A Host Interface Architecture for High-Speed Networks

This paper describes a new host interface architecture for high-speed networks operating at 800 of Mbit/second or higher rates. The architecture is targeted to achieve several 100s of Mbit/second application-to-application performance for a wide range of host architectures. The architecture achieves the goal by providing a streamlined execution environment for the entire path between host application and network interface. In particular, a “Communication Accelerator Block” (CAB) is used to minimize data copies, reduce host interrupts, support DMA and hardware checksumming, and control network access. This host architecture is applicable to a large class of hosts with high-speed I/O busses. Two implementations for the 800 Mbit/second HIPPI network are under development. One is for a distributed-memory supercomputer (iWarp) and the other is for a high-performance workstation (DECstation 5000). We describe and justify both implementations. Keyword Codes: B.4.1; C.2.1

[1]  Shekhar Y. Borkar,et al.  iWarp: an integrated solution to high-speed parallel computing , 1988, Proceedings. SUPERCOMPUTING '88.

[2]  H. T. Kung,et al.  The design of nectar: a network backplane for heterogeneous multicomputers , 1989, ASPLOS III.

[3]  Samuel J. Leffler,et al.  The design and implementation of the 4.3 BSD Unix operating system , 1991, Addison-Wesley series in computer science.

[4]  Mark J. Karol,et al.  Queueing in high-performance packet switching , 1988, IEEE J. Sel. Areas Commun..

[5]  David R. Cheriton,et al.  The VMP network adapter board (NAB): high-performance network communication for multiprocessors , 1988, SIGCOMM '88.

[6]  David A. Borman,et al.  Computing the internet checksum , 1989, CCRV.

[7]  Armando P. Stettner.  The design and implementation of the 4.3BSD UNIX operating system , 1988 .

[8]  L. W. Tucker,et al.  Architecture and applications of the Connection Machine , 1988, Computer.

[9]  J. Lumley,et al.  A high-throughput network interface to a RISC workstation , 1992, IEEE Workshop on the Architecture and Implementation of High Performance Communication Subsystems.

[10]  Robbert van Renesse,et al.  Amoeba A Distributed Operating System for the 1990 s Sape , 1990 .

[11]  Bob Beach.  UltraNet: an architecture for gigabit networking , 1990, [1990] Proceedings. 15th Conference on Local Computer Networks.

[12]  D. Farber.  Gigabit Network Testbeds , 1990, Computer.

[13]  Matthew Mathis,et al.  Deployment of a Hippi-Based Distributed Supercomputing Environment at the Pittsburgh Supercomputing Center , 1992, Proceedings. Workshop on Heterogeneous Processing.

[14]  Alistair Cockburn.  Efficient implementation of the OSI transport protocol checksum algorithm using 8/16-bit arithmetic , 1987, CCRV.

[15]  David A. Mosher,et al.  User-Process Communication Performance in Networks of Computers , 1988, IEEE Trans. Software Eng..

[16]  Peter Steenkiste,et al.  Analyzing communication latency using the Nectar communication processor , 1992, SIGCOMM '92.

[17]  G. Chesson,et al.  Protocol engine design , 1988 .

[18]  H. T. Kung,et al.  Parallelizing a new class of large applications over high-speed networks , 1991, PPOPP '91.

[19]  Computer Staff.  Gigabit network testbeds , 1990, Computer.

[20]  Michael Burrows,et al.  Performance of Firefly RPC , 1990, ACM Trans. Comput. Syst..

[21]  A.E. Klietz,et al.  Network supercomputing , 1992, IEEE Network.

[22]  David R. Cheriton,et al.  The VMP network adapter board (NAB): high-performance network communication for multiprocessors , 1988, SIGCOMM 1988.

[23]  Brian Zill,et al.  Protocol implementation on the Nectar Communication Processor , 1990, SIGCOMM '90.

[24]  David A. Borman,et al.  High speed networking at Cray research , 1991, CCRV.

[25]  Michael Burrows,et al.  Performance of Firefly RPC , 1989, SOSP '89.