A spanning multichannel linked hypercube: a gradually scalable optical interconnection network for massively parallel computing

Abstract

Abstract-A new, scalable interconnection topology called the Spanning Multichannel Linked Hypercube (SMLH) is proposed. This proposed network is very suitable to massively parallel systems and is highly amenable to optical implementation. The SMLH uses the hypercube topology as a basic building block and connects such building blocks using two-dimensional multichannel links (similar to spanning buses). In doing so, the SMLH combines positive features of both the hypercube (small diameter, high connectivity, symmetry, simple routing, and fault tolerance) and the spanning bus hypercube (SBH) (constant node degree, scalability, and ease of physical implementation), while at the same time circumventing their disadvantages. The SMLH topology supports many communication patterns found in different classes of computation, such as bus-based, mesh-based, and tree-based problems, as well as hypercube-based problems. A very attractive feature of the SMLH network is its ability to support a large number of processors with the possibility of maintaining a constant degree and a constant diameter. Other positive features include symmetry, incremental scalability, and fault tolerance. It is shown that the SMLH network provides better average message distance, average traffic density, and queuing delay than many similar networks, including the binary hypercube, the SBH, etc. Additionally, the SMLH has comparable performance to other high-performance hypercubic networks, including the Generalized Hypercube and the Hypermesh. An optical implementation methodology is proposed for SMLH. The implementation methodology combines both the advantages of free space optics with those of wavelength division multiplexing techniques. A detailed analysis of the feasibility of the proposed network is also presented.

A class of highly scalable optical crossbar-connected interconnection networks (SOCNs) for parallel computing systems

“…Ultimately, though, such a system is limited to some fixed maximum number of wavelengths due to the fixed tuning range of the tunable optical source and/or the various optical parameters of the system. Wavelength reuse [24], [25] can be used to support larger system sizes by adding additional tunable optical sources at each processor. If the signals from each of these additional optical sources are kept optically isolated from each other, then the number of optical channels is multiplied by the number of tunable optical sources at each processor.…”

“…These network topologies include: a traditional Crossbar network (CB), the Binary Hypercube (BHC) [27], the Cube Connected Cycles (CCC) [28], the Torus [29], the Spanning Bus Hypercube (SBH) [30], and the Spanning Multichannel Linked Hypercube (SMLH) [25]. Each of these networks will be compared with respect to degree, diameter, number of links, bisection bandwidth, and average message distance.…”

All-optical crossbar switch using wavelength division multiplexing and vertical-cavity surface-emitting lasers

“…One way to extend a WDM optical interconnect beyond the limits of the number of channels supported is by wavelength reuse. 17,18 As the name implies, wavelength reuse optically isolates portions of the interconnect network to allow for the use of the same set of wavelengths simultaneously in multiple parts of the system for different purposes. One method of utilizing wavelength reuse to extend a WDM interconnected system is to connect various portions of the system by means of optically separated networks.…”

“…It is believed that this optical interconnect could be used to construct inexpensive optical interconnects for medium-scale parallel processors and that it could be extended through wavelength reuse 17,18 to support much larger highly parallel architectures.…”

Untitled

“…Other networks that have employed the positive features of the hypercube in combination with optics are, e.g., the Spanning Multichannel Linked Hypercube (SMLH) network [18], the Optical Transpose Interconnection System (OTIS) hypercube [19] and the GlobalLocal Optical Reconfigurable Interconnect (GLORI) network [20].…”