Distance Interconnects Research Articles

Performance Trade-Offs for Wavelength Striping Optical Switching Using a Novel Star Architecture

This work describes various performance trade-offs that arise from the use of a technique for optical switching under various network topologies. Such switching operation can be summarized as follows: (a) user data are divided into fixed-length fragments, (b) each fragment is assigned to a different wavelength, and (c) all wavelengths are simultaneously switched to the egress links. This concept of dividing user data into several wavelengths to be simultaneously switched is called wavelength striping and its purpose is to reduce latency and increase throughput for short distance interconnects. We depart from previous work where a building block implementing this basic switching function has been built around semiconductor optical amplifiers (SOAs). In this paper, we investigate diverse trade-offs that arise from the use of this switching approach in different network topologies. One of the main issues addressed in this paper is the relation between cascadability and bit error rate (BER). In this case, our results indicate that a switch fabric can cascade up to five stages without exceeding a BER of 10−9 and without incurring in power budget problems. We also show that the performance degradation, introduced by cascading SOAs, can be compensated with a star interconnect architecture that is introduced. Other issues addressed in this paper are the effect of scalability on cost and the effect of latency on TCP performance and reliability.

Encircled Angular Flux Representation of the Modal Power Distribution and Its Behavior in a Step Index Multimode Fiber

The optical properties of inline optical components inserted at the input to a multimode optical fiber (MMF) strongly affect the propagating modal power distribution (MPD). To realize stable systems with predictable reproducible performance and to encourage widespread use of MMF systems, improvements to the system design process and to the characterization process need to be made and new interface standards need to be defined. To this end, we have developed a new reproducible MPD measurement and representation together with its theoretical definition. By modifying the encircled flux (EF) representation, which is based on the near-field pattern of a graded-index multimode optical fiber (GI-MMF), we define the encircled angular flux (EAF) for step-index multimode optical fibers (SI-MMFs) based on their far-field patterns (FFPs). Using a SI-MMF, which is used for low cost short distance interconnects, as an example, we show the changes in the MPD along the fiber, reveal an unusual insertion loss-increasing phenomenon due to the evanescent tails of the core modes extending into the cladding, and characterize the equilibrium mode distribution (EMD) in the fiber. The EAF representation enables these phenomena to be quantified. We also propose an EAF template that consists of the target EMD and its tolerance. If device system designers use the EAF template to set the launch conditions, they can perform a fair assessment of the components, and they can design the system performance even if some components are replaced with others from a different supplier manufactured by a different method. We call this concept “Total MPD Management.”

An Energy Efficient Multi-Gbit/s NoC Transceiver Architecture With Combined AC/DC Drivers and Stoppable Clocking in 65 nm and 28 nm CMOS

This paper presents a network-on-chip (NoC) SerDes transceiver architecture for long distance interconnects in the mm range within MPSoCs. Its source synchronous clocking scheme enables application in GALS systems and allows completely stoppable transceiver clocking for low idle power consumption. A capacitive line driver with combined resistive driver for well defined DC swing is employed and analyzed in detail by simulation studies. It is shown that proper DC swing definition is mandatory for robust operation of long links at high data rates. Prototypes of the transceiver over 6 mm bufferless on-chip interconnect are implemented in both 65 nm and 28 nm CMOS technologies. The 65 nm realization achieves an efficiency of 173 fJ/bit/mm at 90 Gbit/s at 1.25 V and 93 fJ/bit/mm at 45 Gbit/s low speed mode at 0.9 V. The 28 nm realization achieves 81 fJ/bit/mm at 72 Gbit/s at 1.05 V and 64 fJ/bit/mm at 36 Gbit/s low speed mode at 0.95 V. The transceiver can be seamlessly integrated as black box point-to-point connection into heterogeneous MPSoC NoCs to enable ultra-compact toplevel floorplan realization and increased energy efficiency. An example of a 20-core MPSoC in 65 nm CMOS technology with 10 serial NoC transceivers is presented.

Demonstration of Latency Reduction in Ultra-Short Distance Interconnections Using Optical Fanout

In this paper, we present a hybrid optoelectronic system capable of improving the performance of short-distance broadcast architectures such as arrays of gates, cells, or processors. By incorporating a layer of the optical fanout in the electrical fanout link, we utilize the "electrical isolation" advantage of the optical interconnects in addition to the absence of the parasitics. Our proof-of-concept circuits show improved signal latency even for a small electrical fanout to 32 minimum-sized inverters, corresponding to an electrical link length of 200mum in a 0.25-mum technology. This is the first physical demonstration of the advantages of using optics in silicon very large scale integrated (VLSI) systems at these distances. We also demonstrate an advantage in the energy-delay-squared product for fanout to more than 250 minimum-sized inverters corresponding to a distance of 1.5 mm

GaAs VCSEL's at /spl lambda/=780 and 835 nm for short-distance 2.5-Gb/s plastic optical fiber data links

Selectively oxidized GaAs vertical-cavity surface-emitting lasers for /spl lambda/=780- and 835-nm emission wavelength and 120-/spl mu/m-core diameter step index plastic optical fiber (POF) are investigated for short distance interconnects. 2.5-Gb/s pseudorandom data transmission over up to 2.5 m of plastic fiber is demonstrated with a bit-error rate (BER) of better than 10/sup -11/. Furthermore, bias-free data transmission at 2.5 Gb/s over 1-m fiber length again at a BER of better than 10/sup -11/ is reported.

High-performance small vertical-cavity lasers: a comparison of measured improvements in optical and current confinement in devices using tapered apertures

We analyze the scaling characteristics of the optical and current confinement for three different vertical-cavity surface-emitting laser (VCSEL) structures with tapered apertures. The improvements in scaling have allowed devices with apertures <3 /spl mu/m to have wall-plug efficiencies over 20% at output powers as low as 150 /spl mu/W. The combination of low threshold (<200 /spl mu/A), single modedness, and good wall-plug efficiency even at low output powers makes these devices excellent candidates for short distance (<1 m) interconnects within computers.