Programming socket-independent network functions with nethuns

Abstract

Software data planes running on commodity servers are very popular in real deployments. However, to attain top class performance, the software approach requires the adoption of accelerated network I/O frameworks, each of them characterized by its own programming model and API. As a result, network applications are often closely tied to the underlying technology, with obvious issues of portability over different systems. This is especially true in cloud scenarios where different I/O frameworks could be installed depending on the configuration of the physical servers in the infrastructure. The nethuns library proposes a unified programming abstraction to access and manage network operations over different I/O frameworks. The library is freely available to the community under the BSD license and currently supports AF_XDP and netmap for fast packet handling along with the classic AF_PACKET and the pcap library. Network applications based on nethuns need only to be re-compiled to run over a different network API. The experiments prove that the overhead introduced by nethuns is negligible, hence making it a convenient programming platform that eases the coding process while guaranteeing high performance and portability. As proofs of concept, a handy traffic generator as well as the popular Open vSwitch application have been successfully ported and tested over nethuns.

Empowering Cloud Computing With Network Acceleration: A Survey

“…The main difference among these works is that they propose to build this new interface layer at different locations: in kernel space (LITE [120]) or in userspace (X-RDMA [121], nethuns [112], INSANE [113], and Demikernel [111]).…”

“…2) User-level interfaces: X-RDMA [121], nethuns [112], INSANE [113], and Demikernel [111] propose a userspace interface to different network acceleration techniques, both hardware and software based, with the ultimate goal to enable the transparent portability of applications across heterogeneous technologies. The key design principle of these new interfaces is the minimization of network concepts and primitives, to simplify system programming in large-scale production environments.…”

“…Similarly to LITE, X-RDMA focuses on RDMA and has a built-in support for RPC and for a messaging interface. Nethuns [112] and INSANE [113] share the same design principles but with the goal of easing the access to a wider range of acceleration technologies, including also DPDK and XDP, through a unified minimal interface, still expressive enough to enable the efficient implementation of higher-level domainspecific abstractions. Although these works do not directly provide a solution for an efficient virtualization of accelerated I/O, they explicitly address this issue by easing the requirements on the virtual device they require (see Section VI-A).…”

“…From a provider perspective, nearly all the considered contributions require the availability of a virtual accelerated NIC, or its equivalent for software-based techniques, which we discuss in Section VI as an open research challenge. Notable exceptions include nethuns [112] and INSANE [113], which ease these requirements and can work with regular hardwarebased virtualization (see Section II-D1). VSocket [57] and SocksDirect [58] take a step further and propose an integrated solution for I/O virtualization, as discussed in the next Section.…”

“…We begin by observing that many interfaces for accelerated networking (Section V) assume the availability of a virtual accelerated NIC (e.g., a virtual RDMA NIC), or its equivalent for software-based techniques (e.g., a special PMD for DPDK). Such assumption has a significant impact on both the I/O virtualization and serviceability dimensions: if applications did not need direct visibility of the NIC, providers would be able to either use standard solutions or design more efficient ones to support accelerated cloud networking [57], [112], [113].…”