Kernel TCP Virtual Server is a Layer-7 switching software in the Linux Virtual Server framework. It can do application-level load balancing (such as content-based scheduling) inside the Linux kernel. The reason for implementing it inside the kernel is to avoid the overhead of context switching and memory copying between user-space and kernel-space. Although the its scalability is lower than that of IPVS (IP Virtual Server), it is flexible, because the content of the request is known before the request is redirected to the real server.
|Tags||Clustering/Distributed Networks Operating System Kernels Linux Networking|
Release Notes: This version includes one fix for a sock lock issue, which probably only appeared on SMP platforms.
Release Notes: Timeout support was added to the connection handler. The bug in the --listen command option of tcpvsadm was fixed.
Release Notes: This version adds a new HTTP scheduler which sends requests to the server based on the hash value of matched request, and adds a --listen option in the tcpvsadm program.
Release Notes: tcp_vs_phttp.c has been cleaned. tcp_vs_http_trans.h and tcp_vs_http_parser.h are spinoffs from tcp_vs.h.
Release Notes: HTTP transport code was split into tcp_vs_http_trans.c. A large part of the HTTP parsing engine was rewritten, and some tuning was made in tcp_vs_chttp.c and tcp_vs_phttp.c. Issues with calling spin_lock and spin_unlock were fixed.