Abstract: Techniques for orchestrating workloads based on policy to operate in optimal host and/or network proximity in cloud-native environments are described herein. The techniques may include receiving flow data associated with network paths between workloads hosted by a cloud-based network. Based at least in part on the flow data, the techniques may include determining that a utilization of a network path between a first workload and a second workload is greater than a relative utilization of other network paths between the first workload and other workloads. The techniques may also include determining that reducing the network path would optimize communications between the first workload and the second workload without adversely affecting communications between the first workload and the other workloads. The techniques may also include causing at least one of a redeployment or a network path re-routing to reduce the networking proximity between the first workload and the second workload.

Abstract: Techniques for tunneling Layer 2 ethernet frames over a connection tunnel using the MASQUE protocol are described herein. The MASQUE protocol may be extended to include a new entity, configured to proxy ethernet frames using a MASQUE proxy connection, and an associated CONNECT method, CONNECT-ETH. Using the extended MASQUE protocol, an Ethernet over MASQUE (EoMASQUE) tunnel may then be established between various networks that are remote from one another and connected to the internet. An EoMASQUE tunnel, established between separate remote client premises, and/or between a remote client premise and an enterprise premise, may tunnel ethernet packets between the endpoints. Additionally, a first EoMASQUE tunnel, established between a first client router provisioned in a first remote client premise and an EoMASQUE proxy node, and a second EoMASQUE tunnel, established between a second client premise and the EoMASQUE proxy node, may tunnel ethernet packets between the first and second client premise.

Abstract: Techniques for managing migrations of QUIC connection session(s) across proxy nodes, data centers, and/or private application nodes are described herein. A global key-value datastore, accessible by proxy nodes and/or application nodes, may store mappings between a first QUIC connection, associated with a proxy node and a client device, on the frontend of the proxy node and a second QUIC connection, associated with the proxy node and an application node, on the backend of the proxy node. With the global key-value datastore being accessible by the proxy nodes, when a proxy node receives a QUIC packet on the front end or the back end, the proxy node may determine where to map this connection to on the opposite end. Additionally, with the global key-value datastore being accessible to the application nodes, when an application node receives a QUIC packet, the application node may determine the client device associated with the connection.

Abstract: Techniques for dynamically load balancing traffic based on predicted and actual load capacities of data nodes are described herein. The techniques may include determining a predicted capacity of a data node of a network during a period of time. The data node may be associated with a first traffic class. The techniques may also include determining an actual capacity of the data node during the period of time, as well as determining that a difference between the actual capacity and the predicted capacity is greater than a threshold difference. Based at least in part on the difference, a number of data flows sent to the data node may be either increased or decreased. Additionally, or alternatively, a data flow associated with a second traffic class may be redirected to the data node during the period of time to be handled according to the first traffic class.

Abstract: Systems, methods, and computer-readable media are provided for reusing execution environments and code of serverless functions while ensuring isolation in serverless computing environments. In some examples, a method can include, in response to a first request to run a serverless function, executing, at an execution environment on a network, computer-readable code configured to perform the serverless function; after the computer-readable code has executed, modifying a pointer to an area of memory used to store a first state of the serverless function to reference a different area of memory; in response to a second request to run the serverless function, reusing, at the execution environment, the computer-readable code to perform the serverless function; and based on the pointer referencing the different area of memory, using the different area of memory to store a second state of the serverless function.

Abstract: Techniques for dynamically load balancing traffic based on predicted and actual load capacities of data nodes are described herein. The techniques may include determining a predicted capacity of a data node of a network during a period of time. The data node may be associated with a first traffic class. The techniques may also include determining an actual capacity of the data node during the period of time, as well as determining that a difference between the actual capacity and the predicted capacity is greater than a threshold difference. Based at least in part on the difference, a number of data flows sent to the data node may be either increased or decreased. Additionally, or alternatively, a data flow associated with a second traffic class may be redirected to the data node during the period of time to be handled according to the first traffic class.

Abstract: A method includes receiving a DNS request, notifying a serverless orchestrator system of data associated with the DNS request, provisioning a function on a serverless function node based on the DNS request, notifying a load balancer regarding the serverless function node, providing a response to the DNS request and routing an API request associated with the DNS request to the serverless function node.

