Title :
A Transport-Friendly NIC for Multicore/Multiprocessor Systems
Author :
Wu, Wenji ; DeMar, Phil ; Crawford, Matt
Author_Institution :
Comput. Div., Fermilab, Batavia, IL, USA
fDate :
4/1/2012 12:00:00 AM
Abstract :
Receive side scaling (RSS) is an NIC technology that provides the benefits of parallel receive processing in multiprocessing environments. However, RSS lacks a critical data steering mechanism that would automatically steer incoming network data to the same core on which its application thread resides. This absence causes inefficient cache usage if an application thread is not running on the core on which RSS has scheduled the received traffic to be processed and results in degraded performance. To remedy the RSS limitation, Intel´s Ethernet Flow Director technology has been introduced. However, our analysis shows that Flow Director can cause significant packet reordering. Packet reordering causes various negative impacts in high-speed networks. We propose an NIC data steering mechanism to remedy the RSS and Flow Director limitations. This data steering mechanism is mainly targeted at TCP. We term an NIC with such a data steering mechanism “A Transport-Friendly NIC” (A-TFN). Experimental results have proven the effectiveness of A-TFN in accelerating TCP/IP performance.
Keywords :
cache storage; local area networks; multi-threading; multiprocessing systems; network interfaces; parallel processing; transport protocols; Ethernet flow director technology; NIC data steering mechanism; NIC technology; RSS limitation; TCP/IP performance; application thread; cache usage; critical data steering mechanism; high-speed network; multicore system; multiprocessing environment; multiprocessor system; network interface card; packet reordering; parallel receive processing; receive side scaling; received traffic; transport-friendly NIC; Context; IP networks; Instruction sets; Linux; Multicore processing; Sockets; Synchronization; Network interface card; TCP/IP.; high performance networking; multicore; protocol stack;
Journal_Title :
Parallel and Distributed Systems, IEEE Transactions on
DOI :
10.1109/TPDS.2011.195