Several research contributions and commercial ventures related to wormhole routing, a switching technique used in direct networks, are discussed. The properties of direct networks are reviewed, and the operation and characteristics of wormhole routing are discussed in detail. By its nature, wormhole routing is particularly susceptible to deadlock situations, in which two or more packets may block one another indefinitely. Several approaches to deadlock-free. routing, along with a technique that allows multiple virtual channels to share the same physical channel, are described. In addition, several open issues related to wormhole routing are discussed. >

A survey of wormhole routing techniques in direct networks

A senior undergraduate or graduate level computer science textbook on algorithm design for distributed computer systems.

/pdf/an-introduction-to-distributed-algorithms-2scptn142v.pdf

An Introduction to Distributed Algorithms

More and more scientists and engineers are becoming interested in using supercomputers. Earlier barriers to using these machines are disappearing as software for their use improves. Meanwhile, new parallel supercomputer architectures are emerging that may provide rapid growth in performance. These systems may use a large number of processors with an intricate memory system that is both parallel and hierarchical; they will require even more advanced software. Compilers that restructure user programs to exploit the machine organization seem to be essential. A wide range of algorithms and applications is being developed in an effort to provide high parallel processing performance in many fields. The Cedar supercomputer, presently operating with eight processors in parallel, uses advanced system and applications software developed at the University of Illinois during the past 12 years. This software should allow the number of processors in Cedar to be doubled annually, providing rapid performance advances in the next decade.

Parallel Supercomputing Today and the Cedar Approach (並列処理ハ-ドウェアと言語特集)

Efficient routing of messages is a key to the performance of multicomputers Multicast communication refers to the delivery of the same message from a source node to an arbitrary number of destination nodes While multicast communication is highly demanded in many applications, most of the existing multicomputers do not directly support this service; rather it is indirectly supported by multiple one-to-one or broadcast communications, which result in more network traffic and a waste of system resources The authors study routing evaluation criteria for multicast communication under different switching technologies Multicast communication in multicomputers is formulated as a graph theoretical problem Depending on the evaluation criteria and switching technologies, they study three optimal multicast communication problems, which are equivalent to the finding of the following three subgraphs: optimal multicast path, optimal multicast cycle, and minimal Steiner tree, where the interconnection of a multicomputer defines a host graph They show that all these optimization problems are NP-complete for the popular 2D-mesh and hypercube host graphs Heuristic multicast algorithms for these routing problems are proposed >

Multicast Communication in Multicomputer Networks.

A fault-tolerant communication scheme that facilitates near-optimal routing and broadcasting in hypercube computers subject to node failures is described. The concept of an unsafe node is introduced to identify fault-free nodes that may cause communication difficulties. It is shown that by only using 'feasible' paths that try to avoid unsafe nodes, routing and broadcasting can be substantially simplified. A computationally efficient routing algorithm that uses local information is presented. It can route a message via a path of length no greater than p+2, where p is the minimum distance from the source to the destination, provided that not all nonfaulty nodes in the hypercube are unsafe. Broadcasting can be achieved under the same fault conditions with only one more time unit than the fault-free case. The problems posed by deadlock in faulty hypercubes are discussed, and deadlock-free implementations of the proposed communication schemes are presented. >

A fault-tolerant communication scheme for hypercube computers

It was long a truism in computer systems design that balanced systems achieve the best performance Message passing parallel processors are no different To quantify the balance of a hypercube design, an experimental methodology was developed and the associated suite of benchmarks was applied to several existing hypercubes The benchmark suite includes tests of both processor speed in the absence of internode communication and message transmission speed as a function of communication patterns

Benchmarking hypercube hardware and software

Improvements in the performance of communication networks will soon alter the criteria for multicomputers efficient algorithm design. To study the performance of existing and proposed network implementations, we conducted parametric simulation studies of several message routing paradigms, assuming an equivalent hardware implementation. Using the results of these simulations, we have developed qualitative analytic models that provide insight into network behavior.

Networks for parallel processors: measurements and prognostications

Any new computer system organization raises many questions about hardware, system software, algorithms, and programming; multicomputers are no exception. The commercial emergence of hypercubes has made it possible to move many research questions from theoretical to experimental venues. However, hardware availability does more than permit implementation of previously tested ideas. The feasibility of many application and operating system algorithms can only be determined experimentally. We believe it is difficult, if not impossible, to experimentally examine these issues singly — they interact in both obvious and subtle ways. This paper describes the Picasso hypercube operating system, a part of the Picasso project at the University of Illinois. Motivated by the need for a flexible operating system testbed, the design of Picasso has exposed several important, and necessary, features of multicomputer operating systems designed for research. We present the design and current performance of Picasso, discuss the lessons learned, and conclude with an overview of future research plans.

Dirk C. Grunwald

Papers

Benchmarking hypercube hardware and software

Networks for parallel processors: measurements and prognostications

Picasso: an experiment in hypercube operating system design