Future Wei Technologies, Inc.

Huawei Cloud Computing Technologies Co., Ltd.

Cisco Technology, Inc.

Farhad P Sunavala

Hong Zhang

Henry Louis Fourie

Ying Xiong

Mehdi Frederic Sif

Mark Xiang-Quan Huang

Robert A Mackie

Bhaskar Bhupalam

Walter Glenn Dixon

Hongbo Tian

Houxiao Han

Bhagatram Yaugand Janarthanan

Kevin Franklin Clayton

Mark Huang

Prakash Ramchandran

Jorge Manuel Gonzalez

Honglin Li

Fei Rao

Galen Kim Davis

A computer-implemented method for accessing user resources in a virtual private cloud (VPC) using a serverless function within a network architecture includes instantiating a first warm application container for hosting the serverless function. The first warm application container includes a runtime language library without function code of the serverless function. A virtual machine for hosting a Port Address Translation (PAT) gateway is instantiated. The PAT gateway includes a first interface to the VPC and a second interface to the first warm application container. In response to detecting a trigger event for triggering the serverless function, the function code of the serverless function is mounted within the first warm application container. During execution of the function code from the first warm application container, VPC-addressed network packets associated with the serverless function are routed to the VPC via the second interface and the first interface within the PAT gateway.

Mechanism to reduce serverless function startup latency

A system, computer-implemented method, and/or non-transitory computer-readable media may provide a warm start technique for cloud-hosted functions. Included is a non-transitory memory storage comprising instructions, and one or more processors in communication with the memory, wherein the one or more processors execute the instructions to identify a cloud-hosted function configured to be executed within a cloud computing environment. The cloud-hosted function is instantiated within the cloud-hosted computing environment.

Instantiation of cloud-hosted functions based on call to function graph

A system, computer-implemented method, and/or non-transitory computer-readable media may provide a warm start technique for cloud-hosted functions. A call is received from a web application. In response to the received call, a function graph is activated and includes a collection of functions associated with the web-application. The function graph also includes a plurality of states and transitions between the states. In response to the activation of the function graph, a cloud-hosted function associated with a particular state in the plurality of states included in the function graph is determined Before the particular state is triggered within the function graph to execute the cloud-hosted function, the determined cloud-hosted function is instantiated.

A computer-implemented method for processing serverless functions includes mapping a received event to an event state of a plurality of event states in a function graph according to a mapping rule, the function graph including one or more actions for the event state. A data package of a previous event state of the plurality of event states is filtered to generate a filtered data package, using a payload filter associated with the event state. The actions of the event state are executed, where the one or more actions satisfy the mapping rule. A modified data package is sent to a computer system to trigger executing one or more serverless functions associated with the executed one or more actions. The modified data package is based on a data package of the event and the filtered data package. One or more responses are received based on execution of the serverless functions.

Event-driven serverless function orchestration

An architecture, method, and/or non-transitory computer-readable media provide a service graph based serverless cloud architecture. The service graphs can orchestrate cloud-hosted functions into a coordinated micro-service application, and thus control the execution of the cloud-hosted functions in a prescribed manner. Additionally, service graphs permit a user to arrange cloud-hosted functions to execute in sequence or concurrently, manage error conditions by re-invoking function calls, handle scaling to accommodate varying event loads, and the like. Service graphs also allow a user to define rendezvous points (i.e., states) to wait for pre-defined events before executing a cloud-hosted function and progressing through the service graph. The advantages of coordinating the execution of cloud-hosted functions with service graphs is that the service graphs provide a consistent framework for managing cloud-hosted functions without the user having to solve the coordination problems on their own.

Service graph based serverless cloud platform

The disclosure relates to technology for coordinating execution of serverless functions. One or more events are received from one or more event sources. The one or more events are mapped to one or more event states of a function graph according to a mapping rule, the one or more event states including one or more actions, and the one or more actions which satisfied the mapping rule are executed. The received events are sent to a computer system to trigger the computer system initiating one or more computing resource groups for executing one or more serverless functions associated with the one or more actions. One or more responses are received from the serverless functions.

