NZ722392B2 - Packet switch using physical layer fiber pathways - Google Patents

Abstract

system for communicating traffic within a network, comprising a fiber optic pathway bundle providing one or more fiber optic pathways, each including a unique physical layer identifier; and at least one edge grooming switch having one or more ports, each port including a unique physical layer identifier. The edge grooming switch is operatively connected to the fiber optic pathway bundle, capable of receiving traffic on the edge grooming switch ports and redirecting the traffic onto the fiber optic pathways. The system also comprises at least one path grooming switch capable of receiving traffic from the fiber optic pathway bundle, and redirecting the traffic to one or more path grooming switch ports each including a physical layer identifier; and a controller for controlling the edge grooming switches, the path grooming switches and the fiber optic pathway bundle for optimizing the fiber optic pathways using corresponding physical layer identifiers to dynamically configure the fiber optic pathways into bundles forming a connection of a speed based on a destination. tifier. The edge grooming switch is operatively connected to the fiber optic pathway bundle, capable of receiving traffic on the edge grooming switch ports and redirecting the traffic onto the fiber optic pathways. The system also comprises at least one path grooming switch capable of receiving traffic from the fiber optic pathway bundle, and redirecting the traffic to one or more path grooming switch ports each including a physical layer identifier; and a controller for controlling the edge grooming switches, the path grooming switches and the fiber optic pathway bundle for optimizing the fiber optic pathways using corresponding physical layer identifiers to dynamically configure the fiber optic pathways into bundles forming a connection of a speed based on a destination.

Description

Patent Applicationfor PACKET SWITCH USENG PHYSICAL LAYER WEE PATHWAYS ENVENTQRS: CRQSS-REFERENCE T0 RELATED APPLICATHONS {hill} This application claims priority to co~pending US. Provisional Application No. tili930,979, filed on January 24, 2M4, entitled "Packet Switch Using Physical Layer Fiber ys and Software Control to Groom Traffic in a Network"; US. Provisional Application No. filiQ'J’ZJZl, filed on iii-lurch 28, 2014, entitled "Built In Redundancy of Alternate Links Within 3 "; US. Provisional Anuiicstion No. 621’020394, filed on July 3, 20M, entitled "Data Center Patti Switch With improved Path interconnection Architecture"; US. Provisional Aeplieation No, 62/057,008, fiierl on September 29, 29M, entitled m For increasing Fiber Port Density in Data Center Applications"; and US. Provisional Application No, 62/657,098fi, filed on September 29, 20M, entitled "Data Center Network", each of which is incorporated herein in their entirety by nce.

