CN113473272A - Non-blocking optical interconnection architecture for data center switch - Google Patents
Non-blocking optical interconnection architecture for data center switch Download PDFInfo
- Publication number
- CN113473272A CN113473272A CN202110748591.0A CN202110748591A CN113473272A CN 113473272 A CN113473272 A CN 113473272A CN 202110748591 A CN202110748591 A CN 202110748591A CN 113473272 A CN113473272 A CN 113473272A
- Authority
- CN
- China
- Prior art keywords
- optical
- board
- cross
- service board
- service
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04Q—SELECTING
- H04Q11/00—Selecting arrangements for multiplex systems
- H04Q11/0001—Selecting arrangements for multiplex systems using optical switching
- H04Q11/0005—Switch and router aspects
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04Q—SELECTING
- H04Q11/00—Selecting arrangements for multiplex systems
- H04Q11/0001—Selecting arrangements for multiplex systems using optical switching
- H04Q11/0005—Switch and router aspects
- H04Q2011/0052—Interconnection of switches
- H04Q2011/0056—Clos
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Optical Communication System (AREA)
Abstract
The invention discloses a non-blocking optical interconnection architecture for a data center switch. The service board is connected with the cross board through the optical waveguide backboard; the service board comprises a plurality of optical modules, a framer, an optical engine and a plurality of optical connectors, wherein the plurality of optical modules are connected to the framer, the framer is connected with the optical engine, and the optical engine is connected with the plurality of optical connectors; the optical connectors of the service board are connected with the optical connectors of the cross board through the optical waveguide backplane, the cross board comprises a plurality of optical connectors, an optical engine and a cross chip, and the optical connectors are all connected to the optical engine which is connected with the cross chip. The invention is an optical switch framework with large capacity, high reliability and low delay, breaks through the bottlenecks of bandwidth, power consumption and the like of the traditional electric back plate, and has higher switching rate, lower occupied space and lower cost.
Description
Technical Field
The invention belongs to a non-blocking optical interconnection framework for a data center switch, and relates to a realization framework and a method for carrying out data exchange by utilizing optical waveguide backboard interconnection.
Background
With the development of services such as cloud computing, big data, internet of things and the like, the bandwidth demand of users is remarkably increased, and real-time and large-amount data transmission needs a network to provide higher bandwidth and lower delay. The traditional electrical interconnection mode is limited by the problems of low exchange rate, high power consumption, time-consuming photoelectric conversion and the like, and is difficult to be applied to future exchange networks with large capacity and high bandwidth. In order to solve the problem, the optical interconnection key technology should be focused, and the overall performance of the switching system is further improved. Meanwhile, compared with an optical switch, the electrical switch is only different in that the electrical switch is interconnected by an electrical link on an electrical backplane, the optical switch is interconnected by an optical waveguide on an optical waveguide backplane, external network interfaces of the electrical switch and the optical switch are the same, and only internal structures of the electrical switch and the optical switch are different, so that the optical switch can be compatible with an existing network without changing an existing network structure. In addition, the optical switch has the advantages of large capacity, high reliability, low power consumption and the like.
Currently, the architecture based on the optical waveguide backplane is diversified, and the attention is focused on a large-capacity system. In the design of a large-capacity switching system, network congestion can bring higher queuing delay, and when the cache is insufficient, high packet loss rate is easily generated, and the factors influence the reliability, delay and throughput of the switching network. Non-blocking networks have attracted considerable attention from researchers due to their highly reliable, low-delay nature. Clos architecture has high reliability, expandability and other superior performance, and becomes the mainstream choice of large capacity switching network.
Disclosure of Invention
In order to solve the problems in the background art, the invention provides a non-blocking optical interconnection architecture for a data center switch, which applies all-optical interconnection with high capacity, high reliability and low power consumption, solves the technical problems of low switching rate, high power consumption and the like of the traditional electric switch, and achieves the aim of realizing low power consumption and high-speed flow switching on the premise of high throughput.
