eSeMeshA: eBPF-based Service Mesh Acceleration for Cloud-Native Infrastructures

Service meshes are indispensable for enabling advanced communication features in cloud-native infrastructures. Though, due to the tight coupling of service meshes with the data path, they introduce significant overheads, resulting in degraded networking performance, including increased latency, poor throughput, and inefficient resource utilization. This paper presents an eBPF-based Service Mesh Acceleration framework, eSeMeshA, a proposal designed to mitigate these bottlenecks in intra-node communication by employing an in-kernel method to bypass costly data paths in modern service mesh architectures, such as Istio Ambient Mesh and Cilium, while also maintaining full support for legacy sidecar-based approaches, like Istio Sidecar. Experimental results show the potential of eSeMeshA to achieve significant gains, over 70% in throughput, up to 41.9% latency reduction, and 35.8% jitter improvement, all with minimal memory overhead of at most 30.4 MB and 23.2% less CPU usage, demonstrating its effectiveness as a scalable and efficient solution for cloud-native networking.