BACKGRQUNE) {i392} Field {993} The present disclosure relates generally to communications systems, and more particularly to patching and ing systems used in Communications systems, [@943 Deserigtion of the d Art [@053 Communication networks (or ks) inciude runny network devices, such as switches, s, e devices, computers, telephone systems, servers, and other hardware and software that form infrastructures used for communications, which includes data, voice and video communications, Networks can be implemented in many types of environments incinding, for exampte, homes, hnsiness offices, data centers, and carriers, flit:., {thin} In most networks, many active network devices are ted to each other through network switches and routers. Network switches and routers have the ability to iearn how to reach each other through teaming methods and protoeois which are weii known in the industry. Alternatively, the k switches and renters are ahie to be configured by network administrators in a way that attows the network switches and routers to transport packets from one tocation to another" The switches and routers are constantiy iearning about other network devices added to or removed from the network and storing this topology information in a tahie within the switch or renter. When a packet is received by a switch or renter at a port or interface, the switch or renter inspects the packet for certain information, such as the destination address, or other information and then performs a tahie iooknp to determine where to send the packet The packet is then transmitted out of a port of a switch or roster towards the ed destination. The above process of receiving and transmitting of a packet at a switch or renter is known as packet sing, or is sometimes referred to as a "hop" in the network. {09?} To assist k administrators in managing packet processing, there are Network Management s S) avaiiahie that can draw a iogicai tion map between switches and renters and associated network s within the network. However, network management s iiy do not controi the flow of traffic (or s transmitted or received), or the mapping of aetnai traffic that passes through a network, Hence, most network strators do not know the actoai path that a packet or set of packets takes through the network. [€398] Networks are growing and changing at a rapid pace. This growth in networks is caused by many factors, some of which are growth in handwidth usage with video or high bandwidth applications on the network, such as sociai networking, the increase in number of servers to support the sociai networking or business applications, server virtnaiization, and the introduction of cloud based computing. Networks at carriers, data centers, and enterprises are faced with a ring increase in the number cf switches needed to snppcrt the grcwth, {£399} in a traditicnal netwcrlt, switches and renters ccnnect in each other and tc servers, ccinpnters, sterage devices, tclephcne systems, and other snch s. Fig. i shcws a simple netwcrit, where each switch ct renter 2 has several parts i, and where the parts i are ccnnected tn ether netwerlt devices, which can incinde ether es, renters, servers, ceinpnters, telephcnes, and other devices Packets are received at the ports 1 and then transmitted en tlher optic er copper cahles 3 in the lc in Fig. i, switch er rcuter 2A receives the packets than] ports l and then transmits the packets on ccpper or fiber cptic cahics 3 towards switch at renter ZR. Switch ct roster 28 es the packets, performs packet prccessing which incindes a table lcckup and then transmits the packets cn fiber cptic ct ccpper cables 3 en tcwartls switch at renter 2C. Switch cr renter 2C receives the packet, tines a tahie lock—up and transmits the packet on to its destinaticn port 4. Fig" 2 depicts a there ccrnplex network than that in Fig, i and shows a pluraiity cl? es andlcr renters 2 in the netwcrlt. Each switch ct renter 2 is involved in packet prccessing {691%} Fig. i and Fig° 2 show haw hop~hy~hcp pacltet precessing cccnrs in a network.

As the size cf the netwcrit grcws and mate switches and renters are added, packets have tn pass thrcngh mere hcps, and hence sdditicnal packet precessing steps are made, and the ccrnpleaity of the lt grows Fig. 3 depicts an even mere cctnpiex hetwcrlr that includes a pinraiity cf physical iayer patch panels in the network. The pcrts at each switch ct renter 2 are connected directly tc a part of a patch panel St A corresponding part on the patch panel 5 is then ted tn a pert ct ancther switch or renter 2 via a pert on a patch panei 5.

{Mill} The use of patch panels further complicates the netwcrlt. As the number of switches and renters 2 in the it increases, the nurnher cf patch panel parts aisc ses, necessitating there patch panels he prcvided‘ This adds tn the space requirements fer the netwctit as weil as the overall complexity of the network architecture.

ElllllZl With the increase in numher of switches, routers, etc, the complexity of the network has grown significantly. More es in the network, translates to more hops, in these more complex networks, packets endure additional processing, which increases the latency or delay for the packet to travel from one point to another in the network, and also increases the cost of maintaining and operating the network. This additional complexity also s in space, power and heat issnes within the network, e the additional es? routers and patch panels need physical space, consume power, and generate heat. ltlllflfil in the industry, the network devices referenced shove are typically connected using physical layer connectivity precincts, such as fiber optic cahles or copper cahles, and patch panelst For every connection in the logical layer (of the (El model), there are several tion points in the physical layer {of the gill model), which means that for every switch port there are several (additional) corresponding physical layer ports ented at patch panels and cross connects. {llllltll As the number of switches, routers, server hardware, etc, in the k grows, the number of patch panels will also grow. This adds to the cost of the network and puts more strain on space requirements as well as day—to—day maintenance and operating requirements. {hillfil it would he highly desirable to design a network with less devices and at the same time address the growth requirements in carrier, data center and enterprise networks.

BRIEF SUMMARY {little} The t disclosure relates to a system and method for communicating traffic within a network, The system includes a plurality of switches for receivingjtranstnitting traffic within the networlt, a programmable physical layer pathway for providing one or more pathways hetween the plurality of switches and a controller for controlling the plurality of switches and the mmable physical layer pathway for optimizing a flow or a plurality of flows of traffic within the network, wherein the ller defines the y across the network based on destinations of traffic or packets received by the plnreiity of switches BREEF DESCREPTEGN {3F THE DRAWENGS {991?} Fig. 1 depicts a simpie network according to the prior art; [0018} Fig. 2 depicts a more complicated network according to the prior art; [(30.19] Fig, 3 depicts a more complicated k incinding patch panels according to the prior art; {MEN Fig. 4 depicts a network system according to an embodiment of the present disclosure; {9921] Figs. 5A and SB depict an MPO connection ing to an embodiment of the present disclosure; {@922} Fig, 6 depicts a network system including patch panel s according to an ment of the present disclosure; and {@235} Fig. 7 depicts a network system according to an embodiment of the present disclosure.

