“随着物联网技术的不断发展，边缘计算架构越来越受到重视。相比传统云计算架构，边缘计算架构将数据处理与存储分布在离物联网终端设备更近的位置，能够有效减少数据在传输过程中的延迟与减少对云端资源的依赖。在物联网设备的不断增加的背景下，越来越多的物联网设备从边缘计算网关处接入，更多的网络请求需要在边缘计算网关处负载均衡给边缘计算集群处理。因此本文对边缘计算网关接入与负载均衡能力提升的问题展开了研究，主要研究成果如下：（1）针对传统Linux内核协议栈处理网络数据包性能受到限制问题，从软件层面使用...-DPDK中文网

With the continuous development of IoT technology, edge computing architecture is increasingly valued. Compared with traditional cloud computing architecture, edge computing architecture distributes data processing and storage closer to IoT terminal devices, which can effectively reduce the latency of data transmission and dependence on cloud resources. In the context of the increasing number of IoT devices, more and more IoT devices are accessing through edge computing gateways, and more network requests need to be load balanced at the edge computing gateway for processing by the edge computing cluster. Therefore, this article researched the improvement of the access and load balancing capabilities of the edge computing gateway, and the main research results are as follows:

(1) To address the issue of limited performance of traditional Linux kernel protocol stack processing network packets, we optimized using DPDK technology at the software level. A user-mode protocol stack based on DPDK technology was designed to solve the problem that DPDK technology does not provide a complete protocol stack. The protocol stack first directly obtains data from the network card through polling and stores it in user space. It then uses a lock-free queue to pass data and synchronize the order between the data and control modules. Finally, the design of the EPOLL interface method for multi-I/O multiplexing improves the convenience of porting various Linux network applications to this protocol stack. Efficient and portable access capabilities of the edge computing gateway have been achieved.

(2) To address the performance defects of centralized polling traversal of dynamically weighted load balancing algorithms, we proposed a distributed dynamic weighted load balancing method based on DPDK. This method uses DPDK’s user-mode protocol stack to accelerate the data processing capabilities of the edge computing gateway and edge computing nodes. The edge computing cluster nodes push the load weights according to demand, reducing the bandwidth and performance loss of the edge computing gateway. Additionally, a method based on information entropy is used to quickly detect fluctuations in the computing resources of edge computing nodes, providing threshold support for pushing strategies to edge computing nodes. The load balancing strategy of the edge computing gateway can accurately match the computing resource situation of edge computing nodes, realizing the improvement of the load balancing ability of the edge computing gateway.

(3) We designed and implemented an access and load balancing system based on DPDK protocol stack. Performance testing shows that the proposed improvement methods are reasonable and effective, which can effectively improve the network packet processing capability of the edge computing layer.