Electrical computers and digital processing systems: multicomput – Remote data accessing – Using interconnected networks
Reexamination Certificate
1998-12-29
2002-11-12
Rinehart, Mark H. (Department: 2152)
Electrical computers and digital processing systems: multicomput
Remote data accessing
Using interconnected networks
C709S223000
Reexamination Certificate
active
06480888
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to the field of interworked networks such as coupled asynchronous transfer mode networks and frame relay networks and to providing improved concentration and efficiency in such networks via deployment in an interworked network of a loopback arrangement of a virtual path concentrator node.
In interworked networks, frequently a high speed central network, for example, an asynchronous transfer mode (ATM) network, the interworked network provides high-speed interconnection for a large number of slower remote ports, for example frame-relay (FR) ports. Referring to
FIG. 1
, a number of FR ports are tied through permanent virtual connections (PVC) and a network to network interface (NNI) to an ATM network. The virtal circuits (VCs) terminate at an ATM port. The ATM network forms virtual paths (VP) that tie together remote ports throughout the ATM network.
FIG. 1
illustrates the general case of two networks, Networks A and B, connected via NNI ports according to the prior art. Each of the depicted Networks A and B has a set of access nodes or switches (AS) providing connectivity to customer premises equipment (CPE), not shown, of various customers. For example, Network A comprises Access Nodes
101
-
1
to
101
-N coupled to the left to customer premises equipment (not shown). Each network also has gateway switches (GS) which host the NNI ports used for routing of internetwork connections. For example, Network A shows AS's
101
-
1
to
101
-
5
connected via PVC's, shown as solid lines, to gateway switch
105
-
1
and AS
101
-N connected via three PVC's to GS
105
-
3
. Additional switches (not shown) may serve as routing nodes between the access and gateway nodes. The categorization of node types (access, routing, gateway) is only for clarity of presentation. Any switch can concurrently provide all three functions.
As depicted, a PVC may couple an AS of network A via a GS, an NNI port and a GS of Network B to an AS of Network B. Each PVC is provisioned as multiple segments. A segment is shown in network A leading from access node (AS) to gateway (GS) of Network A. A companion segment is provisioned in network B from access node (AS) to gateway (GS) of Network B. The NNI ports of each network are connected directly together in order to provide end-to-end connectivity between network A and network B access points.
The NNI ports may have two provisioning limitations: total assigned bandwidth and connection count. Although the bandwidth of a port is fixed, total assigned bandwidth is generally a “soft” limit as ports can be overbooked when adequate traffic management mechanisms allow it. That is, although a port may only offer (for the case of DS3) 36 MHZ of payload bandwidth, a total of 72 MHZ worth of PVCs may be provisioned on the port if they have utilization factors of less than 50 percent.
The connection count limit, however, is a “hard” limit stemming from the port card design. If a port is limited to, for example, 1000 connection terminations, that limit will prevent further provisioning on the port irrespective of the bandwidth utilization.
For the case of a DS3 port with no overbooking, a 1000 PVC connection limit would result in an average rate of 36 Kbps per PVC (36 MHZ of payload bandwidth divided by 1000 PVCs). If PVCs actually average 18 Kbps, port bandwidth would only be 50 percent utilized. The remainder would be wasted.
Further growth of frame relay services, for example, is constrained by a number of factors including platform element connection count limits as discussed above and associated routing algorithm performance issues. High connection counts impact the performance of frame relay service routing and failure recovery processes. These processes run slower and suffer from more “collisions” as connection counts increase. Without relief from the impact of very high connection counts, multiple independent networks may have to be created, operated and maintained in the near future.
Consequently, there is a need in the art to relax some of the constraints such as connection count limits which threaten the expansion and scaling of frame relay service. Thus, there is an opportunity in the art to improve the efficiency of routing traffic with varying bandwidth requirements and for different customers through an improved interworked network architecture.
SUMMARY OF THE INVENTION
According to the principles of the present invention, there are provided virtual path concentrator nodes deployed, for example, between access nodes and gateway switches of a given interworked network which provide circuit aggregation capability. An interworked network comprises a first network having access nodes and a gateway node and a second network having access nodes and a gateway node coupled to the gateway node of the first network via a network to network interface such that a virtual path concentrator node is coupled between an access node and the gateway node of the first network, the virtual path concentrator node providing for the purpose of creation and loading of virtual path circuits between virtual path concentrator nodes and gateway nodes of the first network. Virtual path identifier assignment at endpoints is provided according to a global assignment process of one or more global VPIs for each destination node as described in copending, concurrently filed U.S. patent application Ser. No. 09/221,856, of P. Nicoll and J. Pedersen for Global Addressing and Identifier Assignment in Inter-Worked Networks, which is incorporated herein by reference as to its entire contents. This global VPI is used by every node in the interworked network for routing including the deployed VPCNs. Either the first or the second networks of the interworked network may be an asynchronous transfer mode network or a frame relay network or a combination of asynchronous transfer mode and frame relay networks.
While there is presented herein an example of two interworked networks (A and B) as the generalized case, using the VPCN for circuit aggregation may also be applied within a single network, for example, either network A or network B, to reduce connection counts and thereby minimize loading upon network routing and failure recovery algorithms. The present invention should not be deemed to be limited in respect to only intereworked networks. The loading reduction in any single network or an interworked network is achieved because a virtual path circuit (VPC) carrying multiple customer virtual channel connections (VCC's) counts as a single circuit from the perspective of network routing and failure recovery algorithms.
In one customer application and assuming two hundred fifty permanent virtual circuits may be handled per customer router port, twelve virtual paths may be created at four frame relay virtual path concentrator nodes according to the present invention to route three thousand permanent virtual circuits over two (provided for diversity) frame relay to asynchronous transfer mode network to network interface ports. Similar arrangements could be used to route an additional three thousand permanent virtual circuits to another of the customer's data centers and so on.
The basic principles of applying loopback for providing virtual path connections and saving connection count may be applied to the design of an interworked network. As will be further described herein, an interworked network may be designed to comprise a central core of virtual path concentrator nodes for providing virtual paths between the nodes with surrounding frame relay and/or asynchronous transfer mode networks taking advantage of the central core for saving connection count. Dramatic improvements in internetwork operating efficiency and maintenance will result from building such a network. Flow control and operations, administration and maintenance loops may be provisioned as end-to-end, customer to VPCN, and VPCN to VPCN (among others) closed loops for traffic congestion signaling and improved maintenance. Other advantages of the present inven
AT&T Corp.
Rinehart Mark H.
Vu Thong
LandOfFree
Virtual path concentrator node with loopback does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Virtual path concentrator node with loopback, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Virtual path concentrator node with loopback will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2929067