Using our rapid prototyping technology called Deep Proton Writing (DPW), we have in recent years made a
wide range of micro-optical components with a large depth (500-μm) for a variety of applications. One of these
components is a pluggable out-of-plane coupler for printed circuit board-level optical interconnections. Whereas
DPW is capable of rapidly fabricating high-quality master components, the technology is not suitable for low-cost
mass fabrication. Therefore, we investigate the replication of
out-of-plane coupling components using hot
embossing, through the fabrication of a metal mould of the DPW master by applying electroplating. We compare
these hot embossed replicas with components replicated using the elastomeric mould vacuum casting technology.
One of the grand challenges in solving the interconnection bottlenecks at the Printed Circuit Board (PCB) and
Multi-Chip-Module (MCM) level, is to adequately replace the PCB and intra-MCM galvanic interconnects with
high-performance, low-cost, compact and reliable micro-photonic alternatives. Therefore we address the following
components in this paper: 1) out-of-plane couplers for optical waveguides embedded in PCB, 2) peripheral fiber
ribbons and two-dimensional single- and multimode fiber connectors for high-speed parallel optical connections,
and 3) intra-MCM level optical interconnections via free-space optical modules.
For the fabrication of these micro-optical interconnect modules, we are focusing at the Vrije Universiteit
Brussel on the continuous development of a rapid prototyping technology, which we call Deep Proton Writing
(DPW). The special feature of this prototyping technology is that it is compatible with commercial low-cost
mass replication techniques such as micro injection moulding and hot embossing. Laser ablation is used at
Ghent University for the fabrication of PCB-embedded waveguides and integrated micro-mirrors. The main
advantage of this technology is that it is compatible with present-day PCB manufacturing.
For the free-space MCM-level optical interconnect module, we furthermore give special attention to the
optical tolerancing and the opto-mechanical integration of the components. We use both a sensitivity analysis to
misalignment errors and Monte Carlo simulations. It is our aim to investigate the whole component integration
chain from the optoelectronic device to the micro-opto-mechanical components constituting the interconnect
module.
One of the remaining challenges to solve the interconnection bottlenecks at the Printed Circuit Board (PCB) and Multi-Chip-Module (MCM) level, is to adequately replace the galvanic interconnects with high-performance, low-cost, compact and reliable micro-photonic alternatives. At our labs of the Vrije Universiteit Brussel we are therefore optimizing and deploying a rapid micro-optical prototyping technology for micro-optical interconnect modules, which we call Deep Proton Writing (DPW). An advantage of the DPW process is that it can create steep micro-optical surfaces, micro-holes, micro-lenses and alignment features in one irradation step. Hence, relative accuracies are very well controlled. In this report, we will address more specifically the following components, made each with the DPW technology: 1) out-of-plane couplers for optical wave-guides embedded in PCB, 2) peripheral fiber ribbons and two dimensional single- and multimode fiber connectors for high-speed parallel optical connections, and 3) intra-MCM level optical interconnections via free-space optical modules. We will give special attention to the optical tolerancing and the opto-mechanical integration of components in their packages. We use both a sensitivity analysis to misalignment errors and Monte-Carlo simulations. It is our aim to investigate the whole component integration chain from the optoelectronic device packaging to the micro-opto-mechanical assembly of the interconnect module.
Nowadays, multiprocessor systems are reaching their limits due to the large interconnection bottleneck between chips, but recent advances in the development of optical interconnect technologies can allow the use of low cost, scalable and reconfigurable networks to resolve the problem. In this paper, we make an initial evaluation of the performance gain on general network reconfigurability. In a next stage, we propose an optical system concept and describe a passive optical broadcasting component to be used as the key element in a broadcast-and-select reconfigurable network. We also discuss the available opto-electronic components and the restrictions they impose on network performance. Through detailed simulations of benchmark executions, we show that the proposed system architecture can provide a significant speedup for shared-memory machines, even when taking into account the limitations imposed by the opto-electronics and the presented optical broadcast component.