In order to achieve the purpose, the invention adopts the following technical scheme:
the invention comprises a service board, an optical waveguide backboard and a cross board, wherein the service board is connected with the cross board through the optical waveguide backboard;
the service board comprises a plurality of optical modules, a framer, an optical engine and a plurality of optical connectors, wherein the number of the optical modules and the number of the optical connectors are the same, the plurality of optical modules are all connected to a multi-port on one side of the framer, the multi-port on the other side of the framer is respectively connected with the multi-port on one side of the optical engine, and the multi-port on the other side of the optical engine is connected with the plurality of optical connectors; the optical connectors of the service board are connected with the optical connectors of the cross board through the optical waveguide backplane, the cross board comprises a plurality of optical connectors, an optical engine and a cross chip, the optical connectors are all connected to a multi-port on one side of the optical engine, and a port on the other side of the optical engine is connected with the cross chip.
Data of an external network is input into an optical module of an original service board, is converted into an electric signal by the optical module of the original service board, is forwarded to a framer of the original service board for processing and distribution to a port of an optical engine of the original service board, is modulated by the optical engine of the original service board and then is converted into an optical signal, the optical signal is sent to an optical waveguide backboard through an optical connector on the original service board connected with the optical waveguide backboard, is sent to an optical engine of a cross board after being received by the optical connector on the cross board through the optical waveguide backboard, is sent to a cross chip after being subjected to photoelectric conversion by the optical engine of the cross board, is subjected to packet exchange by the cross chip, is sent to the optical waveguide backboard through another optical connector on the cross board connected with the optical waveguide backboard, and is sent to the optical engine of a target service board after being received by the optical connector on the optical waveguide backboard, the optical signals are converted into electric signals by the optical engine of the target service board and then are sent to the framer of the target service board, and the electric signals are distributed to the optical module of the target service board by the framer and then are sent to an external network.
The optical waveguide back plate comprises a plurality of paths of mutually independent connecting components, optical connectors are arranged at two ends of each path of connecting component, and the optical connectors at two ends of each path of connecting component are in communication connection; each path connecting assembly is used for connecting one optical connector of the service board with one optical connector of the cross board.
In the service board, every two optical modules form a group and are used for being responsible for 1 node to send and receive data.
The two optical connectors in the service board and the two optical connectors in the cross board are in a group and are used for forwarding and receiving 1 group of data.
The strict non-blocking optical interconnection structure comprises two service boards and a cross board, wherein each optical connector in each service board is connected with one optical connector in the cross board; the data received by the optical module of each service board is processed and distributed by the framer, then is converted by the optical engine and sent to the cross board from the optical connector which needs to be routed out, is sent to the optical connector of the service board which needs to be forwarded from the optical connector of the cross board after being exchanged by the cross chip of the cross board, and is processed and distributed to the optical module by the optical engine and the framer of the service board which needs to be forwarded.
The optical interconnection structure comprises N service boards and M cross boards, wherein each service board is provided with 2M optical connectors and 2M optical modules, and each cross board is provided with 2N optical connectors; in each service board, every two optical connectors are used as a group of service board connection groups; in each cross board, every two optical connectors are used as a group of cross board connection groups; each group of service board connection groups of the service boards are respectively connected to one group of cross board connection groups of different cross boards, and each group of cross board connection groups of the cross boards are respectively connected to one group of service board connection groups of different service boards;
the data received by the optical module of each service board is processed and distributed by the framer, then is converted by the optical engine and sent to the cross board of the required route from the optical connector of the required route, is switched by the cross chip of the cross board and then is sent to the optical connector of the service board required to be forwarded from the optical connector of the cross board, and then is processed and distributed to the optical module by the optical engine conversion and the framer of the service board required to be forwarded.
