Deep Learning Workbench Security¶
Deep Learning Workbench (DL Workbench) is a web application running within a Docker* container.
Run DL Workbench¶
Unless necessary, limit the connections to the DL Workbench to localhost (127.0.0.1), so that it is only accessible from the machine the Docker container is built on:
- The script starting the DL Workbench from the package” ensures that the container and the web
application are accessible only from the <tt>localhost</tt> by default.
- * When using <tt>docker run</tt> to @ref workbench_docs_Workbench_DG_Run_Locally “start the DL Workbench from Docker Hub”, limit connections for the host IP 127.0.0.1.
For example, limit the connections for the host IP to the port <tt>5665</tt> with the <tt>-p 127.0.0.1:5665:5665</tt> command . Refer to <a href=” https://docs.docker.com/config/containers/container-networking/#published-ports “
- >Container networking</a> for
details.
<h1>Authentication Security</h1>
DL Workbench uses @ref workbench_docs_Workbench_DG_Authentication “authentication tokens” to access the application. The script starting the DL Workbench creates an authentication token each time the DL Workbench starts. Anyone who has the authentication token can use the DL Workbench.
When you finish working with the DL Workbench, log out to prevent the use of the DL Workbench from the same browser session without authentication.
To invalidate the authentication token completely, @ref workbench_docs_Workbench_DG_Docker_Container “restart the DL Workbench”.
<h1>Use TLS to Protect Communications</h1>
@ref workbench_docs_Workbench_DG_Configure_TLS “Configure Transport Layer Security (TLS)” to keep the authentication token encrypted.