The CLAM approach to multithreaded communication on shared-memory multiprocessors: design and experiments

Abstract

We present results on the experimental design and development of a Connectionless, Lightweight, and Multiway (CLAM) communications environment. The system provides efficient and scalable multiprotocol support for distributed applications that use multimodal data. We present motivation behind design decisions for the CLAM system, and describe two simple, but effective scheduling algorithms for the simultaneous support of multiple, threads-based user-space protocols. One algorithm is readily portable to shared-memory multiprocessors, and enables two or more protocols to coexist within an OS-level process. We present experimental results on the performance of both algorithms.

Scheduling communication in multithreaded programs: experimental results

“…CLAM). The CLAM [13,30] multiprotocol suite currently supports point-to-point connectionless transport, both unreliably as well as reliably. CLAM also provides protocol modules for reliable and unreliable multicast communication.…”

“…As mentioned earlier, we are not concerned with whether transmission is reliable or unreliable in this study of scheduling. Details on the implementation and performance of reliable protocols based on these scheduling algorithms can be found in [12,13]. Here, we examine the behavior of a single receive thread which attempts to run and retrieve incoming packets frequently enough to keep packet loss and latency low.…”

“…The result is an integrated programming environment where protocol and application threads operate within the same space, with low scheduling overheads. Details on the performance of the uniprocessor and multiprocessor thread-based protocols that are supported by our scheduling subsystem can be found in [12,13]. These results include detailed performance comparisons with other systems.…”

Efficient Multithreaded User-Space Transport for Network Computing: Design and Test of the TRAP Protocol

“…1). To avoid becoming heavy with cumbersome functionality, CLAM offers a plug-andplay methodology with its protocol suite [27]. It is layered on top of the Ariadne threads library [37] and the UDP protocol, but it can be ported to any best-effort communication system.…”

“…The design relies on a user-space threads system which supports timers, and on any best-effort delivery subsystem. 5 The TRAP protocol is a key component of the CLAM (connectionless, lightweight, and multiway) communications environment [27] for collaborative and scalable, distributed computing. The system offers user-level processes an integrated framework for multiple threadsbased protocols, multiple media and computational threads, for interoperability on heterogeneous networks.…”