Non-blocking is divided into strictly non-blocking and re-arranging non-blocking, that is, fig. 2, 3 and 4 are both non-blocking, wherein the strictly non-blocking architecture is fig. 2, and the re-arranging non-blocking architecture is fig. 3 and 4.
The invention has the beneficial effects that:
the invention can be widely used for realizing high-capacity, high-reliability and low-power-consumption all-optical interconnection, and utilizes the optical waveguide backboard to connect the service board and the cross board, thereby realizing high-capacity data exchange by using a single switch, overcoming the defects of low exchange rate, high power consumption and the like of the traditional electric backboard, reducing the blocking rate and queuing time delay by a non-blocking framework, further improving the performance of the network, and having higher application value in practical engineering.
The invention is an optical interconnection framework with high capacity, high reliability and low delay, and uses a framer and a cross chip, thereby breaking through the bottlenecks of bandwidth, power consumption and the like of the traditional electric back plate on the basis of being compatible with the existing network, and leading a switch to have higher switching rate, lower occupied space and lower cost.
Drawings
FIG. 1 is a system architecture data exchange flow diagram of the present invention.
Fig. 2 is a schematic diagram of a strictly non-blocking optical waveguide backplane architecture.
FIG. 3 is a schematic diagram of a re-arrangeable non-blocking optical waveguide backplane architecture.
Fig. 4 is a schematic diagram of a next generation re-arrangeable non-blocking optical waveguide backplane architecture.
Detailed Description
The invention is further illustrated by the following figures and examples.
As shown in fig. 1, the optical waveguide network system comprises a service board, an optical waveguide backplane and a cross board, wherein the service board is connected with the cross board through the optical waveguide backplane;
the business board is used as an input stage and an output stage and is responsible for data transmission of an external network, and the cross board is used as an intermediate stage and is responsible for data exchange; the service board and the cross board are interconnected via the optical waveguide backplane and the optical connectors thereon.
The service board comprises a plurality of optical modules, a framer, an optical engine and a plurality of optical connectors, wherein the number of the optical modules and the number of the optical connectors are the same, the plurality of optical modules are all connected to a multi-port on one side of the framer, the multi-port on the other side of the framer is respectively connected with the multi-port on one side of the optical engine, and the multi-port on the other side of the optical engine is connected with the plurality of optical connectors;
the optical connectors of the service board are connected with the optical connectors of the cross board through the optical waveguide backplane, the cross board comprises a plurality of optical connectors, an optical engine and a cross chip, the number of the optical connectors of the cross board is the same as that of the optical connectors of all the service boards, the plurality of optical connectors are all connected to a multi-port on one side of the optical engine, and the port on the other side of the optical engine is connected with the cross chip.
The service board takes the framer as a core chip, and the cross board takes the cross chip as the core chip.
The optical module of the service board sends and receives external network data, the optical engine is used for photoelectric conversion, the framer is used for data receiving and frame synchronization, and the cross chip is used for packet switching of data.
Data of an external network is input into an optical module of an original service board, is converted into an electric signal by the optical module of the original service board, is forwarded to a framer of the original service board for processing and distribution to a port of an optical engine of the original service board, is modulated by the optical engine of the original service board and then is converted into an optical signal, the optical signal is sent to an optical waveguide backboard through an optical connector on the original service board connected with the optical waveguide backboard, is sent to an optical engine of a cross board after being sent to the optical connector on the cross board through an optical waveguide on the optical waveguide backboard for receiving, is sent to a cross chip after being subjected to photoelectric conversion by the optical engine of the cross board, is subjected to packet exchange by the cross chip, is sent to the optical waveguide backboard through another optical connector on the cross board connected with the optical waveguide backboard, and is sent to the optical engine of a target service board after being received by the optical connector on the optical waveguide backboard, the optical signals are converted into electric signals by the optical engine of the target service board and then are sent to the framer of the target service board, and the electric signals are distributed to the optical module of the target service board by the framer and then are sent to an external network.
