Abstract:

The urgent requirement for the highefficiency of routing protocols on the Internet will be satisfied by exploiting the threadlevelparallelism of BGP on multicores. A multithreaded BGP model is proposed based on neighbor session division, and it achieves the parallel processing of different sessions on multiple threads by the use of data parallelism. Firstly, the framework of the multithreaded BGP is put forward, and is then implemented based upon Quagga BGP. The experimental results on the Intel Xeon quadcore server show that, the multithreaded BGP yields the speedup of 2.73 in comparison with BGP, and can make full use of the computing resources on multicores to provide more performance improvement for the protocol.

Key words: multicore;BGP;multithreaded;parallelism