ATOMIC: A Low-Cost, Very-High-Speed, Local Communication Architecture

@article{Cohen1993ATOMICAL,
  title={ATOMIC: A Low-Cost, Very-High-Speed, Local Communication Architecture},
  author={Danny Cohen and Gregory G. Finn and Robert E. Felderman and Annette L. DeSchon},
  journal={1993 International Conference on Parallel Processing - ICPP'93},
  year={1993},
  volume={1},
  pages={39-46}
}
  • D. Cohen, G. Finn, A. DeSchon
  • Published 16 August 1993
  • Computer Science
  • 1993 International Conference on Parallel Processing - ICPP'93
ATOMIC1 is an inexpensive O(gigabit) speed LAN built by USC/ISI. It is based upon Mosaic technol ogy developed for fine-grain, message-passing, massively parallel computation. Each Mosaic processor is capable of routing variable length packets, while providing added value through simultaneous computing and buffering. ATOMIC adds a general routing capability to the native Mosaic wormhole routing through store-and-forward. ATOMIC scales linearly, with a small interface cost. Each ATOMIC channel… 

Figures and Tables from this paper

Atomic: A High-Speed Local Communication Architecture
TLDR
The architecture and performance of ATOMIC is described, which adds a general routing capability to the native Mosaic wormhole routing through store-and-forward and path concatenation and is flexible both in topology and functionality.
Parallel computer architecture - a hardware / software approach
TLDR
This book explains the forces behind this convergence of shared-memory, message-passing, data parallel, and data-driven computing architectures and provides comprehensive discussions of parallel programming for high performance and of workload-driven evaluation, based on understanding hardware-software interactions.
Switches and switch interconnects
  • L.M. Ni, Wenjian Qiao, Mingyao Yang
  • Computer Science
    Proceedings of the Fourth International Conference on Massively Parallel Processing Using Optical Interconnections
  • 1997
TLDR
Major issues in switch design and switch interconnects are discussed, including cut-through switching, which promises short latency delivery and thus is well suited to distributed/parallel applications.
Incremental Design of Scalable Interconnection Networks Using Basic Building Blocks
TLDR
The original routing function can be adopted to send a message between any source and destination without using store-and-forward and causing deadlock and the way the network incrementally minimizes the amount of rewiring and keeps high bisection density and short diameter of the network.
PathFinder: A Pattern-Based Packet Classifier
TLDR
A performance study shows that the software implementation is about twice as fast as existing mechanisms, and that the hardware implementation is currently able to keep up with OC-12 network links and is likely to operate at gigabit speeds in the near future.
Operating System Support for the Desk Area Network
The paper presents a design overview of several of the ATM Devices produced within the Desk Area Network project and discusses the Operating Systems issues that they give rise to.
Incremental design of scalable interconnection networks using basic building blocks
  • Mingyao Yang, L. Ni
  • Computer Science
    Proceedings.Seventh IEEE Symposium on Parallel and Distributed Processing
  • 1995
TLDR
The original routing function can be adopted to send a message between any source and destination without using store-and-forward and causing deadlock and the way the network incrementally minimizes the amount of rewiring and keeps high bisection density and short diameter of the network.

References

SHOWING 1-10 OF 37 REFERENCES
A family of routing and communication chips based on the Mosaic
TLDR
The speciications, internal design, and performance characteristics of the latest series of Caltech Mesh-Routing Chips (MRCs), slack-protocol-conversion chips, and router-interface chips are described.
The design of nectar: a network backplane for heterogeneous multicomputers
TLDR
The motivation and goals for Nectar are presented and its hardware and software are described and the presentation emphasizes how the goals influenced the design decisions and led to the novel aspects of Nectar.
Sirpent: a high-performance internetworking approach
TLDR
Sirpent#8482; (Source Internetwork Routing Protocol with Extended Network Transfer) is described, a new approach to an internetwork architecture that makes source routing the basis for interconnection, rather than an option as in IP.
The design of the Caltech Mosaic C multicomputer
and Introduction The Caltech Mosaic C is an experimental, ne-grain multicomputer that employs single-chip nodes and advanced packaging technology to demonstrate the performance/cost advantages of the
An Atm Network Interface for High-Speed Experimentation
  • B. Davie
  • Computer Science
    IEEE Workshop on the Architecture and Implementation of High Performance Communication Subsystems
  • 1992
TLDR
An architecture which uses microprocessors for those functions where most flexibility is required, such as the execution of Segmentation and Reassembly (SAR) algorithms, while more specialized hardware is used for critical high-speed functions such as data movement and formatting of ATM cells is defined.
A high-performance host interface for ATM networks
TLDR
A host interface 1 for an asynchronous transfer mode (ATM) network comprises a Segmenter 2 and Reassembler 3, capable of reassembling an ATM cell in less than 2.7 microseconds.
Designing a practical ATM LAN
TLDR
It is shown that the ATM hardware and software components together provide services that are essential for ATM to be considered a realistic alternative to current shared-media LANs.
An integration of network communication with workstation architecture
A workstation may be thought of as a group of cooperatively connected subsystems. Point--to--point channels may be used to create a small--scale Gigabit LAN to which these subsystems are attached as
Building a practical HIPPI LAN
  • J. Renwick
  • Physics
    [1992] Proceedings 17th Conference on Local Computer Networks
  • 1992
It is demonstrated that an effective 800-Mb LAN can be built from HIPPI-SC switches, at a cost that is reasonable for large systems. The limitations of HIPPI cables can be overcome by the use of
HIPPI (high performance parallel interface)
  • J. Hughes
  • Computer Science
    [1992] Proceedings 17th Conference on Local Computer Networks
  • 1992
The HIPPI data channel and how it is used from a standard data channel, a standard internetworking, and a standard protocol viewpoint are described. Performance is also briefly discussed, as are some
...
...