• git
• make
• autoconf
• libtool
• gcc
• g++
• openssl-devel / libssl-dev (only needed in host/tdx mode)
• libcbor-devel / libcbor-dev (only needed in host/tdx mode)
• cargo (only needed in host mode)
• SGX driver, Intel SGX SDK & PSW: Please refer to this guide to install.
  • Requires Intel SGX SDK and PSW version >= 2.18
• SGX DCAP: please download and install the packages from this page.
  • ubuntu 18.04: libsgx-dcap-quote-verify-dev, libsgx-dcap-ql-dev, libsgx-uae-service
  • Requires Intel DCAP version >= 1.15
• For TDX support, please refer to linux-sgx README to build the packages.

Please follow the command to build RATS TLS from the latested source code on your system.

1. Download the latest source code of RATS TLS

mkdir -p "$WORKSPACE"
cd "$WORKSPACE"
git clone https://github.com/inclavare-containers/rats-tls

2. Build and install RATS TLS

cd rats-tls
cmake -DBUILD_SAMPLES=on -H. -Bbuild
make -C build install

{rats-tls-server,rats-tls-client} will be installed to /usr/share/rats-tls/samples/{rats-tls-server,rats-tls-client} on your system. All instances are placed in /usr/local/lib/rats-tls.

If you want to build instances related to sgx(sgx_ecdsa, sgx_ecdsa_qve, sgx_la), please type the following command.

cmake -DRATS_TLS_BUILD_MODE="sgx" -DBUILD_SAMPLES=on -H. -Bbuild
make -C build install

If you want to run instances on libos occlum, please type the following command.

cmake -DRATS_TLS_BUILD_MODE="occlum" -DBUILD_SAMPLES=on -H. -Bbuild
make -C build install

If you want to run TDX instances, please type the following command.

cmake -DRATS_TLS_BUILD_MODE="tdx" -DBUILD_SAMPLES=on -H. -Bbuild
make -C build install

Note that SGX LVI mitigation is enabled by default. You can set macro SGX_LVI_MITIGATION to 0 to disable SGX LVI mitigation.

Right now, RATS TLS supports the following instance types:

For instance priority, the higher, the stronger. By default, RATS TLS will select the highest priority instance to use.

cd /usr/share/rats-tls/samples
./rats-tls-server

By default, the listening port is 1234. -p option can be used to specify the listening port.

cd /usr/share/rats-tls/samples
./rats-tls-client

Notice: special prerequisites for TDX remote attestation in bios configuration and hardware capability.

Check msr 0x503, return value must be 0:

sudo rdmsr 0x503s

Note that if you want to run SEV-SNP remote attestation, please refer to link to set up the host and guest Linux kernel, qemu and ovmf bios used for launching SEV-SNP guest.

Notice: special prerequisites for SEV(-ES) remote attestation in software capability.

• Kernel support SEV(-ES) runtime attestation, please manually apply these patches.
• Start the attestation evidence broker service in host.

Notice: special prerequisites for CSV(2) remote attestation in software capability.

• Kernel support CSV(2) runtime attestation, please manually apply these patches 1 and 2.

The options of rats-tls-server are as followed:

OPTIONS:
   --attester/-a value   set the type of quote attester
   --verifier/-v value   set the type of quote verifier
   --tls/-t value        set the type of tls wrapper
   --crypto/-c value     set the type of crypto wrapper
   --mutual/-m           set to enable mutual attestation
   --log-level/-l        set the log level
   --ip/-i               set the listening ip address
   --port/-p             set the listening tcp port
   --product-enclave/-P  set to enable product enclave
   --verdictd/-E         set to connect verdictd based on EAA protocol

You can set command line parameters to specify different configurations.

For example:

./rats-tls-server --tls openssl
./rats-tls-server --attester sgx_ecdsa
./rats-tls-server --attester sgx_ecdsa_qve
./rats-tls-server --attester sgx_la
./rats-tls-server --attester tdx_ecdsa
./rats-tls-server --crypto openssl

RATS TLS's log level can be set through -l option with 6 levels: off, fatal, error, warn, info, and debug. The default level is error. The most verbose level is debug.

For example:

./rats-tls-server -l debug

RATS TLS server binds 127.0.0.1:1234 by default. You can use -i and -p options to set custom configuration.

./rats-tls-server -i [ip_addr] -p [port]

You can use -m option to enable mutual attestation.

./rats-tls-server -m
./rats-tls-client -m

In the early bootstrap of rats-tls, the debug message is mute by default. In order to enable it, please explicitly set the environment variable RATS_TLS_GLOBAL_LOG_LEVEL=<log_level>, where <log_level> is same as the values of the option -l.

Please refer to this guide to run Rats Tls with Occlum and rune.

In non-sgx enviroment, it's possible to show the error messages as below when running the command ./rats-tls-client --attester sgx_ecdsa. According to Intel DCAP's implementation, when calling to sgx_qv_get_quote_supplemental_data_size(), if the libsgx_urts library is present, it will try to load QvE firstly. If failed, the verification will be launched by QVL. So the error info can be ignored and have no impact on the final attestation result.

[load_qve ../sgx_dcap_quoteverify.cpp:209] Error, call sgx_create_enclave for QvE fail [load_qve], SGXError:2006.
[sgx_qv_get_quote_supplemental_data_size ../sgx_dcap_quoteverify.cpp:527] Error, failed to load QvE.

Direct Dependencies