NFS Gains "Extreme" Performance Boost with LOCALIO Protocol in Linux 6.12
In the Linux 6.12 development cycle, significant updates to the Network File System (NFS) were introduced. These updates focus on optimizing performance when the client and server are on the same machine. The key addition is the LOCALIO protocol extension, which allows NFS to detect when both the client and server share a host. It then bypasses traditional network protocols for certain operations.
Extreme Performance Improvement with LOCALIO
The LOCALIO protocol offers remarkable performance improvements. It eliminates the need for RPC (Remote Procedure Calls) and XDR (External Data Representation) encoding during read, write, and commit operations. By removing these steps, systems experience faster data access, especially in containerized environments where the NFS client and server run together on the same server.
Documentation shows that performance gains can exceed 10x for 4K reads with multiple libaio threads. This makes LOCALIO a strong option for high-performance computing tasks.
Benefits for Containerized Workloads
LOCALIO shines in setups where containerized workloads run with an NFS server. By allowing direct file access, it reduces overhead. Storage operations become more efficient in environments where the NFS server provides storage on the same machine.
Other NFS Improvements in Linux 6.12
Linux 6.12 also introduces the noalignwrites mount option. This option helps prevent lock-less lost writes. The update includes several bug fixes and cleanups to enhance NFS performance and reliability. Together, these improvements make NFS in Linux 6.12 highly optimized.
Click below and ‘share’ this article!