/eRPC

Efficient RPCs for datacenter networks

Primary LanguageC

eRPC is a fast and general-purpose RPC library for datacenter networks. Our NSDI 2019 paper describes the system in detail. Documentation is available online.

Some highlights:

  • Multiple supported networks: Ethernet, InfiniBand, and RoCE
  • Low latency: 2.3 microseconds round-trip RPC latency with UDP over Ethernet
  • Performance for small 32-byte RPCs: ~10M RPCs/sec with one CPU core, 60--80M RPCs/sec with one NIC.
  • Bandwidth for large RPC: 75 Gbps on one connection (one CPU core at server and client) for 8 MB RPCs
  • Scalability: 20000 RPC sessions per server
  • End-to-end congestion control that tolerates 100-way incasts
  • Nested RPCs, and long-running background RPCs
  • A port of Raft as an example. Our 3-way replication latency is 5.3 microseconds with traditional UDP over Ethernet.

Requirements

  • Toolchain: A C++11 compiler and CMake 2.8+
  • See scripts/packages/ for required software packages for your distro. Install exactly one of the following, mutually-incompatible packages:
    • Mellanox OFED for Mellanox NICs
    • For other DPDK-compatible NICs, a system-wide installation from the latest DPDK sources (e.g., sudo make install T=x86_64-native-linuxapp-gcc DESTDIR=/usr).
  • NICs: Fast (10 GbE+) bare-metal NICs are needed for good performance. eRPC works best with Mellanox Ethernet and InfiniBand NICs. Any DPDK-capable NICs also work well. Slower/virtual NICs can still be used for testing and development.
  • System configuration:
    • At least 1024 huge pages on every NUMA node, and unlimited SHM limits
    • On a machine with n eRPC processes, eRPC uses kernel UDP ports {31850, ..., 31850 + n - 1}. These ports should be open on the management network. See scripts/firewalld/erpc_firewall.sh for systems running firewalld.

eRPC quickstart

  • Build and run the test suite: cmake . -DPERF=OFF -DTRANSPORT=infiniband; make -j; sudo ctest.
    • DPERF=OFF enables debugging, which greatly reduces performance. Set DPERF=ON for performance measurements.
    • Here, infiniband should be replaced with raw for Mellanox Ethernet NICs, or dpdk for Intel Ethernet NICs.
    • A machine with two ports is needed to run the unit tests if DPDK is chosen. Run scripts/run-tests-dpdk.sh instead of ctest.
  • Run the hello_world application:
    • cd hello_world
    • Edit the server and client hostnames in common.h
    • Based on the transport that eRPC was compiled for, compile hello_world using make infiniband, make raw, or make dpdk.
    • Run ./server at the server, and ./client at the client
  • Generate the documentation: doxygen

Supported bare-metal NICs:

  • Ethernet/UDP mode:
    • ConnectX-4 or newer Mellanox Ethernet NICs: Use DTRANSPORT=raw
    • DPDK-compatible NICs that support flow-director: Use DTRANSPORT=dpdk
      • Intel 82599 and Intel X710 NICs have been tested
      • Virtual NICs have not been tested
      • raw transport is faster for Mellanox NICs, which also support DPDK
    • ConnectX-3 Ethernet NICs are supported in eRPC's RoCE mode
  • RDMA (InfiniBand/RoCE) NICs: Use DTRANSPORT=infiniband. Add DROCE=on if using RoCE.
  • Mellanox drivers optimized specially for eRPC are available in the drivers directory

Running eRPC without fast bare-metal NICs:

  • Follow these instructions to try out eRPC on a machine without fast Mellanox or DPDK-capable NICs (e.g., on your desktop or in a virtual machine). This is for development only: eRPC is not designed to perform well in these settings. eRPC has been tested on KVM virtual machines and in Amazon EC2.
  • Create an emulated RoCE device with [SoftRoCE] (instructions soon).
  • Compile eRPC with DTRANSPORT=infiniband -DROCE=on

Configuring and running the provided applications

  • The apps directory contains a suite of benchmarks and examples. The instructions below are for this suite of applications. eRPC can also be simply linked as a library instead (see hello_world/ for an example).
  • To build an application, create scripts/autorun_app_file and change its contents to one of the available directory names in apps/. See scripts/example_autorun_app_file for an example. Then generate a Makefile using cmake . -DPERF=ON -DTRANSPORT=raw/infiniband/dpdk.
  • Each application directory in apps/ contains a config file that must specify all flags defined in apps/apps_common.h. For example, num_processes specifies the total number of eRPC processes in the cluster.
  • The URIs of eRPC processes in the cluster are specified in scripts/autorun_process_file. Each line in this file must be <hostname> <management udp port> <numa_node>.
  • Run scripts/do.sh for each process:
    • With single-CPU machines: num_processes machines are needed. Run scripts/do.sh <i> 0 on machine i in {0, ..., num_processes - 1}.
    • With dual-CPU machines: num_machines = ceil(num_processes / 2) machines are needed. Run scripts/do.sh <i> <i % 2> on machine i in {0, ..., num_machines - 1}.
  • To automatically run an application at all processes in scripts/autorun_process_file, run scripts/run-all.sh. For some applications, statistics generated in a run can be collected and processed using scripts/proc-out.sh.

Getting help

  • GitHub issues are preferred over email. Please include the following information in the issue:
    • NIC model
    • Mellanox OFED or DPDK version
    • Operating system

Contact

Anuj Kalia (akalia@cs.cmu.edu)

License

	Copyright 2018, Carnegie Mellon University

    Licensed under the Apache License, Version 2.0 (the "License");
    you may not use this file except in compliance with the License.
    You may obtain a copy of the License at

        http://www.apache.org/licenses/LICENSE-2.0

    Unless required by applicable law or agreed to in writing, software
    distributed under the License is distributed on an "AS IS" BASIS,
    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    See the License for the specific language governing permissions and
    limitations under the License.