BETAELED BEECRETKGN {@324} According to an embodiment of the present disciosnre, a system is provided that utilizes software to communicate with es, routers, patch panels, and copper and fiber optic cahies to guide the movement of packets in the network. The system reduces bandwidth load from core switches? spine switches? central switches, and aggregation switches and uses fiher optic pathways to transport packets across the centrai portions of the network, The system is able to t the tremendous growth in network dth reqnirernents with fewer switches in the network, thus translating to a lower cost of ent acquisition, lower consumption of power, lower s of heat dissipation, and lower requirements for cooling" [6625} Embodiments of the present disclosure may include a control and orchestration system capable of discovering all physical connection paths within a network, controlling one or more switches within the network and capable of reconfiguring the physical connections Traffic can be groomed at one or more of the es within the network for optimal traffic flow. 'l‘he switches may include one or more onders antlfor transceivers for performing optical-touelectrical signal sion andfor electricaluto~ optical signal conversion. Each switch may also include one or more optical cross connects or the like. Each switch may also include a module for communicating with the control and orchestration system. The control and orchestration system controls the transponders andior eivers as well as the optical cross connects allowing system administrators to reconfigure the network from a central location without the need for human intervention at the k interfaces, [liliZdl An embodiment of the present disclosure is shown in Fig" 4 and includes a plurality of switches such as at least one Edge Grooming Switch (£368) 16 and at least one Path Grooming Switch (PGS) 18. EGS lo and PGS 18 are onnected using one or more d fiber cables with high density tiher connections, such as fiher Push 0n (MPG) connectors, MXC tors, or other high density fiber connectors l7.

MP0 or MXC connection l7 may be, for example, a cable (erg, a ribhon cable) having a plurality of fiber optic cables terminating in MPO connectors or the like. EGS l6 and PGS 18 include one or more optical cross connects all as well as software modules 42 allowing communication between the switches and the Controller/Orchestration System l2. 1368 lo, PGS l8 - are software configurable hy Controller/Grchestration System 12.

