tree: 7a5a162205156d7a5427c70905772602eed0883e [path history] [tgz]
  1. access_control/
  2. authentication/
  3. execution/
  4. frontend/
  5. management/
  6. proto/
  7. scheduler/
  8. storage/
  9. utils/
  10. README.md
services/README.md

permalink: /services

Teaclave Services

This directory contains protocol definitions and implementations of services in the Teaclave platform.

Services

The platform includes several services (or subsystem) running inside TEE, and communicated with mutual-attested channels. They coordinate together to provide a safe and secure FaaS platform.

  • Authentication Service: A reference implementation of user authentication infrastructure. Here, we use JSON Web Token (JWT), a simple and widely-used authentication standard, to provide a secure authentication mechanism in the platform. Clients need to get valid token before interacting with the platform.
  • Frontend Service: This is the entry point of all requests from users. It will validate user's identity/token and forward requests to appropriate services.
  • Management Service: This service plays an important role in the whole services. It handles almost all requests, such as registering functions/data, creating tasks, and invoking tasks. Also, the management service will contact the access control service to authorize operations when needed. In addition, task and function information will be persistent into the storage services.
  • Storage Service: Basically, the storage service stores persistent data like function, execution data, and task information in the platform. Here, we deploy a key-value database (an implementation of LevelDB) in TEE and use the protected file system (secured by the enclave) for data persistence.
  • Access Control Service: Provides a flexible access control domain specific language to support access control rules for secure multi-party computation. The access control engine is written in Python and evaluated in SGX. Please read this document to learn more about the design of it.
  • Scheduler Service: Schedules staged tasks ready for execution to a proper execution node with desirable capabilities.
  • Execution Service: A host of different executors interacting with the scheduler service to complete tasks. There could be many execution service instances (or nodes) with different capabilities deployed in a cloud infrastructure.

Structure

A service is consist of two parts: app (untrusted) and enclave (trusted). The app part is responsible for launching and terminating the service, which the enclave part is to serve RPC requests from trusted channels. Typically, a service‘s implementation contains two important structs and one trait. Let’s take the frontend service as an example.

  • TeaclaveFrontendService (struct): Define properties or configurations along with the lifetime of the service. For example, the frontend service need to hold clients (with established trusted channels) to communicate with the authentication service and management service.
  • TeaclaveFrontendError (struct): Define errors which may occur in this service, authentication error, for example.
  • TeaclaveFrontend (trait): Define functions (requests) the service need to handle. The trait will be automatically derived from definitions in the ProtoBuf file and can be imported from the teaclave_proto crate.

RPC and Protocols

We use ProtoBuf to define messages and RPC interfaces of the Teaclave services. You can find detailed protocol definitions in the proto directory. In addition, utility functions, traits, and structures will be automatically generated to help implementing services to handle requests and send responses. This is done by a build time tool called proto_gen.

Topology

These services are communicating through RPC with remote attestation. This topological graph illustrates connections between services.

clients => authentication <-+       +----> storage <----+
                            |       |                   |
clients => frontend ----------> management            scheduler <-- execution
                                    |
                                    +--> access_control


                                                  =>      api endpoint connections
                                                  -> internal endpoint connections

Internal endpoint connections will be established and verified with mutual remote attestation to ensure the integrity and confidentiality of the whole system. Therefore, clients can trust the whole platform and safely interacting with the system through the attested authentication and frontend services.