Electrical interconnection networks connecting the different processors and memory modules in modern large-scale multiprocessor machines are running into several physical limitations. In shared-memory machines, where the network is part of the memory hierarchy, high network latencies cause a significant performance bottleneck.
Parallel optical interconnection technologies can alleviate this bottleneck by providing fast and high-bandwidth links. Moreover, new devices like tunable lasers and detectors or MEMS mirror arrays allow us to reconfigure the network at runtime in a data transparent way. This allows for extra connections between distant node pairs that communicate intensely, achieving a high virtual network connectivity by providing only a limited number of physical links at each moment in time.
In this paper, we propose a reconfigurable network architecture that can be built using available low cost components and identify the limitations these components impose on network performance. We show, through detailed simulation of benchmark executions, that the proposed network can provide a significant speedup for shared-memory machines, even with the described limitations.
LED-based projectors have numerous advantages compared to traditional projectors, such as compactness, larger color gamut, longer lifetime, and lower supply voltage. As LEDs can switch rapidly, there is the possibility to pulse. However, there is also an important disadvantage. The optical power per unit of étendue of an LED is significantly lower than, e.g., an ultra-high-performance (UHP) lamp. This problem can be remedied partly by pulsing the LEDs. If one drives an LED with a pulsed current source, the peak luminance can be higher, albeit the average luminance will not increase. By pulsing two LEDs alternately (50% duty cycle), their increased flux can be added up in time and will generate a higher average flux within the same étendue. We combine the LEDs with a polarizing beam splitter (PBS) and change the polarization of one LED with a switchable retarder. The achieved substantial net gain after all losses is 36%.
One of the grand challenges in solving the interconnection bottlenecks at the Printed Circuit Board (PCB) and Multi-Chip-Module (MCM) level, is to adequately replace the PCB and intra-MCM galvanic interconnects with high-performance, low-cost, compact and reliable micro-photonic alternatives. In our labs at the Vrije Universiteit Brussel we are therefore focusing on the continuous development of a rapid prototyping technology for micro-optical interconnect modules, which we call Deep Proton Writing (DPW).The special feature of this prototyping technology is that it is compatible with commercial low-cost mass replication techniques such as micro injection moulding and hot embossing. We will address more specifically in this paper the following components: 1) out-of-plane couplers for optical wave-guides embedded in PCB, 2) peripheral fiber ribbons and two dimensional single- and multimode fiber connectors for high-speed parallel optical connections, and 3) intra-MCM level optical interconnections via free-space optical modules.
We furthermore give special attention to the optical tolerancing and the opto-mechanical integration of the components. We use both a sensitivity analysis to misalignment errors and Monte Carlo simulations. It is our aim to investigate the whole component integration chain from the optoelectronic device to the micro-opto-mechanical components constituting the interconnect module.
Led based projectors have numerous advantages compared to traditional projectors, such as: compact, larger color gamut, longer lifetime, lower supply voltage, etc. As LED's can switch rapidly, there is the possibility to pulse. However, there is also an important disadvantage. The optical power per unit of etendue of a LED is significantly lower than e.g. an UHP-lamp (approximately 50 times). This problem can be remedied partly by pulsing of the LED’s. If one drives a LED with a pulsed current source, the peak luminance can be higher, albeit that the average luminance will not increase. By pulsing X LED's alternately, their increased flux can be added up in time and will generate a higher average flux within the same etendue. This can be carried out in a number of different configurations. The first configuration uses moving components where a number of LED's (e.g. 8) are mounted on a carrousel and consecutively the pulsed LED is brought in the light path of the projector to fill up the time with its peak flux. An alternative without moving components can be reached with 2 LED's which are combined with a PBS. By alternately pulsing the LED's with 50% duty cycle and changing the polarisation of one LED with a switchable retarder, one can combine the flux of both LED's in the same etendue. Because of its fast switching time ferro-electric retarders are used here. This can be extended further to 4,8,16... LED's, at the price of a larger and more complicated optical architecture.