This allows the network administrator to dynamically reconfigure the system as desired" MPGI’MXC connection 17 has a plurality of optical fibers or Physical layer Pathways as will he described later helow. EGS l6 receives traffic (cg, packets) from copper or fiber optic ports ll, converts the traffic to optical traffic signals (in the case of electrical signal traffic arriving over copper , performs processing via Packet Processor 44 on the packet within the PL? and further groom it to he transmitted and transmits this traffic onto one or more of the Physical Layer Pathways (PLPs) consisting of individual fibers or fiber groups within the MPQIMXC connection 17. P68 i8 receives the entire traffic transmitted by E63 16 via these PLPs in the MPG/MXC tion l7. Depending on the mmed configuration, PGS l8 can make an appropriate connection for the entire traffic from a PL? to transmit this traffic via one or more interfaces id to a final ation or a next node within the network. {@527} Ports ll of EGS l6 may include one or more copper based ports which may ntilize R345 type copper connectors or r connectors. Ports ll may also incinde one or more fiber optic ports which are connected to using one or more of several types of fiber optic connectors including LC, SC, MPO, MTP, MXC, SFP+, QSFP, QSFPni», etc. The ports ll may be implemented ly on E83 id or may be implemented on a cassette or card that can he plugged into the EGS 15. The EGS l6 receives traffic from devices that are connected on these copper or fiber optic ports and grooms the traffic to he transported across the network over PLPs which can be implemented using MTP/MPO, MXC fiher trunks, or other types of trunks. The grooming process is controlled by Controller/Orchestration system 12 which is able to identify, define and assign each Pill), for example each PL? within each connector l7, based on the c’s destination and is rable by k administrators. {@6323} The packets that are sent into the network from an EGS id are received by one or more PGSs l3. P68 18 may have one or more types of output ports l4, For example, these ports may include ill—4S type connectors or similar copper connectors. Ports l4 may also include one or more filter optic ports which are connected to using one or more of several types of fiber optic connectors including but not limited to LC, 3C3 MPO, MTR, SPF-t, Q8}??? etc The ports 14 may be implemented directly on PGS id or may be ented on a cassette or card (tea, a small form factor cage and associated transceiver) that can he plugged into the PGS l8. PGS i8 can perform multiple functions depending on where it is located in the network PGS l8 receives traffic from one or more EGS id devices or from another PGS l8 on one or more PLPs, and then switches this traffic for further transmission. PGS l8 is capable of receiving traffic on a Phi)? and switching the individnal PLPs or fiber pathways. That is, P83 l8 can switch the traffic traveling along a PL? from one MPO tion on to another FL? of another MPO connection for transporting the cattle to another node or its intended destination. PGS l8 is also cseable of receiving traffic from a number of PLPS, and switching the traffic for delivery to a destination switch, router or other device port using a single ace, or a number of interfaces with connectors such as , for example, LC, SC, MPG, Mil", MXC, SF?+, QSFP, QSFP+, etc, 963 18 can redirect its or switch it to a destination fiber port which can be implemented using, for example LC, SC, MPG, M’i‘l’, MXC, SF?+, QSFP, QSFP+, etc, {limit} Fig. 5A shows in more detail Mnlti~hher connection 17 according to an embodiment of the present disclosure. Multimtihei' connection l7 includes a Physical Interface l9, which in this embodiment is an MPO cable with connectors and includes one or more Logical Bundles 2i}, with each Logical Bundle 20 including one or more 3?th 21. A Programmable Light Path (FL?) is a connection that is created by the Qrchestration System from one point in the network to another for the use of packets to tlow through the network. A PL? can also he a Packet Flow from one destination to r. The PLP, created by the tration System traverses over one or more connectors and connections. A Pill) may be the Physical path, such as a it) ths connection from one point to r, passing through one or more EGS, FGS, connectors, and cable connections, A PM" may also be one of many Logical Paths 2i that are contained within a Physical Bundle '20. Within each Multiwtiber connection l7, there is a Logical Bundle 20, created by bundling fiber cable strands together to tone a tion of certain type or speed, for example, 2 fiber strands can be bundled together and used to create a lGG‘bps tion or a 25 Gbps connection (a transmit fiber and a e fiber), 8 fiber strands can be used to create a iii) ths (4 transmit fibers and 4 receive fibers using llleps each) or lilil Gbps tion (4 it fibers and 4 receive iibers using 25 ths each), 2t) fiber strands can be used to create a liltl Glens connection (ill transmit fibers and ill receive fibers using llleps each), or 64 fibers are used to create a 800 Gbps connection (32 transmit fibers and 32 receive fibers using Zfifibps each), or other such connections based on number of fiber strands. Within such a Mnlthfiber connection 17, there can be one or more Logical Bundles Ell. Each PL? 21 they be, for example, a fiber optic fiber or group of fibers within a innltitiher connection.

