Expandable universal network

• US 20040114539A1
• Filed: 12/11/2002
• Published: 06/17/2004
• Est. Priority Date: 12/11/2002
• Status: Active Grant

• Current Assignee: RPX Clearinghouse LLC (RPX Corporation)
• Original Assignee: Nortel Networks Limited (Nortel Networks Corporation)
• Inventors: Beshai, Maged E., Jamoussi, Bilel N.
• Granted Patent: US 7,450,845 B2
• Time in Patent Office: Days
• US Class Current: 370/254
• CPC Class Codes: H04L 49/25 Routing or path finding in a switch fabric
• CPC Class Codes: H04L 49/254 Centralised controller, i.e. arbitration or scheduling
• CPC Class Codes: H04L 49/357 Fibre channel switches

Abstract

A wide-coverage high-capacity network that scales to multiples of petabits per second is disclosed. The network comprises numerous universal edge nodes interconnected by numerous disjoint optical core connectors of moderate sizes so that a route set for any edge-node-pair includes routes each of which has a bounded number of hops. A core connector can be a passive connector, such as an AWG (arrayed wave-guide grating) device, or an active connector such as an optical channel switch or an optical time-shared channel switch. The core capacity is shared. Thus, failure of a proportion of core connectors reduces network connectivity but does not result in a major service discontinuity.

The network uses a source routing scheme where a set of candidate routes for each node pair is determined from the network topology and the candidate routes for a node pair are sorted according to their differential propagation delays. A method of measuring the differential propagation delay is also disclosed.

Claims

1. A network comprising:

a plurality of edge nodes interconnected by a backbone core and an auxiliary core, said backbone core comprising a first set of wavelength-division-multiplexed links and a plurality of primary optical connectors, said auxiliary core comprising a second set of wavelength-division-multiplexed links and a plurality of secondary optical connectors, wherein,

said backbone core provides, for each edge node, at least one backbone path having a bounded number of links to each other edge node; and

said auxiliary core provides, for at least one edge node, at least one auxiliary path to at least another node.

2. The network of claim 1 further including a controller associated with each of said edge nodes, said controller comprising

a route-selection module for selecting candidate routes from said each of said edge nodes to any other of said edge nodes;

a first time-locking module for sending time locking signals from said each of said edge nodes to at least one of said secondary optical connectors;

a propagation-delay-measurement module for determining relative propagation delays along said candidate routes; and

a first connection-scheduling module for scheduling signal transfer from an input side to an output side of said each of said edge nodes.

3. The network of claim 2 wherein said route-selection module at a first edge node is adapted to maintain a description of a set of routes to a second edge node.

4. The network of claim 3 wherein said first time-locking module comprises a time counter and means for communicating a reading of said time counter to external nodes.

5. The network of claim 4 wherein said propagation-delay-measurement module is adapted to exchange time-measurements with at least another propagation-delay-measurement module and estimate the differential propagation delay along any two routes in said set of routes.

6. The network of claim 5 wherein said time-measurements are derived from said time counter.

7. The network of claim 4 wherein said first connection-scheduling module is a fine time-slot scheduling module adapted to schedule time slots in a predefined time frame.

8. The network of claim 2 further including a controller associated with each of said secondary optical connectors, said controller comprising

a second time-locking module for exchanging time locking signals with at least one of said edges;

and

a second connection-scheduling module for scheduling signal transfer from an input side to an output side of said each of said secondary optical connectors.

9. The network of claim 8 wherein said second connection-scheduling module is a coarse time-slot scheduling module adapted to schedule time slots in a predefined time frame.

10. The network of claim 7 wherein said edge time-slot scheduling module schedules Ω

₁ time slots per time unit, the value of Ω

₁ determined according to the inequality

Ω

₁≧

(C×

T)/(Ψ

×

R×

δ

)

where

C is the total access capacity of said each of said edge nodes expressed in bits per second;

T is the duration of said time frame in seconds;

Ψ

is the mean connection time in seconds;

R is a channel capacity expressed in bits per second; and

δ

is the duration in seconds of a fine time slot.

