Non-blocking, scalable optical router architecture and...

Details Non-blocking, scalable optical router architecture and... Non-blocking, scalable optical router architecture and...

2000-10-27

2003-12-16

Negash, Kinfe-Michael (Department: 2633)

Optical communications

Multiplex

Optical switching

C398S045000, C398S049000, C370S351000, C370S352000, C370S389000

Reexamination Certificate

active

06665495

ABSTRACT:

TECHNICAL FIELD OF THE INVENTION
The present invention relates generally to telecommunications systems and methods, and more particularly, a non-blocking, scalable optical router having an architecture that optimizes bandwidth management to allow for non-blocking switching and routing of optical data packets.
BACKGROUND OF THE INVENTION
The emergence of the Internet and the reliance by business and consumers on the transfer of data in all daily activities requires telecommunications networks and components that can deliver ever increasing amounts of data at faster speeds with higher quality levels. Current telecommunications networks fail to meet these requirements.
Existing electrical and electro-optical switching routers are limited in the switching speeds that are attained and the data capacity that can be processed between switches in a non-blocking manner. Current electrical switching architectures are generally limited to a switching speed of 40-100 Gigabits. In an attempt to overcome this limitation, current electrical and optical routers use aggregation of slower switches to increase the overall switching speed of the router. For example, a system may combine a hundred one (1) Gigabit routers to increase the switching speed of the system. However, while the overall speed and capacity will exceed one Gigabit, this aggregation will not result in full 100 Gigabit per second speed and capacity, resulting in a decreased efficiency (less than full realization of switching capability). Furthermore, aggregation increases costs due to the increased number of routers and increases complexity due to interconnect and routing issues. In addition to the issues surrounding data routing speed, electronic telecommunication routing systems all face difficult transference issues when interfacing with optical data packets. Another technique used in electrical telecommunication routing systems to increase data routing speed is parallel processing. However, this technique has its own limitations including control complexity (it is difficult to control the multiple routers operating in parallel). In any of these techniques involving multiple routers to increase the processing speed, a single control machine must arbitrate among the many multiple machines which increases control complexity, cost and ultimately uses an electronic control machine that is limited by electronic processing speeds.
FIGS. 1 and 2
will illustrate the limitations of these prior art systems.
FIG. 1
shows a typical prior art local network cluster
10
that uses an interconnect structure with multiple routers and switches to provide the local geographic area with a bandwidth capability greater than that possible with any one switch in the router
10
. Network
10
includes four routers
12
, which will be assumed to be 300 Gigabit per second routers, each of which serves a separate area of 150 Gbps of local traffic. Thus, the 300 Gigabit capacity is divided by assigning 150 Gigabits per second (Gbps) to the incoming traffic on local traffic links
16
and assigning 50 Gbps to each of three links
14
. Thus, each link
14
connects the router
12
to every other router in the network
10
, thereby consuming the other
150
gigabit capacity of the router
12
. This interconnectivity is in the form of a “mesh” that allows each router
12
to communicate directly with every other router
12
in the network
10
.
This configuration has a number of limitations. While the four local geographic area produce a total of 600 Gbps of capacity, the network
10
requires four routers
12
of 300 Gbps each, or 1200 Gbps of total router capacity, to provide the interconnectivity required to allow direct communication between all routers
12
. Additionally, even though fully connected, each router
12
does not have access to all of the capacity from any other router
12
. Thus, only one third of the local traffic (i.e., only 50 Gbps of the total potential 150 Gbps) can be switched directly from any one router
12
to another router
12
, and the total potential traffic demand is 600 Gigabits per second. In order to carry more traffic over a link
14
, a larger capacity would be required at each router
12
(for example, to carry all 150 Gbps over a link
14
between routers, each link
14
would need to be a 150 Gbps link and each router
12
would have to have an additional 300 Gbps capacity). Thus, to get full connectivity and full capacity, a non-blocking cluster network
10
having a mesh configuration would require routers with 600 Gbps capacity each which equates to 2400 Gbps total router capacity (or four times the combined traffic capacity of the local geographic areas).
FIG. 2
shows another prior art cluster router network
18
that aggregates sixteen data lines
20
that each can carry up to one hundred sixty gigabit per second of data that appears to have the potential capacity of 2.5 Terabits (16 lines carrying 160 Gbps each). Each of the data lines
20
is routed through an edge router
22
to an interconnected edge network
24
(e.g., a ring, mesh, ADM backbone or, other known interconnection method) via carrying lines
26
. However, due to inefficiencies in this network configuration (as described above), the full potential of 2.5 Terabits cannot be achieved without a tremendous increase in the size of the edge routers
22
. For example, if the edge routers are each 320 Gbps routers, then 160 Gbps is used to take incoming data from incoming data line
20
and only 160 Gbps of access remains to send data to each of the other fifteen routers
22
in the cluster
18
(i.e., approximately 10 Gbps can be allotted to each of the other fifteen routers, resulting in greater than 90% blockage of data between routers). Furthermore, the capacity of the routers is already underutilized as the overall router capacity of the network cluster
18
is 5 terabits per second (Tbps), while the data capacity actually being serviced is 2.5 Tbps. Even with the router capacity underutilized, the network
18
has over 90% blockage between interconnected routers through the edge network
24
. To increase the capacity between routers in a non-blocking manner, the individual routers would need to be increased in capacity tremendously, which increases cost and further exacerbates the underutilization problems already existing in the network.
Therefore, a need exists for an optical telecommunications network and switching architecture that will provide full, non-blocking routing between service areas that allow full capacity utilization without requiring over-sized routers that result in extreme underutilization of the router capacity and tremendous increase in router costs over the network.
SUMMARY OF THE INVENTION
The present invention provides a non-blocking optical routing system and method that substantially eliminates or reduces disadvantages and problems associated with previously developed optical-routing systems and methods.
More specifically, the present invention provides a system and method for providing non-blocking routing of optical data through a telecommunications network that allows full utilization of available capacity. The network includes a number of data links that carry optical data packets to and from an optical router. The optical router includes a number of ingress edge units coupled to an optical switch core coupled further to a number of egress edge units. The ingress edge units receive the optical data packets from the data links and aggregate the optical data packets into “super packets” where each super packet is to be routed to a particular destination egress edge unit or port. The super packets are sent from the ingress edge units to an optical switch fabric within the optical switch core that routes each super packet through the optical switch fabric to the super packet's particular destination egress edge unit in a non-blocking manner (i.e., without contention or data loss through the optical switch fabric). This routing is managed by a core controller that monitors the flow of incoming optical data packet

Affiliated with

Also associated with

Rating

Non-blocking, scalable optical router architecture and... does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Non-blocking, scalable optical router architecture and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Non-blocking, scalable optical router architecture and... will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3144553