FSM.
FSM (Flomesh Service Mesh) is a lightweight service mesh for cloud-edge integration. It leverages OSM (Open Service Mesh) as the control plane and Pipy as the data plane. It features high performance, low resource, simplicity, ease of use, scalability and broad compatibility with diverse architectures including x86, ARM64, Loongson and RISC-V.
Based on the OSM control plane, FSM supports full SMI (Service Mesh Interface) specification. When being used in conjunction with ErieCanal, which brings in the Ingress/Gateway API and cross-cluster service discovery, the entire "FSM + ErieCanal" suite provides you a complete intra-Kubernetes-cluster and multi-cluster solution for both east-west and north-south traffic management and service governance.
High performance, low resources.
Data plane proxy Pipy is developed in C++. It uses an asynchronous network framework; allocated resources are pooled and recycled. Provides high performance with a shallow resource footprint, suitable for large-scale, high-density cloud scenarios and edge scenarios with limited processing power and performance.
Multi-Cluster.
With ErieCanal, which supports Ingress, Gateway API, and cross-cluster service discovery, it provides Kubernetes intra-cluster + multi-cluster "east-west + north-south" traffic management and service governance capability. It enables cross-cluster interoperability and traffic scheduling to solve the traffic management challenges in multi-cluster use cases.
ARM Ready.
Supports x86, ARM64, Longxin, and RISC-V computing architectures to improve hardware compatibility and meet the needs of diverse computing scenarios.
Easy to scale.
The data plane uses a programmable proxy Pipy and a built-in JavaScript engine, which allows customization and extension of the service mesh functionality using JS syntax scripts.
The control plane is based on OSM (Open Service Mesh) developed by Go, and the open design makes it easy to extend the functionality.
Observability.
Generates comprehensive telemetry data for network traffic within the mesh, providing non-intrusive observability for application containers, including metrics, distributed traces, and logs. Easily customize various metrics and provide multiple log collection and processing methods through the ease of scalability of the data plane.
Visual display of telemetry data is provided through the console.
Multi-protocol Support.
Pipy natively supports HTTP1, HTTP2/gRPC, TCP, UDP, MQTT, Dubbo, Thrift, and other protocols; the control plane currently supports HTTP1, HTTP2/gRPC, TCP, UDP traffic management, and more protocols can be supported through extensions.
Support for SMI.
SMI (Service Mesh Interface) provides a standard interface to the Kubernetes Service Mesh, a basic set of features for standard Service Mesh scenarios, the flexibility to provide new Service Mesh features, and a space for innovation in the Service Mesh technology ecosystem that is easy to use.
Dashboard.
Provides a powerful console that facilitates managing the Service Mesh and configuring service governance features in addition to rich data presentation.