The disclosure relates to technology for coordinating execution of serverless functions. One or more events are received from one or more external sources. The one or more events are mapped to one or more event states of a processing graph according to mapping rules, the one or more event states including one or more actions, and the one or more actions are executed in response to the one or more events satisfying conditions of the mapping rules. An event response is the received in reply to the received one or more events, where the event response is based on execution of one or more tasks corresponding to a sequence of the executed one or more actions.

A method for service function chaining across subnetworks includes receiving a packet at a virtual switch integration bridge from a first service function (SF) that is in a service function chain (SFC) and that is on a first subnetwork, determining a next SF in the SFC in a different subnetwork, and sending the received packet directly from the virtual switch integration bridge to the next SF.

Service function chaining across multiple subnetworks

A method and a system for intelligent mobile function distribution are disclosed. An embodiment includes discovering, by an Intelligent Mobile Function Distribution (iMOD) entity, resources available in a mobile network and moving mobile functions across the mobile network in accordance with the discovered resources, thereby allowing the mobile network to adapt to changing conditions.

Methods and systems for intelligent mobile function distribution

Packets of data can be processed by a chain or sequence of service functions which include both virtual and physical service functions. In one approach, the original header and/or payload of a packet is analyzed at a software switch to determine whether the packet should be processed by the chain. If such processing is indicated, a next service function to process the packet is identified. The packet is then forwarded to a virtualized or physical device which implements the next service function. A portion of the header, such as a MAC destination address, may be modified to direct the packet to the next function. Further, in some cases, a service function chain (SFC) header with a chain identifier may be added to, or removed from, the packet.

Integrating physical and virtual network functions in a service-chained network environment

In one embodiment, a method for mobile network function virtualization (MNFV) includes creating an evolved packet core (EPC) cluster and associating a sub-network with the EPC cluster. The method also includes booting a virtual machine (VM) and attaching the VM to the EPC.

System and method for mobile network function virtualization

A virtual network function includes network interface via a single client network interface coupled to a network functions infrastructure. The network interface is provided using a virtual network function abstraction layer in the form of an implementation of a virtual network function in a virtual machine nested within a second virtual machine having a virtual switch. Routing between any number of ports on the virtual network function and a single network interface is enabled by providing a virtual interface in the switch mapped to each network interface needed by the virtual network function. Data from a client network via the network interface for the infrastructure is routed to the proper virtual network function interface in the virtual switch.

Single network interface for multiple interface virtual network functions

A packet forwarding method that includes receiving a packet from a traffic classifier at a host that comprises a virtual machine (VM), and a first hypervisor for the host that provides a first set of routing instructions that provides routing instruction to forward the packet to a port of the VM, and wherein the VM comprises a plurality of virtual network functions (VNFs), and a second hypervisor for the VM that provides a second set of routing instructions that provides routing instructions to forward the packet to the plurality of VNFs based on an input, processing, by the host, the packet using the plurality of VNFs in accordance with the first set of routing instructions and the second set of routing instructions, and forwarding, by the host, the processed packet to a network device.

Service function bundling for service function chains

A first set of frames is received in a data flow and tagged with respective serial numbers identifying an ordering of frames in the first set. The tagged first set is forwarded to a general processor for processing prior to being sent to a remote network element. A data flow policy is received from the general processor based on processing of the first set. A second set of frames of the data flow are received subsequent to forwarding the first set to the general processor. It can be verified, based on the assigned serial numbers, that each frame in the first set has been sent to the remote network element prior to sending frames in the second set. The second set is processed and sent to the remote network element by the network processor, bypassing the general processor, based on the data flow policy.

Growing community of inventors

San Ramon, CA, United States of America

Farhad P Sunavala