject 2: Alvin Tan, Tess Despres, Jean-Luc Watson, Shishir Patil

Solution ResultsProblem

Current data muling systems utilize metadata that can reconstruct mobility traces or social networks.

Background

Use a state-of-the-art anonymous communication protocol, Express[1] (depicted below), to implement a privacy preserving data muling system.

What is data muling?

What are the privacy concerns? We demonstrate mobility tracking through simulation using only sensor/mule IDs & time of transmission.

Simulation:

System implementation:We evaluate Express against a baseline plaintext implementation on hardware.

Next Steps

● Energy evaluation of our system on embedded HW● Bidirectional communication ● Batch writes to reduce latency ● Protection against malicious mules

MQTT Publish

MQTT Subscribe

Write to memcache

Express Client

Server A

Write to in-memory

tables

Server B

Private write(hide data source)

Public read (can charge $)

hello!

We also use a simulation to evaluate scalability with:

● 100 mobile mules (10-20m BLE range)● 1000 stationary sensors (1-15m/s move speed)● 128 byte data payloads

[1] Eskandarian, S., Corrigan-Gibbs, H., Zaharia, M., & Boneh, D. (2021). Express: Lowering the cost of metadata-hiding communication with cryptographic privacy. In 30th USENIX Security Symposium (USENIX Security 21) (pp. 1775-1792).

<data>Sensor A, 9:17am - Mule B