Partition-Aware Packet Steering Using XDP and eBPF for Improving Application-Level Parallelism

Abstract

A single CPU core is not fast enough to process packets arriving from the network on commodity NICs. Applications are therefore turning to application-level partitioning and NIC offload to exploit parallelism on multicore systems and relieve the CPU. Although NIC offload techniques are not new, programmable NICs have emerged as a way for custom packet processing offload. However, it is not clear what parts of the application should be offloaded to a programmable NIC for improving parallelism. We propose an approach that combines application-level partitioning and packet steering with a programmable NIC. Applications partition data in DRAM between CPU cores, and steer requests to the correct core by parsing L7 packet headers on a programmable NIC. This approach improves request-level parallelism but keeps the partitioning scheme transparent to clients. We believe this approach can reduce latency and improve throughput because it utilizes multicore systems efficiently, and applications can improve partitioning scheme without impacting clients. CCS CONCEPTS • Software and its engineering → Operating systems; Communications management; Multiprocessing / multiprogramming / multitasking.

XDP-Based SmartNIC Hardware Performance Acceleration for Next-Generation Networks

“…Nevertheless, none of the previous work has been implemented or tested in hardware offloading mode. Enberg et al [34] propose a combined application and hardware packet steering implementation using eBPF and XDP. It provides a practical approach for accelerating network-intensive applications.…”

Treehouse: A Case For Carbon-Aware Datacenter Software

“…Although more work is needed to understand how best to integrate these technologies with the kernel, studies have shown that operating system overheads can be slashed while still providing traditional kernel functions such as centralized scheduling, file system semantics, and performance isolation [37,49,62]. A complementary approach is to move user-defined functions written in a type-safe language into the Linux kernel, to allow customization closer to the hardware [18,24,48,64].…”

BPF for storage

“…With the rise of fast networks, Berkeley Packet Filter (BPF) has gained in popularity for efficient packet processing since it eliminates the need to copy each packet into userspace; instead, applications can safely operate on packets in the kernel. Common networking use cases include filtering packets [4,5,25,33], network tracing [2, 3, 32], load balancing [4,15], packet steering [30] and network security checks [12]. It has also been used as a way to avoid multiple network crossings when accessing disaggregated storage [38].…”