For developers using BioLib Encrypted Cloud, one benefit is that the application source code and databases will not be sent to the end-users' machine. BioLib Encrypted Cloud utilizes modern encryption technologies to make applications available in a way that protects both the end-users' data and the developers' code. You can find general information about BioLib Encrypted Cloud here.
As a developer, to specify that your code should only be executed in BioLib Encrypted Cloud, include the
client_side_include tag in your
client_side_include, specify which modules of your application are allowed to run on the client's computer. All other modules will be executed in BioLib Encrypted Cloud.
Say you have an application that makes inference with a TensorFlow model that you wish to keep secret. It includes the following modules:
If you wish to allow the main module to run on the client's computer, specify this under the
biolib_version: 1 modules: main: path: src/__main__.py executor: 'bl-python3:*' tensorflow: path: build/model.pb executor: 'bl-tensorflow:*' client_side_include: - src/__main__.py # <-- This file will run on the client's computer # all other files will not be sent to the client arguments: ...
Currently, only TensorFlow frozen graph (
.pb) and Open Neural Network Exchange (
.onnx) files are supported in BioLib Encrypted Cloud.