As shown in fig. 1, the optical waveguide backplane includes multiple independent connection assemblies, optical connectors are disposed at two ends of each connection assembly, and the optical connectors at two ends of each connection assembly are connected through optical waveguides and other communications; each path of connecting assembly is used for connecting one optical connector of the service board with one optical connector of the cross board, and the number of the connecting assemblies is the same as that of the optical connectors of all the service boards.
The optical connector is used for transmitting optical signals and plays a role of only connection transmission.
In the service board, every two optical modules form a group and are used for transmitting and receiving 1 data. And two optical connectors in the service board and two optical connectors in the cross board are in a group and are used for forwarding and receiving 1 group of data.
As shown in fig. 2, the non-blocking optical interconnect architecture includes two service boards and only one cross board, each optical connector in each service board is connected to one optical connector in the cross board; the data received by the optical module of each service board is processed and distributed by the framer, then is converted by the optical engine and sent to the cross board from the optical connector which needs to be routed out, is sent to the optical connector of the service board which needs to be forwarded from the optical connector of the cross chip after being exchanged by the cross chip of the cross board, and is processed and distributed to the optical module by the optical engine and the framer of the service board which needs to be forwarded.
As shown in fig. 3 and 4, the non-blocking optical interconnect architecture includes N service boards each having 2M optical connectors and 2M optical modules, and M cross boards each having 2N optical connectors; in each service board, every two optical connectors are used as a group of service board connection groups; in each cross board, every two optical connectors are used as a group of cross board connection groups; each group of service board connection groups of the service boards are respectively connected to one group of cross board connection groups of different cross boards, and each group of cross board connection groups of the cross boards are respectively connected to one group of service board connection groups of different service boards;
the data received by the optical module of each service board is processed and distributed by the framer, then is converted by the optical engine and sent to the cross board of the required route from the optical connector of the required route, is switched by the cross chip of the cross board and then is sent to the optical connector of the service board required to be forwarded from the optical connector of the cross chip, and then is processed and distributed to the optical module by the optical engine conversion and the framer of the service board required to be forwarded.
The service board and the cross board are interconnected using a Clos topology. In the Clos-based three-stage switch, if l is more than or equal to 2n-1, the switch is a strict non-blocking architecture, and if l is more than or equal to n, the switch is a rearrangeable non-blocking architecture. The optical waveguide backplane interconnection architecture is designed based on physical limitations of the chip, the number of optical waveguide backplane channels and the like.
Example 1: strictly non-blocking optical interconnect architecture for data center switches
As shown in fig. 2, the architecture of the present embodiment includes 2 service boards and 1 cross board.
The maximum optical path number of the optical engine is 12, and the cross chip respectively receives or transmits 6 optical signals at most. Each service board is provided with 6 Optical Modules (OM) connected with an external network, each 2 optical modules are responsible for sending and receiving data of 1 node, and each service board is respectively connected with 3 nodes. Each service board has 6 unidirectional optical connections with the cross-board, which optical channels are established by Optical Connectors (OC).
The architecture can provide interconnection for 6 nodes, is limited by a light engine, has a transmission rate of 112Gbps per channel and has a total crossing capacity of 1.344 Tbps.
In this architecture, for 6 sets of data streams with different input and output ports, at the input stage traffic boards, the data streams from different input ports are transmitted over different optical links to the intermediate stage crossbar where, similarly, the data streams destined for different nodes will also be forwarded over different optical links to the destination traffic board. The intermediate stage crossbar can forward data streams with different pairs of output ports to the corresponding OCs without blocking.
In this architecture, congestion will only occur when output ports collide; only one cross board is provided, and the service board can forward the data packet to any OC on the service board according to a load balancing strategy, so that the complexity of the architecture is greatly reduced, and the delay is also reduced.
Example 2: repeatable non-blocking optical interconnect architecture for data center switches
As shown in fig. 3, the architecture of the present embodiment includes 6 service boards and 2 cross boards.
