README
¶
Use Cases
|- usecases
|-- dna
|-- hypnogram
|-- models
The use cases demonstrate two practical applications of SPADE for DNA and hypnogram
data. Besides, the models directory includes the entities' implementation for the centralized
system model. To implement the client-server model, we utilized standard gRPC and protocol buffer libraries.
NOTE: To enable database management for use cases, we employed go-sqlite3, which requires gcc to be installed beforehand. Here, you can check the requirements for go-sqlite3 installation.
System Model
|-- models
|--- analyst.go
|--- user.go
|--- server.go
|--- config.go
|--- handlers.go
Our experiments implemented a centralized system model involving three key entities:
User, Server/Curator, and Analyst. The code for each entity contains a Start...()
function where we set up the SPADE scheme and share the public parameters. The order
of starting entities should be, as in real-life applications, starting by the server, then users,
and analyst.
Furthermore, there are config.go and handlers.go, including general configuration structures,
and handler functions for gRPC and SQLite, respectively.
Use Case #1: Hypnogram Data
|-- hypnogram
|--- dataset
|--- app.go
|--- app_test.go
|--- config.go
We have a dataset of 590 files for the first use case, each including 1000 entries. The app.go
contains the hypnogram application for initializing the server, users, and analyst by using
the default configuration in config.go and calling each entity from models.
Note: to initialize users, we are reading all the files from the dataset by assuming each file
represents a user.
To test the application and get the results, there is a app_test with different test functions,
one for each entity, where you can run it by the following commands (please note that the order of
running them should be as it mentioned):
Testing Instruction
move to the hypnogram directory by $ cd usecases/hypnogram/
- Start the server
go test -run TestInitServer - Start users
go test -run TestInitUsers - Start analyst
go test -run TestInitAnalyst
By running the mentioned test, the server will start listening to get requests from users and store the hypnogram ciphers locally in an SQLite database. Later, by receiving the analyst's request, the server will send him decryption keys along with the user's cipher. The analyst, by receiving the keys and ciphers, will be able to partially decrypt the cipher and get the results.
Use Case #2: DNA sequences
|-- dna
|--- dataset
|--- app.go
|--- app_test.go
|--- config.go
We have a dataset of 5012 files for the first use case, each including 78214 entries. The app.go
contains the DNA application for initializing the server, users, and analyst by using
the default configuration in config.go and calling each entity from models.
Note: to initialize users, we are reading all the files from the dataset by assuming each file
represents a user.
To test the application and get the results, there is a app_test with different test functions,
one for each entity, where you can run it by the following commands (please note that the order of
running them should be as it mentioned):
Testing Instruction
move to the dna directory by $ cd usecases/dna/
- Start the server
go test -run TestInitServer - Start users
go test -run TestInitUsers - Start analyst
go test -run TestInitAnalyst
By running the mentioned test, the server will start listening to get requests from users and store the hypnogram ciphers locally in an SQLite database. Later, by receiving the analyst's request, the server will send him decryption keys along with the user's cipher. The analyst, by receiving the keys and ciphers, will be able to partially decrypt the cipher and get the results.
Protocol Sequence Diagram
sequenceDiagram
autonumber
actor u as User(s)|id
participant c as Curator/Server
actor a as Analyst(s)
%% start the protocol
note over c: generate q, g <br> n=maximum number of users, <br> m=maximum number of elements
c ->> c: msk, mpk = setup(n, m)
c -->> u: (q, g, mpk)
c -->> a: (q, g, mpk)
u ->> u: generate random #alpha; <br> regKey = g^ #alpha; <br> ct = Enc(mpk, #alpha; , m)
u -->> c: (id, regKey, ct)
c ->> c: stores regKeys[id] = regKey <br> stores cts[id] = ct
%% analyst query
a -->> c: (id, v)
c ->> c: dkv = keyDer(id, v, msk, regKeys[id])
c -->> a: (dkv, c[id])
%% end of protocol
Directories