Slim and Fast: Low-Overhead Container Overlay Network With Fast Connection Setup

Abstract

Large-scale cloud applications today are often deployed using multiple containers, and container overlay networks is the de facto method to provide connectivity among these containers. However, the existing tunneling-based overlay network incurs significant performance overhead due to the need of transformation for every packet. Recent work <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Slim</i> , through manipulating connection-level meta-data, allows containers to use host OS sockets directly thus they can achieve good performance without extra packet tunneling. Nevertheless, the connection setup is significantly slowed down, which requires an extra round-trip communication between both sides to pass the mapping information of the host OS socket and the container socket. This greatly hurts the performance of many cloud applications that must process short connections at high speed. We propose <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SlimFast</i> , a low-overhead container overlay network which provides a fast connection setup. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SlimFast</i> directly uses the host OS socket for container communication as <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Slim</i> . However, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SlimFast</i> needs no extra communication during connection setup. We reserve a dedicated host port for the container overlay network and use socket mapping table to locally find the right container socket during connection setup. We implement <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SlimFast</i> which is compatible with existing container applications. Experiments show that, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SlimFast</i> can improve the connection setup time by about 2.1x compared with <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Slim</i> , meanwhile maintaining low-overhead during data transmission as <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Slim</i> . This brings significant performance improvement to real applications. Particularly, testbed results show that <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SlimFast</i> improves the throughput of Nginx proxy and Memcached by about 0.9x and 2.2x, respectively.