TITLE OF THE INVENTION;
COMPUTERNETWORKFORLAYERTHREE CORRELATIONWITH LAYERONETOPOLOGY
Cross-reference to Related Applications
The present application is related to co-pending U.S. Patent Application Ser. No. 60/287,106 entitled "DYNAMIC BANDWIDTH PROVISIONING FOR INTERNET PEERING" filed in the name of Alex Mashinsky on April 27, 2001 and U. S. Patent
Application Ser. No. 60/300,527 entitled "METHOD AND SYSTEM FOR PROVISIONING OPTICAL CIRCUITS IN A MULTI-NETWORK, MULTI-VENDOR ENVIRONMENT" filed in the name of Alex Mashinsky on June 22, 2001.
Field of the Invention
The present invention relates generally to high bandwidth communication systems and, more particularly, to a method and system for offloading traffic from lower to higher bandwidth networks based on quality of service and other factors.
Background of the Invention
Current peering arrangements between networks encourage carriers to hand off traffic from their own network to the peering partner as soon as possible. By so doing, the first carrier is able to free up a greater amount of on-net resources, thus improving their own network performance, while passing the cost of routing the traffic on to the peering partner. A problem with handling network traffic in this manner is that typically peering partners have little or no Quality of Service (QoS) agreements between them, and traffic that is handed off from one carrier to the next is typically routed using only a "Best Efforts" routing scheme.
The best effort routing requires the optical signals to be converted, for example, from optical ■ signals into electrical signals so that they may be routed at layer three of the network. Because of this, it is difficult for one carrier to offer QoS guarantees to their customers when handing that customer's traffic off to a peering partner. FIG. 1 depicts a prior art peering arrangement between two carriers, CARRIER X and
CARRIER Y, where traffic originating from ROUTER A traveling to ROUTER B is depicted by a dotted line with an arrow. Conversely, traffic originating from ROUTER B traveling to ROUTER A is depicted by a solid line with an arrow. As described above the traffic flows from the two routers take different paths through the respective carrier networks, each carrier routing the traffic across a different path, in an attempt to hand the traffic off to the other as quickly as possible.
This existing method of routing network traffic between peering partners was not particularly problematic in the past because much of the traffic that passed across data networks did not demand a very high quality of service. Email, Web browsing, and the like are generally not "real-time" applications that demand threshold levels of network performance for proper execution. But an increase of network capacity and an explosion of interest in network applications have increased demand for such "real-time" applications, including Voice Over Internet Protocol (VoIP), video conferencing, streaming video, person- to-person (P2P) applications, and the like. FIG. 2 depicts a pie chart breaking out total network data traffic into different priority groups. The majority of traffic passing across data networks remains in a relatively low priority bucket, constituting data generated by various applications that do not demand high performance routing. As depicted in FIG. 2, this traffic resides in the normal "best efforts" category.
An increasing percentage of the traffic now falls into various higher priority buckets. As described above, this traffic represents data generated by applications that demand high performance routing which may fall into one of the two separate higher priority buckets depicted. Currently all data traffic is treated the same and routed under the best efforts scheme, but a need clearly exists for carriers to be able to offer "differentiated service" - service that differentiates between different types of data, and routes the data accordingly. Such differentiated service would not only make more "real time" applications possible, but it would also decrease the traffic load currently handled by normal routing, thereby increasing the efficiency of normal routing.
A problem faced by carriers wishing to implement some form of differentiated service is that currently they have no way of determining the priority of traffic carried on their networks. This is because typical routing information included in, for example, an Internet Protocol (IP) data packet does not include a priority setting. Methods and systems for offering differentiated service on data networks have been proposed, such as Multi-Protocol Data Switching (MPLS), Differentiated Services Protocol (DiffServ), and the like, however each face inherent difficulties in implementing routing based on QoS.
Accordingly, a need exists for a method and system that enables carriers to provide differentiated data services. A further need exists for carriers to provide QoS guarantees for traffic handed off to peering partners.
Summary of the Invention
The present invention allows carriers to separate their network traffic into differentiated service buckets and trigger layer one topology changes to facilitate the routing/switching of the traffic.
Systems of the present invention interface with network management systems and real-time provisioning systems. By monitoring both the reahtime state information of network resources and priority information associated with data currently being carried over the network, the inventive system determines whether to trigger the provisioning of a "nailed- up" switched path to carry certain high priority data, or to allow the data to be routed in a typical best efforts manner. An object of the present invention is to allow carriers to change the focus of their peering relationships with other networks from maximum network utilization (i.e., pass off as much traffic onto the peering partner as possible in order to free up the maximum amount of on-net resources), to providing the maximum and appropriate QoS to network traffic. In fact, since setting up a switched path for data traffic across a peering partner's network will cost the carrier a premium, it is likely that carriers will attempt to carry priority traffic as close to the destination as possible before handing it off to a peering partner, in order to reduce the premium as much as possible.
