BI-SGX : Bioinformatic Interpreter on SGX-based Secure Computing Cloud

Introduction

These codes are interpreter which uses Intel(R) SGX's protection features and specializes for bioinformatical computations. The basical framework of these codes are heavily based on Intel(R) Software Guard Extensions (SGX) Remote Attestation End-to-End Sample.

Contrary to Intel's original RA sample, this model uses inverted client-server model. In other words:

ISV (SGX side) runs as server, which waits for SP's request after RA
SP (non-SGX side) runs as client, which send request to ISV after RA

The reason why implemented as this inverted model is, in bioinformatical computation (or secret computation), usually user would like to use SGX features as cloud server with sending or storing their data in protected way.

The main contributions of this code are:

This interpreter provides much friendly coding rule than using SGXSDK itself.
Provides code protect feature without any annoying build/installation.
Provides various bioinfomatically-utilized methods as default in interpreter.
Provides integrated cloud platform which can be used as secure cloud storage and secure cloud computation base, and can make two features corporate.

Installation

Prerequisites

Contrary to Intel's original RA sample, these codes are developed for Linux platform, so Windows OS is unexpected. Perhaps these can be run on Windows, but no performance guarantee.

Ensure your platform's OS is
- CentOS 7.4 (64-bit)
- Ubuntu 16.04 LTS (64-bit)
- Ubuntu 18.04 LTS (64-bit)
Ensure that you have built and installed the Intel SGX packages (Both of these must be SGX 2.x version):
- Intel SGX Software Development Kit and Platform Software package for Linux. You must install SGXSDK since version is 2.8
- Intel SGX Driver for Linux
Ensure that you have installed MySQL and MySQL Connector/C++.

$ sudo apt install mysql-server
$ sudo apt install libmysqlcppconn-dev

Run the following commands to install the required packages to build the RA code sample (this assumes you have installed the dependencies for the Intel SGX SDK and PSW package)
- On CentOS 7.4
```
$ yum install libcurl-devel
```
- On Ubuntu 16.04
```
$ apt-get install libcurl4-openssl-dev
```
Run the following command to get your system's OpenSSL version. It must be at least 1.1.0:

$ openssl version

If necessary, download the source for the latest release of OpenSSL 1.1.0, then build and install it into a non-system directory such as /opt (note that both --prefix and --openssldir should be set when building OpenSSL 1.1.0). For example:

$ wget https://www.openssl.org/source/openssl-1.1.0i.tar.gz
$ tar xf openssl-1.1.0i.tar.gz
$ cd openssl-1.1.0i
$ ./config --prefix=/opt/openssl/1.1.0i --openssldir=/opt/openssl/1.1.0i
$ make
$ sudo make install

Configure and compile

First, prepare the build system (GNU* automake and autoconf) by running bootstrap, and then configure the software package using the configure command. You'll need to specify the location of OpenSSL 1.1.0. See the build notes section for additional options to configure.

$ ./bootstrap
$ ./configure --with-openssldir=/opt/openssl/1.1.0i
$ make

Or you can also execute these 3 commands at once by executing following command:

$ source make.sh

Both make clean and make distclean are supported.

User agent

In this project, libcurl is supported for SP as user agent on Linux to communicate with IAS. Using wget is not recommended, but if you want, you can change setting following Intel`s original RA sample's README.

Usage

Set up DB for secret data storing

There must be two tables within one database.

One is table named userinfo, which is for management of login info. The format of userinfo is following:

+-----------+------+------+-----+---------+-------+
| Field     | Type | Null | Key | Default | Extra |
+-----------+------+------+-----+---------+-------+
| username  | text | YES  |     | NULL    |       |
| pass_hash | text | YES  |     | NULL    |       |
| privilege | text | YES  |     | NULL    |       |
+-----------+------+------+-----+---------+-------+

And another is named stored_data, which is for storing secret data. The format of stored_data is following:

+-----------+----------+------+-----+---------+-------+
| Field     | Type     | Null | Key | Default | Extra |
+-----------+----------+------+-----+---------+-------+
| dataname  | text     | YES  |     | NULL    |       |
| owner     | text     | YES  |     | NULL    |       |
| data      | longblob | YES  |     | NULL    |       |
| datatype  | text     | YES  |     | NULL    |       |
| cipherlen | int(11)  | YES  |     | NULL    |       |
+-----------+----------+------+-----+---------+-------+

So ISV have to create above tables as aforementioned formats before start BI-SGX.

After started program, ISV must initialize DB login info. As default, this initialization will be done by following context:

host = "localhost";
user = "BI-SGX";
password = "bisgx_sample";
database = "`BI-SGX`";

But hardcoded login context is extremely insecure, so you should edit void BISGX_Database::initDB() in isv.cpp to manually enter login context using like cout. More secure login method will be impremented in the future.

Start programs

You can run ISV (SGX server) code by entering command:

$ ./run-isv

To run SP (non-SGX client) code, enter command:

$ ./run-sp

Default IP address is localhost and default port is 7777. You can also use ./run-client to start ISV and ./run-server to start SP, but their names are inconsistent with their actual roles.

Send data from SP to ISV

After complete RA, you can send your file to ISV from SP.

If you are data owner (use data storage feature)

Firstly, you have to prepare login context as login.ini. Username, password, and authority-type should be described like following:

testuser # username 
testpass12345 # password
O # authority type. "O" for data owner and "R" for researcher.
FASTA # option. designate data type of data to upload or designate other processes to require.

Secondly, you have to prepare data to store to cloud server. Currently 2 data types are acceptable; One is to be set of int or double, and another is to be set of char. The separator must be newline (\n).

After completing remote attestation, you will be required to input filename of dataset. Then cloud DB storing will be executed and the result status will be returned from cloud server.

For more detailed usage, please see at BI-SGX's wiki.

If you are researcher (use interpreter feature)

Firstly, you have to prepare login.ini as with the case of data owner. Note that authority type is R for researcher.

After completing remote attestation, you will be required to input filename of interpreter code. Then interpreter code will be executed at cloud's enclave and the result will be returned from cloud server.

Specification/grammer of BI-SGX

See at BI-SGX's wiki for interpreter's specifications and grammers.

Implemented features

Inverted client-server communication model
Remote Attestation between SP and ISV
Cryptographic features to send secret in secure
Load secret into enclave
DB storing feature for data owner
Interpreter which runs inside enclave

TODO

Adopt more effective algorithms for processing internal tables and internal codes in interpreter.

Website

bi-sgx.net

LICENSE

All of these codes are developed and distributed under Intel Sample Source Code license. See the LICENSE file for detail.

And many of BI-SGX's interpreter implementation owe to following book: ISBN978-4-7973-6881-9「明快入門インタプリタ開発」

mohitWrangler/BI-SGX