140 likes | 301 Views
Fast Forwarding Table Lookup Exploiting GPU Memory Architecture. Date:100/12/14 Publisher:IEEE ICTC 2010 Author:Youngjun Lee,Minseon Jeong,Sanghwan Lee,Eun -Jin Im Kookmin University. Introduction.
E N D
Fast Forwarding Table Lookup Exploiting GPU Memory Architecture Date:100/12/14 Publisher:IEEE ICTC 2010 Author:YoungjunLee,MinseonJeong,SanghwanLee,Eun-Jin Im Kookmin University
Introduction • To speed up the software router processing time,PacketShader employers GPU to exploit the extensive parallelism that GPU can provide. • GPU contains a large number of small ALUs so that we can run the same number of threads as the number of ALUs
Introduction • Three type memory in GPU 1.GlobalMemory:highestbandwidth,largesize,but when two threads access the same location,the memory conflict should be resolved. 2.ConstantMemory and Texture Memory do not show this problem and their access time is shorter,but size is too small to maintain forwarding table.
Introduction • A new way of forwarding table lookup mechanism,which exploits the constant memory so that we can speed up the table lookup procedure.
Fast Forwarding Table Lookup • Maintain the whole forwarding table in the global memory of GPU. • Create indices in the constant memory(64KB).
Fast Forwarding Table Lookup • Currently the number of active BGP entries is about 340,000. • The number of prefixes with length of 16 or less is 4.18%,24 is about 52%.Thus we expand all the prefixes of length less than 24 into multiple /24 prefixes. • The number of prefixes is still less than one million-20 bits.
Fast Forwarding Table Lookup Search:13.1.9.11