Abstract
This chapter first discusses the major functions that are provided by current message layers and then discusses the implementation of these functions that are affected by the services provided by the network hardware and the internal node architecture. These software overheads are determined by the functionality provided by the messaging layer and the hardware features that are available in the network and node interfaces to support this desired functionality. For example, the overhead in transferring message data from the local memory to the network interface is determined by the internal node design and the availability of services such as direct memory access (DMA), interrupts, or memory-mapped interfaces. The services of this messaging layer are the basis on which user-level message-passing libraries can be constructed. The chapter also focuses on the message-passing programming model and does not include issues concerning the support of shared-memory abstractions on distributed-memory architectures. The data transfer requirements for implementing shared-address spaces have similarities with the requirements of message-passing implementations, and several current research efforts are studying the integration of shared-memory and message-passing communication within a common framework.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