The maximum optical path number of the optical waveguide back plate is 24, 2 cross boards can be connected at most, and each cross board can be connected with 6 service boards.
Each service board is provided with 4 Optical Modules (OM) connected with an external network, each 2 optical modules are responsible for sending and receiving data of 1 node, and each service board is respectively connected with 2 nodes. Each service board has 4 unidirectional optical connections with the cross-board, which optical channels are established by Optical Connectors (OC).
The architecture can provide interconnection for 12 nodes, is limited by light engines, and has a transmission rate of 112Gbps per channel, so that the total crossing capacity of the equipment is 2.688 Tbps.
In this architecture, the number n of the first-stage input ports is 2, the number l of the second-stage cross plates is 2, and the two are equal, which is a re-arrangeable non-blocking architecture. In the framework, 2 data streams of the service board cannot be sent to the same cross board, and the selection of the middle-level cross board is carried out in the service board to avoid blocking; the 12 sets of data streams with different input and output ports are routed through the network simultaneously without blocking.
Example 3: next generation re-arrangeable non-blocking optical interconnection architecture for data center switch
As shown in fig. 4, the limitations of optical waveguide preparation, wiring and coupling technologies, silicon photonic chips, etc. have led to the switching capability of optical waveguide backplane communication systems. With the development of optical waveguide backplane technology, it can support more optical channels, i.e. can interconnect more nodes, and provide higher throughput for a single board. On this basis, the present embodiment designs the next generation optical waveguide backplane architecture.
Under the same optical waveguide bandwidth and chip design constraints, when the number of interconnected channels is not limited, the maximum throughput which can be realized by a single board is explored.
In the implementation, limited by the light engine device, a maximum of 12 connectors can be arranged on the service board or the cross board, and since the connection is unidirectional, the optical waveguide backplane can interconnect 6 service boards and 6 cross boards at most. Each service board is provided with 12 Optical Modules (OM) connected with an external network, each 2 optical modules are responsible for sending and receiving data of 1 node, and each service board is respectively connected with 6 nodes. There are 12 unidirectional optical connections between each service board and the cross board, the number of optical waveguide backplane interconnections is 72 x 72, and the optical channels are established by Optical Connectors (OC).
The architecture can provide interconnection for 36 nodes, is limited by light engines, and has a transmission rate of 112Gbps per channel, so that the total crossing capacity of the equipment is 8.064 Tbps.
In the architecture, the number n of the first-stage input ports is 6, the number l of the second-stage cross boards is 6, and the two are equal, so that the architecture is a re-arranging non-blocking architecture, and data flows of 36 node pairs can simultaneously traverse the network without blocking.
The non-blocking interconnection method of the data center optical switch of the invention completes the transmission of optical signals by the cooperation of the connection of the service board and the optical waveguide back board and the cross board.
Therefore, the non-blocking optical waveguide backboard interconnection architecture with high capacity, high reliability and low power consumption can effectively overcome the defects of low switching rate and high power consumption of the traditional electric switch, can be compatible with the existing network, and has the characteristics of low delay, low packet loss rate and the like.
The foregoing detailed description is intended to illustrate and not limit the invention, which is intended to be within the spirit and scope of the appended claims, and any changes and modifications that fall within the true spirit and scope of the invention are intended to be covered by the following claims.
Claims (7)
1. A non-blocking optical interconnect architecture for a data center switch, characterized by: the system comprises a service board, an optical waveguide backboard and a cross board, wherein the service board is connected with the cross board through the optical waveguide backboard; the service board comprises a plurality of optical modules, a framer, an optical engine and a plurality of optical connectors, wherein the number of the optical modules and the number of the optical connectors are the same, the plurality of optical modules are all connected to a multi-port on one side of the framer, the multi-port on the other side of the framer is respectively connected with the multi-port on one side of the optical engine, and the multi-port on the other side of the optical engine is connected with the plurality of optical connectors; the optical connectors of the service board are connected with the optical connectors of the cross board through the optical waveguide backplane, the cross board comprises a plurality of optical connectors, an optical engine and a cross chip, the optical connectors are all connected to a multi-port on one side of the optical engine, and a port on the other side of the optical engine is connected with the cross chip.