Matti—fiber connections 17 can have l2, 24, 48, 64, or 72 hills (cg, fibers) within the same connection, A fiber can be red to carry traffic at different rates such as lths, lthhps, 25631353, or other rates. Additionally, multiple fibers can be grouped as handles to transmit at higher rates such as dhfibps, lGGGbps, or other rates. [titliltll As shown in Fig. 518, each Physical interface l9 may include a plurality of Logical Bundles 20 each having a plurality of PLPs 2i. A Logical Bundle 2% may include no PLPs 2i but may be d for future use, Referring back to Fig. 4, it will be appreciated that EGS l6 thus actually has access to many PLPs and the ControlleriOrchestration system l2 enables system administrators to assign the number of PLPs that are used for each destination. The EGS id is able to receive l? packets on one or more of its ports ll, and then process and groom the packets to be transmitted on the appropriate PL? 2i. Controller/Orchestration system l2 is aware of the network architecture, and traffic can be groomed accordingly to reduce the number of hops. {@931} As shown in the embodiments described above with respect to Fig. 4, the copper based or fiber optic ports ll, id of EGS id and PGS l8, respectively, may be implemented directly on the switches themselves. According to another embodiment as shown in Fig. 6, the copper based or fiber optic ports or other types of interfaces as described above may be mounted to te packages (cassette or card) such as Patch Panel Modules 22 which can be inserted into £68 16 and PGS lit. Patch Panel Modules 22 may themselves be configurable by the Controller/Orchestration system 12. This embodiment allows the network administrator to cotnbine the functionality of the ing system with the functionality of a patching system. The network administrator or network designer has the choice of using the Patch Panel Module 22 or existing external ng systems. The patch panel connections can he uniquely identified, where the connector or adapter implemented on the switch or other device can be MPO, LC, SC, etc. each having a unique ill? and the cable connecter that is ed into the port can also have a unique ll) ated with it. To be able to ain the unique identifiers, Patch Panel s 22 may he designed with ninth wire technologies interfaces, REED tagging technology interfaces, tion point it) (Chill) technology interfaces, or other managed intelligence technologies or the Patch Panel Modules 22 may he designed with one or more of these different technology interfaces in order to provide the capabilities of supporting more than one particular managed intelligent technology These unique ills over the k wide interface allow physical connectivity ation for the entire network to be readily obtained by ller/Orchestration system l2. {3%le Fig, 7 depicts how embodiments of the t disclosure can reduce the number of devices in the k, when compared to Figure 3 (prior art) which contains a very large number of devices in the network, By managing traffic appropriately and taking advantage of the physical layer as part of the l ing network, embodiments of the present disclosure can reduce the number of total devices in the network, provide a relatively simple configuration, and obtain a higher level of knowledge of how traffic flows through the network. For example, packets received at interfaces ll of EGS ltiA are groomed according to their destination and appropriate PLPs are set up between EGS 16A and PGS lSA, PGS lSB, and P68 lSC. Traffic received from interface ll can be directly mapped to one or more PLPs in a single multi—tiher connection l7, or the traffic can undergo inspection and be routed to one or more PLPs in one or more of several multi—fi‘ner connections l7A— 17C. {@933} Each MPO, MXC, or other multiuiiher connection l7 has a unique fier; much like a MAC address and the Controller/Orchestration system l2 is able to read these unique cable identifiers from the MPQIMXC connections l7 using ninth wire technologies interfaces, RFED tagging technology interfaces, connection point ll) {{3le) technology aces, or other d intelligence technologies in addition, various other types of connections carrying one or more fiber pathways having unique identifiers may be utilized, where each pathway is also uniquely identified and reported throughout the system, For example, connections can he used on a switch or other device such as Nit: card, or an l cross connect, or other device using MFG, LC, SC or other types of fiber optic adapters and connectors. [9934} The Controllerlflrchestration system l2 enables system administrators to define pathways or l’Ll’s and ensure that they have a system and network design that delivers desired results. Furthermore, if network administrators desire to reconfigure PLPs, they are able to do so using software rather than physically sending human heings to disconnect and reconnect network interfaces and connections, although there will still be some types of changes that can he performed hy al human intervention at the switch, router, EGS, EPS or other devices. Furthennore, the network managers are able to test the performance of a link using software and without human intervention on the line, which is a capability that was not possible hefore. {@935} The shoyewdescriheti embodiments thus simplify the network hy grooming c at fewer points than in a ional network that has traditional switches and s. The simplification comes as a result of having fewer switches and the ability of EGS and P68 devices to replace several "hops" of the ional switch and router network. {9936} Although described herein utilizing MPG connections, other types of connections may he userl incluriing, for example, MT? tions. The ahovewuescrihecl embodiments can use the MTP, MPO, or MXC connections at the switch (EGS, PGS) to transnort one or more connections of lGB, lthB, 2563, 406:8, lGOGB through the same connection. This reduces the number of nhysicnl ports required on the switch, An MTP, MPO, or MXC connection located directly on a switch and carrying multiple lGB, lQGB, 25GB, 4063, 10063, or other speed tions has not heretofore been availahle. {will} As bed shove, embodiments of the present disclosure may use unique port and connection identifiers to document each connection point in the system. Any of several identification technologies can he used to accomplish the identification process, including but not limited to Connection Point it) (CHE) technology. While CPll) technology is known in the art, it has not heretofore been used on a switch or NIC card previously. Furthermore, CPI?) technology has not heretofore been implemented in a system carrying multiple s of ill, 25, 40, lllll, 460 GB aneously. [9933} The use of traffic grooming as described herein may involve the s of grouping many small communications flows into larger units, which can he processed as single entities. in this way, multiple flows of traffic destined for a common node can he pieced en the same PLP. Traffic grooming may else inveive rte-amplification, reshatiihg etidier tetiming ef signals within the network. {£36339} it wiii be understeeti that varieties medifieatiens can he made to the emhetiiments 0f the present diseiesure witheut éepertieg {mm the spirit and scope thereet‘. Therefore, the above ption sheuid het he construed as ng the diseiesui'e, hut mereiy as emhedimeets theteef. These skilled in the art wiii eevisien other cations within the scope and spirit ef the invention as defined by the eieims appended hereto. Fer example, the netwerit interfaces eeniempieted hy the present sure can utilize vetiehs eemmenieaiien groteeeis for netwerk eemmehieetiens. Further, the tietwerh aces may use vatieus emhediments et‘ transceivers and ters for eemmunieetien paths. As another exampie, the data center iietwerk architectures eeetempieted hy the present diseiesere eat: include singie iayer and multiuiayer switching layers.

