Optical-core network with selective signal queueing
View Patent ↗A core network shared by a large number of edge nodes comprises core nodes interconnected by core channels. Selected core channels are provided with buffers to enable temporal alignment of signals arriving at any core node from several other core nodes. A buffer may be provided at either end of a core channel. Otherwise, all other ports of the core node may be bufferless. Providing such buffers enables the construction of high-capacity wide-coverage network comprising a large number of core nodes shared by numerous edge nodes. Methods of temporal coordination among the network nodes are disclosed.
1. A network comprising:
a plurality of optical core nodes, each having an optical switch fabric including a plurality of inward ports receiving optical signals from other core nodes; and a plurality of outward ports transmitting optical signals to other core nodes; and
a plurality of edge nodes each subtending to at least one of said core nodes;
wherein at least one channel connects a selected outward port of a first core node to a selected inward port of a second core node and said at least one channel is provided with a signal buffer and a buffer controller; said buffer and buffer controller associated with said selected outward port;
wherein at least one of said core nodes further includes:
a plurality of bufferless inlet ports receiving optical signals from subtending edge nodes; and
a plurality of bufferless outlet ports transmitting optical signals to subtending edge nodes;
and wherein said buffer controller has a time-locking circuit operable to exchange time-locking signals with controllers of external nodes.
2. The network of claim 1 wherein said buffer is associated with said selected inward port.
3. The network of claim 2 wherein said buffer controller sorts buffered signals according to the outlet ports and outward ports to which the signals are switched through the optical switch fabric of said first core node.
4. The network of claim 1 wherein said buffer controller sorts buffered signals according to the outlet ports and outward ports to which the signals are switched through the optical switch fabric of said second core node.
5. A network comprising:
a plurality of optical core nodes, each having an optical switch fabric including a plurality of inward ports receiving optical signals from other core nodes; and a plurality of outward ports transmitting optical signals to other core nodes; and
a plurality of edge nodes each subtending to at least one of said core nodes;
wherein at least one channel connects a selected outward port of a first core node to a selected inward port of a second core node and said at least one channel is provided with a signal buffer and a buffer controller;
wherein at least one of said core nodes further includes:
a plurality of bufferless inlet ports receiving optical signals from subtending edge nodes; and
a plurality of bufferless outlet ports transmitting optical signals to subtending edge nodes;
wherein said buffer comprises an optical-to-electrical converter, an electronic memory, and an electrical-to-optical converter;
and wherein each of said core nodes has a core controller including a master time counter and a master timing circuit operable to transmit and receive timing signals.
6. The network of claim 5 wherein said master timing circuit distributes a time reference to said at least one of said inward ports.
7. The network of claim 6 wherein at least one of said edge nodes is time-locked to at least one of said core nodes.
8. The network of claim 7 wherein said core controller is adapted to receive connection requests from, and transmit connection schedules to, subtending edge nodes.
9. The network of claim 8 wherein said core controller further includes a scheduling mechanism for computing said connection schedules.
10. The network of claim 9 wherein said connection requests include time-slot allocations in a time-division-multiplexed frame.
11. The network of claim 9 wherein said connection requests include burst-transfer requests.
12. The network of claim 9 wherein said core controller is operable to receive flow-rate-allocation requests and schedule a corresponding number of time slots per time-division-multiplexed frame.
13. The network of claim 9 wherein said core controller is operable to receive flow-rate-allocation requests and compute corresponding burst-transfer permits.
14. The network of claim 13 wherein said core controller is further operable to receive burst-transfer requests and compute corresponding burst-transfer schedules.
15. The network of claim 14 wherein said core controller of a core node is further operable to allocate an entire channel from the core node to a selected edge node subtending to another of said core nodes.
16. A network comprising:
a plurality of edge nodes; and
a plurality of optical core nodes, at least one of said core nodes having:
an optical switch fabric; and
a core controller;
wherein at least one of said edge nodes is time-locked to said at least one of said core nodes;
and wherein said core controller is operable to:
receive requests for transfer of time-limited signals from said at least one of said edge nodes;
compute a transfer schedule for said time-limited signals; and
allocate a continuous path from said at least one of said core nodes to carry time-limited signals destined to a plurality of edge nodes subtending to another of said core nodes.
17. The network of claim 16 wherein said time-limited signal is an optical signal carrying at least one time slot belonging to a time-division-multiplexed frame.
18. The network of claim 16 wherein said time-limited signal is an optical signal carrying a data burst of an arbitrary size.
19. The network of claim 16 wherein said continuous path terminates in a selected edge node from among said plurality of edge nodes and said selected edge node transmits time-limited signals to other edge nodes through said another of said core nodes.
20. A meshed network comprising a plurality of optical core nodes each having:
an optical switch fabric;
a plurality of inward ports receiving optical signals from other core nodes; and
a plurality of outward ports transmitting optical signals to other core nodes;
a plurality of edge nodes each subtending to at least one of said core nodes and connecting to data sources and data sinks;
wherein at least one of said core nodes further includes:
a plurality of inlet ports receiving optical signals from subtending edge nodes; and
a plurality of outlet ports transmitting optical signals to subtending edge nodes;
wherein
at least one outward port of each core node has a core channel connecting to an inward port of another core node; and
at least one of said inward ports is provided with a buffer and a buffer controller for time aligning signals received at said buffer.
21. The meshed network of claim 20 wherein each of said core nodes further includes a core controller operable to communicate with core controllers of other core nodes to exchange connection-scheduling data.
22. A meshed network comprising a plurality of optical core nodes each having:
an optical switch fabric;
a plurality of inward ports receiving optical signals from other core nodes; and
a plurality of outward ports transmitting optical signals to other core nodes;
a plurality of edge nodes each subtending to at least one of said core nodes and connecting to data sources and data sinks;
wherein at least one of said core nodes further includes:
a plurality of inlet ports receiving optical signals from subtending edge nodes; and
a plurality of outlet ports transmitting optical signals to subtending edge nodes;
wherein
at least one outward port of each core node has a core channel connecting to an inward port of another core node; and
at least one of said outward ports is provided with a buffer;
wherein said buffer is associated with a buffer controller for time locking said buffer to an external node.
23. The meshed network of claim 22 wherein each of said core nodes further includes a core controller operable to communicate with core controllers of other core nodes to exchange connection-scheduling data.