11. The network of claim 8 wherein said coarse time-slot scheduling module schedules Ω

₂ time slots per time unit, the value of Ω

₂ determined according to the inequality

Ω

₂≧

(C*×

T)/(Ψ

×

R×

Δ

)

where

C* is the total access capacity of said each of said secondary optical connectors expressed in bits per second;

T is the duration of said time frame in seconds;

Ψ

is the mean connection time in seconds;

R is a channel capacity expressed in bits per second; and

Δ

is the duration in seconds of a coarse time slot.

12. The network of claim 1 wherein said at least one auxiliary path does not overlap any of said at least one backbone path.

13. The network of claim 1 wherein at least one of said first plurality of optical connectors is a static wavelength router comprising arrayed-waveguide-grating demultiplexers and multiplexers.

14. The network of claim 1 wherein said second plurality of optical connectors includes at least one optical channel switch.

15. The network of claim 1 wherein said second plurality of optical connectors includes at least one optical time-division-multiplexed switch.

16. The network of claim 15 wherein each of said edge nodes has a plurality of input ports and a plurality of output ports and wherein

data received at each input port is segmented in narrow segments of equal sizes, and

data transmitted from each output port is structured in wide segments each of which containing an integer-multiple of narrow segments.

17. A backbone network comprising:

edge nodes arranged in a recursive structure of node groups of order J, J being an integer greater than zero and each edge node constituting a zero-order group, wherein

node groups of order (j−

1), 0<

j≦

J, are interconnected by m_j×

m_j connectors to form a node group of order j so that each node has j non-overlapping paths each traversing (j−

1) intermediate nodes and

$\left(\underset{k=1}{\overset{j}{\sum <br><br>}}<br><br>\text{\hspace{1em}<br><br>}<br><br>m_k-2\times <br><br>j\right)$

non-overlapping paths each traversing J intermediate nodes to each other non-adjacent node in any other (j−

1)^th-order group.

18. The backbone network of claim 17 wherein the number of nodes in a node group of order J does not exceed

19. The backbone network of claim 17 wherein at least one connector is an arrayed-waveguide-grating wavelength router.

20. The backbone network of claim 17 wherein at least one connector is an optical channel switch.

21. The backbone network of claim 17 wherein at least one connector is an optical switch adapted to switch time-slotted optical signals arranged in a time frame.

22. The backbone network of claim 17 wherein said network order J is determined from

23. A method of measuring differential one-way propagation delays along each route in a route set having at least two candidate routes from a first node to a second node, the method comprising steps of:

for each of said at least two candidate routes

sending a timing message at said first edge node along each of said at least two candidate routes;

recording time information at each subsequent edge node along the route; and

recording a queueing delay at each subsequent edge node along the route;

computing differential propagation delays at said second node; and

communicating a result back to a source node.

24. The method of claim 23 wherein each of said candidate routes has a specified number of hops and including the further step of providing a route definition including an identifier of an output port in each node traversed by the route.

25. The method of claim 24 including the further step of providing, in said timing message, a message identifier, an identifier of said first node, an identifier of said second node, a forwarding record for each node along said candidate route, and an index that points to a forwarding record holding the output-port identifier of a current node along said candidate route.

26. The method of claim 25 including the further step of updating said index at each edge node along a route.

27. A method of operation for an edge node adapted to interface with core connectors, each connector having an input side and an output side, said core connectors including static connectors having fixed input side to output side connectivity and adaptive connectors that modify input side to output side connectivity according to traffic-load variation, the edge node operable to perform steps comprising:

time-locking to said adaptive connectors; and

computing differential one-way propagation delays along candidate routes traversing said core connectors towards a second edge node.

28. The method of claim 27 including the further steps of said edge node determining an estimate of an extreme value of the rate of data directed to said second edge node and selecting a number of said candidate routes according to said estimate to form a route set.

29. The method of claim 28 including the further step of sorting said routes according to a merit based at least in part on said differential one-way propagation delays.

30. The method of claim 29 including the further step of said first edge node selecting a route from said route set according to said merit.