Cisco Technology, Inc.

Microsoft Technology Licensing, LLC

Yixing Ruan

Carlos M Pignataro

James Neil Guichard

Javed Asghar

Paul M Quinn

Reinaldo Penno

Pradeep S Sindhu

Rajiv Asati

Deepak Goel

Wael Noureddine

Srihari Raju Vegesna

Philip A Thomas

Ayaskant Pani

Wojciech Dec

Narendra Jayawant Gathoo

Gopesh Goyal

Robert William Bowdidge

Pranav Dharwadkar

Kenneth Gray

Kirill Kogan

Hari K Lalgudi

Chran Ham Chang

A fabric control protocol (FCP) and packet forwarding mechanisms are described that maximize utilization of bandwidth within massive, large-scale data centers having multi-stage data center switch fabric topologies, such as topologies that include a third switching layer formed by super spine switches. Automatic generation of data plane forwarding information referred to as FCP path information enumerates, for each data processing unit (DPU), the available FCP paths. The FCP path information may be based on unique combinations of peak points of the switch fabric for a given DPU with FCP colors assigned to network links that are used to multi-home the DPU to the switch fabric.

Fabric control protocol for large-scale multi-stage data center networks

Embodiments of the present disclosure are directed to augmenting a Network Service Header (NSH) metadata of a data packet with a virtual routing and forwarding identifier (VRF-ID) and forgoing augmenting a virtual private network (VPN) label into a multiprotocol label switched (MPLS) metadata of the data packet. A provider edge router can use the VRF-ID to identify a next hop for the data packet as a service to be applied prior to forwarding the data packet to a VPN site.

Extending NSH services into the VPN L2/L3 domain

Aspects of the embodiments are directed to augmenting a control packet with an interface identifier, the interface identifier identifying an interface at a physical network forwarding element; and transmitting the control packet with the interface identifier to the physical network forwarding element. The interface identifier can be included in metadata of a network service header (NSH). The NSH is encapsulated with the control packet, which is transmitted with the control packet. The NSH can be extracted and the interface identifier used to identify a user interface (or a presenting interface) based on a metadata lookup.

Using a network service header to manage a network-as-a-system

A method for applying network services to data traffic forwarded between virtual private network (VPN) sites includes: receiving a data packet addressed to a target site associated with the VPN, determining services to be applied to the data packet according to a service chain, where the determining is a function of at least one of the VPN, the origin site or the target site, adding an indication of a VPN forwarding context onto the data packet, encapsulating the data packet with Network Service Header encapsulation, where a header for the encapsulated data packet indicates at least the service chain; forwarding the encapsulated data packet in accordance with the service chain, receiving the encapsulated data packet at the end of the service chain, terminating the service chain, removing the encapsulation, and forwarding the data packet to a target destination per the indication of a VPN forwarding context.

NSH service plane for L3VPN forwarding

In one embodiment, a provider edge (PE) device in a computer network determines an IPv4 address and link-layer address for each adjacent customer premise equipment (CPE) device, and assigns each CPE device a unique IPv6 address. The PE device stores a key-pair mapping between the unique IPv6 address and combined IPv4 and link-layer address for each adjacent CPE, the mapping bound by a CPE session context, and uses the CPE session context to convert between IPv4 and IPv6 for all network traffic to and from a particular CPE device.

Distributed mapping of address and port (MAP) between a provider edge device and customer premise equipment devices

Mechanisms for programming and performing combined interface and non-interface specific associative memory lookup operations for processing of packets are disclosed. One system includes multiple interfaces, a content-addressable memory, multiple memory entries and a lookup mechanism. The content-addressable memory includes multiple interface independent entries, multiple first interface dependent entries corresponding to the first interface, and multiple second interface dependent entries corresponding to the second interface. The lookup mechanism is configured to initiate lookup operations in order to produce the interface independent and interface dependent results.

Combined interface and non-interface specific associative memory lookup operations for processing of packets

Methods and apparatus are disclosed for use with an associative memory, such as for, but not limited to implementing access control list and quality of service features in a communications or computing device. Multiple lists, such as access control lists, may be manipulated to typically produce a single list of entries with continuation indications, or the lists might be provided from another source. An associative memory is programmed with entries with each entry typically including a corresponding continuation level indication or flag. One or more lookup words are then generated and provided to the associative memory for a particular packet (or other entity) corresponding to the different continuation levels. In one implementation, a modified version of the Order Dependent Merge technique is used to generate the list of entries and to identify the corresponding continuation level or other continuation indication and the corresponding action to be performed for each entry.

Methods and apparatus for performing continue actions using an associative memory which might be particularly useful for implementing access control list and quality of service features

Growing community of inventors

San Jose, CA, United States of America

Yixing Ruan