duyuefeng0708/smart_contract

Introduction

This project targets the ubiquitous keyword search function and aims to deliver encrypted search services for blockchain-based applications.

We provide two encrypted search services for private blockchain and public blockchain, respectively. Both services ensure the confidentiality of data and query, without affecting the correctness and verifiability of blockchain.

Note that the service for public blockchain further considers malicious threats in an untrusted distributed environment.

Publication

• Chengjun Cai, Jian Weng, Xingliang Yuan, Cong Wang, "Enabling Reliable Keyword Search in Encrypted Decentralized Storage with Fairness", under major revision of IEEE Transactions on Dependable and Secure Computing (TDSC)

• Shengshan Hu, Chengjun Cai, Qian Wang, Cong Wang, Luo Xiangyang, and Kui Ren, "Searching an Encrypted Cloud Meets Blockchain: A Decentralized, Reliable and Fair Realization", in the 37th International Conference on Computer Communications (INFOCOM'18)

Requirements

Recommended environment for private blockchain network:

Two (or more) AWS EC2 instances using t2.medium (2 vCPU, 4 GB RAM) in Ubuntu 16.04 LTS.

Recommended public blockchain test network:

ROPSTEN 

Required tools:

ethereum (geth)
truffle
python 3.7 (additional packages: Cryptography and BitVector)

Configuration of private blockchain network

Geth genesis.json template

{
    "config": {
        "chainId": 8888,
        "homesteadBlock": 0,    
        "eip155Block": 0,
        "eip158Block": 0
    },
    "coinbase" : "0x0000000000000000000000000000000000000000",
    "difficulty" : "0x1000",
    "extraData" : "",
    "gasLimit" : "0x7a1200",
    "nonce" : "0x0000000000000042",
    "mixhash" : "0x0000000000000000000000000000000000000000000000000000000000000000",
    "parentHash" : "0x0000000000000000000000000000000000000000000000000000000000000000",
    "timestamp" : "0x00",
    "alloc": {
    }
}

$ geth init template.json

Contract deployment

Put contract file (e.g., Mycontract.sol) into the truffle suite, i.e., in the folder 'contracts'.

Make sure it is authenticated in the geth console

> geth attach
> personal.unlockAccount(eth.coinbase, PASSWORD, 0)
> exit

Use the truffle command migrate to deploy the contract to the blockchain network

$ truffle migrate 

Dataset

We use two datasets to test the encrypted search services on private blockchain and public blockchain, respectively.

• Enron Email Dataset (CMU)

• Synthetic financial dataset (with specification provided by Hong Kong Applied Science and Technology Research Institute (ASTRI)

Execution examples for Application I (Private blockchain)

Format input dataset

For a financial dataset with 128-dimensional vectors, we first extract each row as a separate file. The keyword inside each file is represented as

attribute:value

For example, keyword UID:1 is built with attribute UID and its value in that specific row, i.e., 1.

We can then put all formatted files into a folder

Build encrypted index

The encrypted index is built with python file index.py

$ python index.py <Folder> 

The built index is stored in the json file label.json

The relation between filename and file id can be found in fileid.txt in

Store built encrypted index to the blockchain

We load the label.json and send the encrypted index them as transactions to the blockchain via javascript file StoreIndex.js

We execute the command with the truffle suite

$ truffle exec DSSE/StoreIndex.js

Generate encrypted search token

Given a plaintext keyword input, we generate the corresponding encrypted search token

$ python searchtoken.py <Keyword>

The generated search token is stored in the json file searchtoken.json

Execute encrypted search on the blockchain

We load the searchtoken.json and send the search token them as a transaction to the blockchain via javascript file Search.js

We execute the command with the truffle suite

$ truffle exec DSSE/Search.js

The search result is then presented

Execution examples for Application II (Public blockchain)

Generate encrypted file index

Put all formatted files into a folder

Client generates the file index

$ python client_democ.py <Folder> 

The index digest and set hash(es) of all added file(s) are presented in the console log, and they will then be recorded on the blockchain via contract function Storeindex

The built file index is stored as binary file fileindex-demo.p

Confirm index digest and set hash(es)

Service peer load the file fileindex-demo.p, and execute service_peer_check.py

$ python service_peer_check.py

The computed index digest and set hash(es) of all added file(s) are presented and recorded on the blockchain via contract function Storeindex

If client and service peer have confirmed on the same index digest and set hash(es), the process continues. Otherwise, the process aborts.

Generate encrypted search token

Given a query keyword, client generates the encrypted search token

$ python searchtoken_demo.py <Keyword>

The generated search token is presented in the console and recorded on the blockchain via contract function Search.

Execute encrypted search

Service peer loads the search token and execute search operation

$ python service_peer_democ.py <Search Token>

The search result and result digest are presented on the console, and service peer will record the result digest to the blockchain via contract function Search.