Inter-multi-core buffer dynamic migration method and device based on DPDK

Abstract

The invention discloses an inter-multi-core buffer dynamic migration method and device based on DPDK. Each CPU core is correspondingly provided with a buffer cache pool, a plurality of buffer cache pools share one buffer sharing pool, each buffer cache pool is provided with a shared waterline, and a buffer below the shared waterline is used as a shareable buffer; when any CPU core applies for a buffer, if the residual buffer amount in the buffer cache pool of the core does not meet the application and the available buffer amount in the buffer sharing pool is insufficient, inter-core buffer migration is triggered, that is, buffer locks of other CPU cores are obtained, then shareable buffers in buffer cache pools of other CPU cores are obtained, the buffer resource dynamic migration is achieved, and the buffer resource utilization rate is increased.

Description

technical field [0001] The invention belongs to the technical field of data communication, and more specifically relates to a DPDK-based multi-core buffer dynamic migration method and device. Background technique [0002] The forwarding platform of traditional network equipment is built based on dedicated network chips, such as cavium's cn78xx network processor and the private NP of various communication manufacturers. The technical solution based on dedicated network chips leads to a high degree of hardware dependence, and the flexibility of business development is limited by the hardware, which further results in longer business development cycles, delayed market delivery cycles, and further increased product development costs. Data Plane Development Kit (DPDK for short) technology is based on general-purpose processors (such as x86 and arm) to build a multi-core soft forwarding platform. This technical solution not only guarantees data processing performance and throughpu...

AI Technical Summary

Problems solved by technology

[0008] During use, when the buffer in the buffer_cache of a core is insufficient or empty, there may still be a certain amount of idle buffer in the buffer_cache of other cores, that is, the buffer resources may be unevenly distributed among the buffer_cache of each core.

However, since the buffer_cache is private to each core, that is, each core can only obtain buffer resources from the buffer_cache of the core, so when the buffer resources in the buffer_cache of a core are tight, and there are not enough buffer resources in the shared pool mempool, even The remaining cores have buffer surplus, and this core cannot apply for buffers from the buffer_cache of other cores for forwarding.

This allocation management defect leads to insufficient effective utilization of buffer resources. When buffer resources are available, the core will still lose packets due to insufficient buffers, resulting in service interruptions and increasing the stability and reliability of the forwarding plane.

Images