隐私计算平台互联互通协议参考实现

本仓库存放隐私计算平台互联互通协议参考实现的代码

构建

interconnection-impl引用了spu仓库代码，需要根据spu构建前提在编译环境上安装好依赖库

然后执行以下构建指令：

bazel build ic_impl/ic_main

运行

ECDH-PSI

本地同时执行以下两条指令：

bazel run ic_impl/ic_main -- -rank=0 -algo=ECDH_PSI -protocol_families=ECC \
        -in_path ic_impl/data/psi_1.csv -field_names id -out_path /tmp/p1.out \
        -parties=127.0.0.1:9530,127.0.0.1:9531

bazel run ic_impl/ic_main -- -rank=1 -algo=ECDH_PSI -protocol_families=ECC \
        -in_path ic_impl/data/psi_2.csv -field_names id -out_path /tmp/p2.out \
        -parties=127.0.0.1:9530,127.0.0.1:9531

SS-LR

运行SS-LR之前，需要先启动Beaver服务。Beaver服务的代码位于SPU仓库中，需要将SPU代码克隆到本地，然后编译并启动Beaver服务：

git clone [email protected]:secretflow/spu.git
cd spu && bazel run libspu/mpc/semi2k/beaver/ttp_server:beaver_server_main -- -port=9449

启动Beaver服务后，本地同时执行以下两条指令：

bazel run ic_impl/ic_main -- -rank=0 -algo=SS_LR -protocol_families=SS \
        -dataset=ic_impl/data/perfect_logit_a.csv -has_label=true \
        -use_ttp=true -ttp_server_host=127.0.0.1:9449 \
        -parties=127.0.0.1:9530,127.0.0.1:9531

bazel run ic_impl/ic_main -- -rank=1 -algo=SS_LR -protocol_families=SS \
        -dataset=ic_impl/data/perfect_logit_b.csv -has_label=false \
        -use_ttp=true -ttp_server_host=127.0.0.1:9449 \
        -parties=127.0.0.1:9530,127.0.0.1:9531

环境变量传参

为满足北京金融科技产业联盟的调度层互联互通标准对算法组件接口的要求，interconnection-impl支持SS-LR算法从环境变量读取配置参数

当某个参数在环境变量和命令行选项都被指定时，优先选择读取环境变量参数

环境变量定义

环境变量	参考值	描述
runtime.component.parameter.algo	ss_lr	algorithm
runtime.component.parameter.protocol_families	ss	comma-separated list of protocol families
runtime.component.parameter.batch_size	21	size of each batch
runtime.component.parameter.last_batch_policy	discard	policy to process the partial last batch of each epoch
runtime.component.parameter.num_epoch	1	number of epoch
runtime.component.parameter.l0_norm	0	l0 penalty term
runtime.component.parameter.l1_norm	0	l1 penalty term
runtime.component.parameter.l2_norm	0.5	l2 penalty term
runtime.component.parameter.optimizer	sgd	optimization algorithm to speed up training
runtime.component.parameter.learning_rate	0.0001	learning rate parameter in sgd optimizer
runtime.component.parameter.sigmoid_mode	minimax_1	sigmoid approximation method
runtime.component.parameter.protocol	semi2k	ss protocol type
runtime.component.parameter.field	64	field type, 64 for Ring64, 128 for Ring128
runtime.component.parameter.fxp_bits	18	number of fraction bits of fixed-point number
runtime.component.parameter.trunc_mode	probabilistic	truncation mode
runtime.component.parameter.shard_serialize_format	raw	serialization format used for communicating secret shares
runtime.component.parameter.use_ttp	true	whether to use beaver service
runtime.component.parameter.ttp_server_host	ip:port	remote ip:port or load-balance uri of beaver service
runtime.component.parameter.ttp_session_id	interconnection-root	session id of beaver service
runtime.component.parameter.ttp_adjust_rank	0	which rank do adjust rpc call to beaver service
system.storage	file://path/to/root	root path of input / output file
runtime.component.input.train_data	{"namespace":"data","name":"perfect_logit_a.csv"}	relative path and name of input file
runtime.component.parameter.skip_rows=1	1	number of skipped rows from dataset
runtime.component.parameter.has_label	true	if true, label is the last column of dataset
runtime.component.output.train_data	{"namespace":"output","name":"result_a"}	relative path and name of output file

FAQ

若构建失败并提示 Host key verification failed，解决方式如下:

• 首先，向github账户添加SSH key

• 然后在本地执行以下指令(以选择 RSA 作为 SSH key 类型为例)：

ssh-keyscan -t rsa github.com >> ~/.ssh/known_hosts