We present the latest results of a students case-study on the design, implementation and testing of an on-campus free-space laser datalink. Although the concept of such a laserlink dates from the sixties, the technology has gained much attention lately because it provides a powerful alternative for high bandwidth short-distance datalinks in dense urban areas, university campuses, disaster zones; where line-of-sight is available, and construction with buried cables is costly.
From an educational point of view, the multidisciplinary topic of free-space optical communication is particular suited to illustrate several photonics courses to Electrotechnical Engineering students. The topic not only involves the study of opto-electronics, optical system design and optical datacommunications, but also covers selected topics in the domains of automation, modulating electronics, opto- and electromechanics. By imparting our first-year master students with this practical hands-on project experience, we provide them with the necessary practical and theoretical skills for their future photonics career.
After completion this year the optical link will serve a threefold goal. At first instance it will be used in the practical classes for future photonics students. The demonstrator will also promote photonics to the general public during our university info days and will serve as a basic platform for new future students projects, where
we intend to extend the link in the direction of coarse WDM.
Although over the past few years state-of-the-art point-to-point optical interconnects have shown the potential to fulfill the ever increasing demand for higher data communication bandwidth, still electronic interconnects are favoured over optical interconnects because electronics is a much more mature and established technology. However, when photonic interconnects could allow more complex and richer sets of interconnect patterns, by e.g. allowing for one-to-many optical interconnects (signal broadcasting) and reconfigurable point-to-point optical interconnects, they might outperform electronics both in terms of bandwidth and ease of reconfiguration. In this paper we do a concept study of several approaches to bring signal broadcast within an existing free-space (FS) plastic micro-optical interconnect intra-chip component. The original component consists of a combination of a refractive microlens array and a classical high-quality microprism. The idea of signal broadcasting can be realized by incorporating a fan-out diffractive optical element (DOE) at certain positions in this component. In a first design we integrate the DOE on the deflection edge of the microprism. For a second design we focus on the replacement of the refractive microlens array by their diffractive counterparts. In this approach the fan-out functionality of the DOE is combined with the lens functionality of the diffractive microlens arrays. In a third approach we target multi-faceted diffractive microlens arrays to implement the fan-out functionality. All presented designs can bring signal broadcast to the intra-chip optical interconnect level, although some of them will turn out to be more attractive for practical implementation in demonstrators. We compare and discuss the advantages and disadvantages of the proposed designs.
We study the potentialities of three dimensional micro-optical pathway blocks combining refractive microlens arrays, reflective micro-prisms and diffractive fan-out elements, to enhance the functionalities of short-distance intra-chip optical interconnects. As an example, we demonstrate the possibility to enhance the point-to-point interconnection functionality to that of broadcasting data over a chip. We also investigate the limitations imposed by the physical dimensions of the refractive and diffractive micro-optical components. We then illustrate this example by a quantitatively elaborated design of a fan-out element from a VCSEL array to a detector array with a 1 to 9 signal broadcasting for every source. Furthermore we show that with the use of DOE’s we can achieve a broadcasting functionality that can lead towards reconfigurable optical interconnects, with the aid of wavelength sensitive resonant cavity detectors and WDM-inspired interconnection schemes.
In highly parallel computer systems, reconfigurable interconnect network topologies can improve the performance by adaptively increasing the communication bandwidth where it is most needed. In electrical reconfigurable interconnect networks (e.g. crossbars or multi-stage networks), a high reconfigurability can only be achieved at the cost of both chip area and network latency. The facts that short-distance optical link latencies are rapidly decreasing and that new technologies allow optical reconfigurability, make optical interconnects an interesting alternative to overcome these interconnection issues. Optical interconnection technologies indeed offer several possibilities to increase network connectivity without drastically increasing the chip area and the delay costs. In this paper we study the bandwidth and latency requirements of inter-processor and processor-memory interconnect for shared-memory parallel computers when the processor clock increases up to 10 GHz. We also investigate new enabling technologies and discuss their potential use in architectures based on reconfigurable optical interconnects.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.