Abstract:
In some examples, an optical node includes transition logic to: receive an indication of a data channel to be added across an optical medium, the data channel to occupy a portion of an optical spectrum; in response to a receipt of the indication, divide the data channel into a plurality of sub-channels; and sequentially add each of the plurality of sub-channels across the optical medium in a particular order.
Abstract:
In one embodiment, method of photonic packet switching includes receiving, by a photonic switching fabric from a first top-of-rack (TOR) switch, a destination port request corresponding to a first photonic packet and a first period of time, where the destination port request includes a first output port and determining whether the first output port is available during the first period of time. The method also includes receiving, by the photonic switching fabric from the first TOR switch, the first photonic packet and routing the first photonic packet to the first output port when the first output port is available during the first period of time. Additionally, the method includes routing the first photonic packet to an alternative output port when the first output port is not available.
Abstract:
Embodiments are provided for scalable photonic packet fabric architectures using photonic integrated circuit switches. The architectures use compact size silicon photonic circuits that can be arranged in a combined centralized and distributed manner. In an embodiment, an optical switch structure comprises a plurality of core photonic based switches and a plurality of photonic interface units (PIUs) optically coupled to the core photonic based switches and to a plurality of groups of top-of-rack switches (TORs). Each PIU comprises a N×N silicon photonic (SiP) switch optically coupled to a group of TORs associated with the PIU from the groups of TORs, where N is a number of the TORs in each group. The PIU also comprises a plurality of 1×P SiP switches coupled to the group of TORs associated with the PIU and to the core photonic based switches, where P is a number of the core photonic based switches.
Abstract:
A method and apparatus for routing signals through a photonic switch are provided. Optical Signal-to-Noise Ratio (OSNR) requirements for signals to be concurrently routed through the switch are determined, and incoming signal routing requests are blocked when routing same would violate OSNR requirements. Blocking may occur when a maximum allowed number M of lightpaths of same wavelength would be exceeded by admitting the request. Otherwise, signals are routed along a lightpath which satisfies the OSNR requirements. Cell Extinction Ratio in conjunction with OSNR requirements can be used to determine M. Switching cells can potentially accommodate multiple lightpaths of different wavelengths, but regular switching cells may be inhibited from accommodating multiple lightpaths of same wavelength. Routing solutions which maximize both cell sharing and cell packing may be sought. Routes that violate crosstalk limitation conditions are inhibited.
Abstract:
Crosstalk can be suppressed in photonic switching fabrics by activating unused photonic elements in a manner that manipulates the inactive connections and inhibits the propagation of cross-talk over the switching fabric. For example, unused photonic elements can be set to a cross or bar configuration to block first and second order crosstalk from propagating to the output ports, thereby reducing noise in the output signals. All of the unused elements can be activated in order to maximize crosstalk suppression. Alternatively, fewer than all of the unused elements may be activated to achieve a balance between crosstalk suppression and power conservation. Photonic switch architectures can be configured to use pre-determined cross-talk suppression maps (e.g., patterns of activated unused cells) for the various switching configurations, which may be computed using a recursive algorithm.
Abstract:
In one embodiment, optical node includes an input port configured to receive an input optical frame and a first optical switch coupled to the input port, where the first optical switch is configured to remove the input optical frame to produce a removed input optical frame when an address of the input optical frame is a node address of the optical node and to pass the input optical frame to a second optical switch when an address of the input optical frame differs from the address of the optical node. The optical node also includes the second optical switch coupled to the first optical switch, where the second optical switch is configured to output the input optical frame to an output port when the address of the input optical frame is not the node address of the optical node and the address of the input optical frame is not an empty address.
Abstract:
Methods and apparatus for a reconfigurable optical add-drop multiplexer (ROADM) cluster node are provided. In some embodiments, the ROADM cluster node includes a set of g line chassis for performing line functionality. In some embodiments, the ROADM cluster node further includes a set of h add-drop chassis for performing add-drop functionality. In some embodiments, each of the g line chassis includes a set of N line cards and a set of M interconnect cards. In some embodiments, the ROADM cluster node further includes a set of M interconnect chassis configured for interconnecting each line chassis to each other line chassis. In some embodiments, the set of M interconnect chassis is further configured for interconnecting each line chassis to each of the h add-drop chassis. In some embodiments, the ROADM cluster node separates the line functionality and add-drop functionality. In some embodiments, 1.15N≤M≤1.5N.
Abstract:
Hybrid dilated Benes photonic switching architectures employ an arrangement of two-by-one (2×1) photonic and two-by-two (2×2) photonic elements to enjoy improved cross-talk performance while maintaining moderate cell counts. A jumpsuit switch optical network node architecture comprising multiple stages may operate more efficiently than single stage switching fabrics, by enabling manipulation of connectivity in some stages to achieve load balancing over other stages. Specifically, a first stage of switching fabrics connected to input ports of the optical node may be manipulated to load balance incoming signals over a second stage of switching fabrics coupled to output ports of the optical node. Additionally, a third stage of switching fabrics connected to add ports of the optical node may be manipulated to load balance added optical signals over the second stage of switching fabrics.
Abstract:
In one embodiment, a method of photonic frame scheduling includes receiving, by a photonic switching fabric from a top of rack (TOR) switch, a frame request requesting a time slot for switching an optical frame to an output port of a photonic switch of the photonic switching fabric and determining whether the output port of the photonic switch is available during the time slot, and generating a contention signal including a grant or a rejection, in accordance with the determining. Also, the method includes assigning the time slot to the TOR switch for the output port of the photonic switch, when the contention signal includes the grant, transmitting, by the photonic switching fabric to the TOR switch, the contention signal and receiving, by the photonic switching fabric from the TOR switch, the optical frame during the time slot, when the contention signal includes the grant.
Abstract:
Recursive optimization algorithms can be used to determine which idle photonic switching elements to configure in N×N switching fabrics to achieve crosstalk suppression. Different algorithms are used to achieve different levels of optimization. Embodiment full optimization techniques may configure all inactive cells to reduce crosstalk, and consequently may provide the best noise performance and highest power usage. Partial optimizations may configure fewer than all inactive cells to reduce crosstalk, and may provide sub-optimal noise performance at lower power usages. Differential partial optimization algorithms configure inactive cells in different stages of a photonic switching fabric. Fewer than all cells in a given stage may be configured by some algorithms.