A further object of the present invention is to correlate layer three information relating to network traffic with layer one network topology information. This is accomplished by operating in conjunction with an automated provisioning system, such as the one disclosed in applicant' s co-pending applications 60/287, 106, DYNAMIC BANDWIDTH PROVISIONING FOR INTERNET PEERING, and 60/300,527, METHOD AND SYSTEM FOR PROVISIONING OPTICAL CIRCUITS IN A MULTI-NETWORK, MULTI- VENDOR ENVIRONMENT.
Brief Description of the Drawings
Further aspects of the instant invention will be more readily appreciated upon review of the detailed description of the preferred embodiments included below when taken in conjunction with the accompanying drawings, of which:
FIG. 1 is a schematic illustration of existing multi-carrier networks;
FIG. 2 is a chart illustrating approximate proportions of data traffic in existing multi- carrier networks;
FIG. 3 is a schematic illustration of a "nailed up" carrier path in a multi-carrier network according to certain embodiments of the present invention;
FIG. 4 is a is a second schematic illustration of a "nailed up" carrier path in a multi- carrier network according to certain embodiments of the present invention; and
FIG. 5 is a flow chart depicting an exemplary method for providing QoS routing according to certain embodiments of the present invention;
Detailed Description of the Invention
Disclosed herein is a method and system for dynamically establishing a nailed up lightpath based on QoS issues, network congestion, customer or application requests, and the like. Such a method would provide a way of "off-loading" high priority or QoS intensive traffic from the normally routed network onto the nailed up, direct lightpath. The present invention nails up a fixed lightpath, onto which traffic may be diverted, eliminating the need for the layer three routing, so long as the traffic's determined destination is the same as the terminal point of the fixed lightpath.
The present invention provides methods for determining the need for such a lightpath, setting up the lightpath, diverting appropriate traffic onto the lightpath, and returning the network resources that make up the lightpath back to the network. Certain embodiments of the system are advantageous in that the need for optical-electrical-optical (OEO) data format conversions needed to route data over different network layers is reduced or eliminated.
Referring now to FIGS. 1- 5, wherein similar components of the present invention are referenced in like manner, preferred embodiments of a method and system for using layer three correlation with layer one topology are disclosed.
FIG. 3 depicts the peering arrangement between CARRIERS X and Y, similar to that shown in FIG. 1. In the multi-carrier network, according to some embodiments of the present invention, a "nailed up" switched lightpath between A and B is displayed with connections depicted in bold. With the addition of this lightpath, traffic flowing between terminals A and B may pass through the networks of CARRIER X and Y as it did before, finding it's way through the network based on routing and Border Gateway Protocol (BGP) tables, or it may travel directly along the highlighted fiber optic lightpath.
The lightpath has the benefit of being fixed, which typically means it will have a more predictable QoS than routing through current networks. In addition, if for example, CARRIER X requests that a lightpath be set up through CARRIER Y to reach B, CARRIER X may be assured that all traffic sent along the lightpath will always travel the same route. Previously, CARRIER X had no visibility or control into CARRIER Y's routing.
FIG. 4 depicts a similar situation as described in FIG. 3, only now for clarity the fixed lightpath is represented as being totally separated from the remaining networks. Also depicted in FIG. 4 is the Intelligent Node Controller/IntelligentProvisioning Node (INC/IPN) signaling control plane described in applicant's co-pending applications 60/287,106 and 60/300,527. As depicted in FIG. 4, data routed from ROUTER A to ROUTER B may either
enter the network of CARRIER X and find its way through CARRIER Y's network using standard "best effort" routing, or it may travel along the nailed up lightpath. For example, an IP packet sent from the COMPUTER 100 to ROUTER A to ROUTER B to the COMPUTER 200 may either travel through the networks of CARRIERS X and Y, or it may travel along the nailed up lightpath directly from ROUTER A to ROUTER B.
FIG. 5 is a flow chart depicting a method 500 pf routing high bandwidth traffic according to the present invention. The method may be performed by software residing e.g. at an edge router, such as ROUTER A of FIG. 4. The method may start at one of two different steps - START(A) and START(B). The first step moving from START(A) is to identify QoS issues relating to the routing of traffic within a network or across multiple networks (step 502). For example, ROUTER A may run a traffic analyzing application and determine that traffic generated from COMPUTER 100 that is to be sent to COMPUTER 200 is not currently being routed efficiently. The traffic analyzing application may monitor such factors as network congestion and determine whether inefficient routing based on threshold values. According to this J example, should congestion between ROUTER A and ROUTER B reach a certain level, the next step of the invention is triggered. In another example ROUTER A may determine that traffic routed to ROUTER B is experiencing delay, congestion, and the like based on, for example, information received from COMPUTER 200 informing COMPUTER 100 that data is not being received properly. In any event, upon identifying a QoS issue, the next step of the invention is triggered.