2. A non-blocking optical interconnect fabric for a data center switch as claimed in claim 1, wherein: data of an external network is input into an optical module of an original service board, is converted into an electric signal by the optical module of the original service board, is forwarded to a framer of the original service board for processing and distribution to a port of an optical engine of the original service board, is modulated by the optical engine of the original service board and then is converted into an optical signal, the optical signal is sent to an optical waveguide backboard through an optical connector on the original service board connected with the optical waveguide backboard, is sent to an optical engine of a cross board after being received by the optical connector on the cross board through the optical waveguide backboard, is sent to a cross chip after being subjected to photoelectric conversion by the optical engine of the cross board, is subjected to packet exchange by the cross chip, is sent to the optical waveguide backboard through another optical connector on the cross board connected with the optical waveguide backboard, and is sent to the optical engine of a target service board after being received by the optical connector on the optical waveguide backboard, the optical signals are converted into electric signals by the optical engine of the target service board and then are sent to the framer of the target service board, and the electric signals are distributed to the optical module of the target service board by the framer and then are sent to an external network.
3. A non-blocking optical interconnect fabric for a data center switch as claimed in claim 1, wherein: the optical waveguide back plate comprises a plurality of paths of mutually independent connecting components, optical connectors are arranged at two ends of each path of connecting component, and the optical connectors at two ends of each path of connecting component are in communication connection; each path connecting assembly is used for connecting one optical connector of the service board with one optical connector of the cross board.
4. A non-blocking optical interconnect fabric for a data center switch as claimed in claim 1, wherein: in the service board, every two optical modules form a group and are used for being responsible for 1 node to send and receive data.
5. A non-blocking optical interconnect fabric for a data center switch as claimed in claim 1, wherein: the two optical connectors in the service board and the two optical connectors in the cross board are in a group and are used for forwarding and receiving 1 group of data.
6. A non-blocking optical interconnect fabric for a data center switch as claimed in claim 1, wherein: the strict non-blocking optical interconnection structure comprises two service boards and a cross board, wherein each optical connector in each service board is connected with one optical connector in the cross board; the data received by the optical module of each service board is processed and distributed by the framer, then is converted by the optical engine and sent to the cross board from the optical connector which needs to be routed out, is sent to the optical connector of the service board which needs to be forwarded from the optical connector of the cross board after being exchanged by the cross chip of the cross board, and is processed and distributed to the optical module by the optical engine and the framer of the service board which needs to be forwarded.
7. A non-blocking optical interconnect fabric for a data center switch as claimed in claim 1, wherein: the optical interconnection structure comprises N service boards and M cross boards, wherein each service board is provided with 2M optical connectors and 2M optical modules, and each cross board is provided with 2N optical connectors; in each service board, every two optical connectors are used as a group of service board connection groups; in each cross board, every two optical connectors are used as a group of cross board connection groups; each group of service board connection groups of the service boards are respectively connected to one group of cross board connection groups of different cross boards, and each group of cross board connection groups of the cross boards are respectively connected to one group of service board connection groups of different service boards; the data received by the optical module of each service board is processed and distributed by the framer, then is converted by the optical engine and sent to the cross board of the required route from the optical connector of the required route, is switched by the cross chip of the cross board and then is sent to the optical connector of the service board required to be forwarded from the optical connector of the cross board, and then is processed and distributed to the optical module by the optical engine conversion and the framer of the service board required to be forwarded.