High Speed System Area Networks (InfiniBand) - Challenges and solutions
نویسندگان
چکیده
Key issues/challenges/questions effecting IB and its adoption: • How will IB play with other IO interconnect and networking technologies (pcix, pci2.0, pciexpress, ethernet, fibre channel) • What role will IB play in future data centers. • What are IBs strengths and weaknesses compared to other interconnect technologies. • Why is the IB architecture so complex? • Some segments of the HPC community are interested in very large clusters (256 nodes and above). Are their special challenges building such clusters and how is the InfiniBand community addressing these?
منابع مشابه
HOMME and POPperf High Performance Applications: Optimizations for Scale
The High Order Method Modeling Environment (HOMME) and the modified version of The Parallel Ocean Program (POPperf) are two important applications for atmospheric and weather research. With an emphasis on efficiency, portability, maintainability and most importantly, scalability, HOMME and POPperf have been successfully deployed over the years on a wide variety of highperformance systems, such ...
متن کاملQueueing Network Models of Credit -Based Flow Control
Keywords--Forkjom queues, Bulk-arrival process, Credit-based flow control, Performance evaluation, InfiniBand, Fibre Channel. 1. I N T R O D U C T I O N Credit-based flow control management schemes are becoming increasingly popular in high-speed networks that span a local area. Fibre Channel [1], the technology used to realize storage area networks, employs credit-based methods for both end-to-...
متن کاملVIA over the CLAN Network
The Virtual Interface Architecture is an industry standard for high performance networking in system-area networks, and the same model is proposed for Infiniband. Existing implementations suffer from high complexity, and scaling to higher bandwidths and large numbers of endpoints is likely to be problematic. We present a novel implementation of VIA that consists of a thin software layer over th...
متن کاملDesign and Performance Evaluation of LiMIC (Linux Kernel Module for MPI Intra-node Communication) on InfiniBand Cluster
High performance intra-node communication support for MPI applications is critical for achieving the best performance out of clusters of SMP workstations. Although the performance of system area networks has improved in the recent years, intra-node communication still remains orders of magnitude faster than the network. Present day MPI stacks cannot make use of operating system kernel support f...
متن کاملSurvey Report for High-Speed Network Impact on DDBMS
DBMS as an I/O-intensive software has been focused on how to efficiently make use of cache, main memory, and disk storage based on the observation of their bandwidths in the memory hierarchy. In the past decade, Distributed Database Management Systems (DDBMS) has introduced computer network as another kind of data I/O media. Traditionally, computer networks were put on a lower layer than disk s...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2003