Abstract: Techniques for routing service mesh traffic based on whether the traffic is encrypted or unencrypted are described herein. The techniques may include receiving, from a first node of a cloud-based network, traffic that is to be sent to a second node of the cloud-based network and determining whether the traffic is encrypted or unencrypted. If it is determined that the traffic is encrypted, the traffic may be sent to the second node via a service mesh of the cloud-based platform. Alternatively, or additionally, if it is determined that the traffic is unencrypted, the traffic may be sent to the second node via an encrypted tunnel. In some examples, the techniques may be performed at least partially by a program running on the first node of the cloud-based network, such as an extended Berkeley Packet Filter (eBPF) program, and the like.

Abstract: Techniques for combining independent sessions between application(s) and a VPN, proxy service, or similar system, including inner protocol sessions (e.g., such as QUIC, etc.), coming from a single device to form a single logical session, where the single logical session could share a single authentication/authorization token are described. The techniques include receiving, from a device within a network, a request for a first application to access a service associated with the proxy service or the VPN, sending, to the device, a first authentication request, and receiving, from the device, a message including a token. The techniques may further include authenticating, by the proxy service or the VPN, the token using a unique identifier associated with the device and enabling, by the proxy service or the VPN, the device to access the service via a first session flow.

Abstract: Techniques for managing migrations of QUIC connection session(s) across proxy nodes, data centers, and/or private application nodes are described herein. A global key-value datastore, accessible by proxy nodes and/or application nodes, may store mappings between a first QUIC connection, associated with a proxy node and a client device, on the frontend of the proxy node and a second QUIC connection, associated with the proxy node and an application node, on the backend of the proxy node. With the global key-value datastore being accessible by the proxy nodes, when a proxy node receives a QUIC packet on the front end or the back end, the proxy node may determine where to map this connection to on the opposite end. Additionally, with the global key-value datastore being accessible to the application nodes, when an application node receives a QUIC packet, the application node may determine the client device associated with the connection.

Abstract: Techniques for using global virtual network instance (VNI) labels in a multi-domain network to route network data with a multi-tenant network overlay are described herein. A routing device provisioned in a network domain of the multi-domain network may register with a service discovery system of the network domain for use of network configuration data to establish routes through the multi-domain network with network nodes. Each network domain of the multi-domain network may include an application programming interface (API) server for processing API requests to make changes to configurations of a network domain. A border gateway protocol (BGP) large community may be utilized to encode global VNI labels, network addresses, local next hop nodes, and/or additional network information and sent to routing devices provisioned in separate network domains. A service chain may be signaled by global VNI labels to route network traffic through various services prior to reaching a destination endpoint.

Abstract: Techniques for orchestrating workloads based on policy to operate in optimal host and/or network proximity in cloud-native environments are described herein. The techniques may include receiving flow data associated with network paths between workloads hosted by a cloud-based network. Based at least in part on the flow data, the techniques may include determining that a utilization of a network path between a first workload and a second workload is greater than a relative utilization of other network paths between the first workload and other workloads. The techniques may also include determining that reducing the network path would optimize communications between the first workload and the second workload without adversely affecting communications between the first workload and the other workloads. The techniques may also include causing at least one of a redeployment or a network path re-routing to reduce the networking proximity between the first workload and the second workload.

Abstract: Techniques for integrating disparate headend traffic ingress services with disparate backend services are disclosed herein. The techniques may include receiving, at a classification and forwarding node of a networked computing environment, a data packet encapsulated according to a first encapsulation protocol that is supported by the classification and forwarding node. The techniques may also include determining, by the classification and forwarding node, that the data packet is to be sent to a service from among a group of services associated with the networked computing environment. The classification and forwarding node may also determine whether the first encapsulation protocol is supported by the service. Based at least in part on determining that the service supports a second encapsulation protocol different than the first encapsulation protocol, the classification and forwarding node may encapsulate the data packet according to the second encapsulation protocol and send the data packet to the service.