The first step moving from START(B) is to receive a request for lightpath set up (step 504). The request may be sent from a bandwidth intensive application, a traffic monitoring application, etc. The purpose of the request is to set up a fixed lightpath between two network points in order to guarantee a certain level of quality in the traffic delivery. For
example, a user may launch a video conferencing application from COMPUTER 100 depicted in FIG. 4. The application may send a request to ROUTER A for a high level of QOS in routing the data generated from the video conferencing session to COMPUTER 200 in order to insure a high quality picture. Regardless of the starting point, the method 500 is next used to identify network resources that may be used to set up a fixed lightpath to ease congestion or provide higher levels of QoS based on information received/determined in the previous steps 502, 504 (step 506). The identified network resources may be completely within one carrier's network, or may span more than one carrier's network. For example, referring to FIG. three, ROUTER A may determine that CARRIER X has the resources within its to set up the needed lightpath, or it may be necessary to utilize resources from CARRIER Y's network. For methods of querying networks to determine whether necessary resources are available, see applicant's co-pending applications. Those applications describe a signaling control plane architecture that is useful for querying network elements within a carrier's network to determine available resources and to activate or provision those resources.
If the necessary resources are available, the next step of the method involves setting up a lightpath utilizing the necessary available resources (step 508). For example, referring again to FIG. 3, should Carrier X determine that it is necessary to offer higher QoS between ROUTER A and ROUTER B, and determine that there are appropriate network resources available in both CARRIER X and CARRIER Y's networks, ROUTER A may signal the control plane (disclosed in applicants co-pending application) to set up a lightpath between ROUTER A and ROUTER B. The control plane reaches deep enough into the carriers' networks to actually provision and activate the necessary equipment to set the lightpath up. It is likely that provisioning a lightpath in another carrier's network will cost a premium, which the carrier/customer requesting the lightpath will need to factor into the decision of offering
the service. Applicant's co-pending applications 60/287,106 and 60/300,527 describe , methods of setting up lightpaths that may span multiple networks utilizing a signaling control plane. Such descriptions are specifically incorporated herein by reference.
The next step of the method 500 involves diverting appropriate traffic away from the normal routing in the network and onto the provisioned lightpath (step 510). According to one embodiment of the invention, ROUTER A sends a message to COMPUTER 100, instructing it to append a lightpath ID on all data packets that are to travel over the lightpath (e.g. all data packets generated by a given application). This lightpath ID may be similar to an MPLS label, and is used by the router to determine which traffic should flow along the lightpath instead of through the network. According to other embodiments, the router may simply use existing header information in each IP packet (e.g. destination) to determine whether routing should take place along the lightpath or in the network. Diverting traffic may involve broadcasting, for example, updates to router/BGP tables.
By diverting traffic along the lightpath, a carrier utilizing the invention may provide higher QoS to prioritized traffic, or may simply off-load a portion of the normal traffic being routed through the netowork, thereby eliminating congestion. The invention accomplishes this by taking information that resides at Layer three of the network (e.g. IP information) and modifying the network(s) Layer one topology (establishing a lightpath).
The final step of the invention involves returmng the provisioned circuits of the lightpath to inventory after completion of data routing or a reduction in network traffic (step 512), after which the method 500 ends.
In another aspect of the invention, layer one resources (i.e. physical resources) that are not currently available to layer three may be made available, based on routing information gathered at layer three. For example, when traffic is routed by a layer three routing device, the router typically accesses a static routing table to determine where to send the data next.
This table may be updated periodically in order to maintain data flow over the network layer, but at any given time, the table is typically based on the physical network resources that are currently available to the router. There may be additional resources (i.e. additional capacity along a known link, additional links, etc.) that could be made available, but currently aren't. In existing technologies, manual analysis must be performed in order to determine how many physical resources to provision, thus requiring network specialists and sophisticated software tools to help plan and determine resource availability. The present invention includes a method of automating that process by linking e.g. a layer three traffic analyzing device, with a layer one provisioning system. In this manner, the layer three device may examine the underlying physical resources in the network and determine whether or not data would more be routed more efficiently with, for example, extra bandwidth provisioned along a given link. An estimated cost of provisioning such bandwidth may be factored into such determination.
Although the invention has been described in detail in the foregoing embodiments, it is to be understood that the descriptions have been provided for purposes of illustration only and that other variations both in form and detail can be made thereupon by those skilled in the art without departing from the spirit and scope of the invention, which is defined solely by the appended claims.