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110748591.0A CN113473272A (en) | 2021-07-02 | 2021-07-02 | Non-blocking optical interconnection architecture for data center switch |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110748591.0A CN113473272A (en) | 2021-07-02 | 2021-07-02 | Non-blocking optical interconnection architecture for data center switch |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113473272A true CN113473272A (en) | 2021-10-01 |
Family
ID=77877374
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110748591.0A Pending CN113473272A (en) | 2021-07-02 | 2021-07-02 | Non-blocking optical interconnection architecture for data center switch |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113473272A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024002115A1 (en) * | 2022-07-01 | 2024-01-04 | 华为技术有限公司 | Service processing method, optical module, access site and convergence site |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030210870A1 (en) * | 1999-10-06 | 2003-11-13 | Graves Alan F. | Switch for optical signals |
| CN105392071A (en) * | 2015-11-20 | 2016-03-09 | 天津光电通信技术有限公司 | Optical line intersection system architecture based on circuit domain |
| CN107105355A (en) * | 2016-02-23 | 2017-08-29 | 中兴通讯股份有限公司 | A kind of exchange method and exchange system |
-
2021
- 2021-07-02 CN CN202110748591.0A patent/CN113473272A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030210870A1 (en) * | 1999-10-06 | 2003-11-13 | Graves Alan F. | Switch for optical signals |
| CN105392071A (en) * | 2015-11-20 | 2016-03-09 | 天津光电通信技术有限公司 | Optical line intersection system architecture based on circuit domain |
| CN107105355A (en) * | 2016-02-23 | 2017-08-29 | 中兴通讯股份有限公司 | A kind of exchange method and exchange system |
Non-Patent Citations (1)
| Title |
|---|
| 周宇萌、邱昆、许渤、崔展齐、凌云、陈锡莲: "超大容量光交换机中器件约束的Clos网络路径设计", 《光学学报》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024002115A1 (en) * | 2022-07-01 | 2024-01-04 | 华为技术有限公司 | Service processing method, optical module, access site and convergence site |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11671330B2 (en) | Network interconnect as a switch | |
| US8270830B2 (en) | Optical network for cluster computing | |
| US8103137B2 (en) | Optical network for cluster computing | |
| US8121478B2 (en) | Method and apparatus for implementing non-blocking computer interconnection network using bidirectional optical switch | |
| US20130156425A1 (en) | Optical Network for Cluster Computing | |
| CN101917333B (en) | Region-based photoelectric double-layer network-on-a-chip and routing method | |
| JP2018535613A (en) | Line card chassis, multi-chassis cluster router and routing and packet processing | |
| US8780927B2 (en) | Data network elements, crossbars, and methods providing coupling between remote PHY and MAC devices | |
| CN111865472B (en) | Bufferless optical interconnection architecture and method for data center | |
| CN110568560B (en) | Method for realizing large-scale optical switch topology array chip based on Benes structure and having balanced loss | |
| CN113473272A (en) | Non-blocking optical interconnection architecture for data center switch | |
| CN110430486B (en) | Boundary switching node based on integrated hybrid optical switching network | |
| CN105323660A (en) | Light signal cross system, and cross processing method and device | |
| CN110568559A (en) | Chip architecture based on large-scale optical switch topology array | |
| CN109361967B (en) | Cluster-based data center optical network architecture method | |
| CN114614939B (en) | Flexible clustered data center optical interconnection network | |
| Yamanaka | Breakthrough technologies for the high-performance electrical ATM switching system | |
| TWI716279B (en) | Optical tunnel switching network system | |
| TWI776294B (en) | Optical frame switch | |
| Jiang et al. | Nonblocking optical MINs under crosstalk-free constraint | |
| Wang et al. | Systematic exploration of high-radix integrated silicon photonic switches for datacenters | |
| CN115643224B (en) | FC switch and data transmission method | |
| CN112083530B (en) | NxM blocked low-cost optical cross matrix | |
| Yang et al. | A Non-blocking Network Design for Terabit Capacity Optical Interconnects | |
| Lu et al. | NEST: Towards Extreme Scale Computing Systems |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20211001 |
|
| WD01 | Invention patent application deemed withdrawn after publication |