I/WE

Claims (11)

CLAIM :

1. A system for communicating traffic within a k comprising: a fiber optic y bundle providing one or more fiber optic pathways, wherein each of the one or more fiber optic pathways includes a al layer identifier uniquely identifying the fiber optic pathway; at least one edge grooming switch having one or more ports, each of the one or more edge grooming switch ports es a physical layer identifier uniquely identifying each edge grooming switch port, the at least one edge grooming switch being operatively connected to the fiber optic pathway bundle, the at least one edge grooming switch being capable of receiving traffic on the one or more edge grooming switch ports and redirecting the traffic onto the one or more fiber optic pathways of the fiber optic pathway bundle; at least one path grooming switch connected to the fiber optic pathway bundle and e of receiving the traffic from the one or more fiber optic pathways and redirecting the traffic to one or more path grooming switch ports, wherein each path grooming switch port includes a physical layer identifier uniquely identifying each path grooming switch port; and a controller for controlling the at least one edge grooming , the at least one path grooming switch and the fiber optic y bundle for optimizing the one or more fiber optic pathways for the traffic to flow, wherein the controller identifies each path grooming switch port utilizing the path grooming port physical layer identifier, each edge grooming switch port utilizing the edge grooming port physical layer identifier and each fiber optic pathway utilizing the pathway physical layer identifier and wherein the controller can dynamically configure the one or more fiber optic pathways into logical s that form a connection of a certain speed based on a destination of the traffic received by the at least one edge grooming switch.

2. The system as recited in claim 1, wherein the one or more ports of the edge grooming switch are e of receiving the traffic from copper or fiber optic cables, and wherein the one or more ports on the at least one path grooming switch are capable of transferring the c to copper or fiber optic cables.

3. The system as recited in claim 1, wherein the traffic comprises s of data. 25944400_1

4. The system as recited in claim 1, wherein the fiber optic ys comprise at least one fiber optic cable.

5. The system as recited in claim 1, wherein the fiber optic pathway bundle comprises at least one multifiber cable.

6. The system as recited in claim 1, wherein the fiber optic pathways are connected to the at least one edge grooming switch using high density fiber connections.

7. The system as recited in claim 6, wherein the high-density fiber connections comprise one of MPO, MTP and MXC connections.

8. The system as recited in claim 1, wherein the fiber optic ys are connected to the at least one path grooming switch using high density fiber connections.

9. The system as d in claim 8, wherein the high-density fiber connections comprise one of MPO, MTP and MXC tions.

10. The system as recited in claim 1, wherein the at least one edge grooming switch es traffic on the one or more edge grooming switch ports from copper connections.

11. The system as recited in claim 1, wherein the at least one edge grooming switch receives traffic on the one or more edge grooming switch ports from fiber optic connections. Fiber Mountain, Inc. By the Attorneys for the Applicant SPRUSON & FERGUSON Per: 25944400_1