Abstract: A system receives a first request from a first instance of a network function associated with a first address. The system may determine the first address and, based at least in part on the first address, may identify a second address with which to respond to the first request. The system may then send, to the first instance of the network function, a response to the first request specifying the second address. The system may also receive a second request from a second instance of the network function associated with a third address. The system may determine a fourth address with which to respond to the second request, and may thereafter send a response to the second request to the second instance of the network function, with the response specifying the fourth address.

Abstract: Systems, methods, and computer-readable media are provided for reusing execution environments and code of serverless functions while ensuring isolation in serverless computing environments. In some examples, a method can include, in response to a first request to run a serverless function, executing, at an execution environment on a network, computer-readable code configured to perform the serverless function; after the computer-readable code has executed, modifying a pointer to an area of memory used to store a first state of the serverless function to reference a different area of memory; in response to a second request to run the serverless function, reusing, at the execution environment, the computer-readable code to perform the serverless function; and based on the pointer referencing the different area of memory, using the different area of memory to store a second state of the serverless function.

Abstract: Techniques for migrating on-premises and/or cloud-based workloads to follow a network session as it potentially migrates, due to multipathing techniques, across multiple edge and/or cloud datacenters. The techniques may include determining, by a controller of a network, that a traffic flow between an endpoint device and a workload has migrated to a different path of a multipath flow such that the traffic flow terminates at a different termination point than the workload. Based at least in part on determining that the traffic flow has migrated, the controller may cause a migration of a state of the workload to a location associated with the different termination point. That is, the controller may cause the workload to be migrated in its current state, which may be specific to the endpoint device, to follow the traffic flow.

Abstract: Techniques for dynamically load balancing traffic based on predicted and actual load capacities of data nodes are described herein. The techniques may include determining a predicted capacity of a data node of a network during a period of time. The data node may be associated with a first traffic class. The techniques may also include determining an actual capacity of the data node during the period of time, as well as determining that a difference between the actual capacity and the predicted capacity is greater than a threshold difference. Based at least in part on the difference, a number of data flows sent to the data node may be either increased or decreased. Additionally, or alternatively, a data flow associated with a second traffic class may be redirected to the data node during the period of time to be handled according to the first traffic class.

Abstract: Techniques for using global virtual network instance (VNI) labels in a multi-domain network to route network data with a multi-tenant network overlay are described herein. A routing device provisioned in a network domain of the multi-domain network may register with a service discovery system of the network domain for use of network configuration data to establish routes through the multi-domain network with network nodes. Each network domain of the multi-domain network may include an application programming interface (API) server for processing API requests to make changes to configurations of a network domain. A border gateway protocol (BGP) large community may be utilized to encode global VNI labels, network addresses, local next hop nodes, and/or additional network information and sent to routing devices provisioned in separate network domains. A service chain may be signaled by global VNI labels to route network traffic through various services prior to reaching a destination endpoint.

Abstract: Techniques for detecting inactive peers of a tunneled communication session, while allowing for a scalable tunneled protocol that includes split control plane nodes and data plane nodes are described herein. A method according to a technique described herein may include establishing a communication session between a first node and a second node in a network such that control plane traffic of the communication session flows through one or more control nodes and data plane traffic of the communication session flows through one or more data nodes different than the one or more control nodes. The method may also include receiving, at a control node, an indication from a data node that a probe message is to be generated. The probe message may be configured to determine data plane connectivity in the communication session. Additionally, the control node may generate the probe message and send it to the first node.

Abstract: Techniques for using global virtual network instance (VNI) labels in a multi-domain network to route network data with a multi-tenant network overlay are described herein. A routing device provisioned in a network domain of the multi-domain network may register with a service discovery system of the network domain for use of network configuration data to establish routes through the multi-domain network with network nodes. Each network domain of the multi-domain network may include an application programming interface (API) server for processing API requests to make changes to configurations of a network domain. A border gateway protocol (BGP) large community may be utilized to encode global VNI labels, network addresses, local next hop nodes, and/or additional network information and sent to routing devices provisioned in separate network domains. A service chain may be signaled by global VNI labels to route network traffic through various services prior to reaching a destination endpoint.