Abstract: Techniques are described for providing a distributed application load-balancing architecture that supports multipath transport protocol for client devices connecting to an application service. Rather than having client devices generate new network five-tuples for new subflows to the application servers, the techniques described herein include shifting the burden to the application servers to ensure that the new network five-tuples land in the same bucket in the consistent hashing table. The application servers may receive a hashing function utilized by the load balancers to generate the hash of the network five-tuple. By having the application servers generate the hashes, the load balancers are able to continue stateless, low-level processing of the packets to route them to the correct application servers. In this way, additional subflows can be opened for client devices according to a multipath transport protocol while ensuring that the subflows are routed to the correct application server.

Abstract: Systems, methods, and computer-readable media are provided for load balancing requests and controlling object replication based on object popularity. A request for an object can be received at a dispatcher of a storage system from a client. Candidate storage nodes of the storage system for serving the object can be identified by the dispatcher by generating an ordered list of the candidate storage nodes using a two-dimensional consistent hashing function. Distribution of the request for the object through one or more candidate storage nodes for filling the request for the object can be facilitated according to the ordered list of candidate storage nodes. Specifically, the one or more candidate storage nodes can be configured to facilitate distribution of the request by selectively filling the request to the client using cache admission policies formed based on popularity characteristics of requested objects at the one or more candidate storage nodes.

Abstract: This disclosure describes methods and systems to externally manage network-to-network interconnect configuration data in conjunction with a centralized database subsystem. An example of the methods includes receiving and storing, in the centralized database subsystem, data indicative of user intent to interconnect at least a first network and a second network. The example method further includes, based at least in part on the data indicative of user intent, determining and storing, in the centralized database subsystem, a network intent that corresponds to the user intent. The example method further includes providing data indicative of the network intent from the centralized database subsystem to a first data plane adaptor, associated with the first network, and a second data plane adaptor, associated with the second network.

Abstract: In one embodiment, a traffic analysis service obtains traffic characteristics of network traffic associated with a device in a network. The traffic analysis service uses a machine learning model to infer resource usage by the device based on the obtained traffic characteristics of the network traffic associated with the device. The traffic analysis service controls traffic flows in the network based on the inferred resource usage by the device.

Abstract: This disclosure describes techniques for providing a distributed scalable architecture for Network Address Translation (NAT) systems with high availability and mitigations for flow breakage during failover events. The NAT servers may include functionality to serve as fast-path servers and/or slow-path servers. A fast-path server may include a NAT worker that includes a cache of NAT mappings to perform stateful network address translation and to forward packets with minimal latency. A slow-path server may include a mapping server that creates new NAT mappings, depreciates old ones, and answers NAT worker state requests. The NAT system may use virtual mapping servers (VMSs) running on primary physical servers with state duplicated VMSs on different physical failover servers.

Abstract: Systems, methods, and computer-readable media for load balancing using segment routing and application monitoring. A method can involve receiving a packet including a request from a source device to an application associated with a virtual address in a network, mapping the request to a set of candidate servers hosting the application associated with the virtual address, and encoding the set of candidate servers as a list of segments in a segment routing header associated with the packet. The method can further involve determining that a first candidate server from the set of candidate servers is a next segment in the list of segments, encoding the first candidate server in a destination address field on a header of the packet, and forwarding the packet to the first candidate server.

Abstract: Techniques for communicating packets in a computer network are provided. At a network device, a first stream of packets is obtained from a packet processing node. A second stream of packets is obtained from a packet generating node. A counter is maintained that counts a number of the packets received from the second stream. An internal clock signal of the network device is obtained. A control signal is generated for pacing the first stream of packets based on the counter and the internal clock signal. The first stream of packets is provided to a packet consuming node based on the control signal.

Abstract: Embodiments provide techniques for delivering a paced stream of video data packets. One embodiment includes receiving a data stream of video data packets formatted according to a Society of Motion Picture and Television Engineers (SMPTE) standard. A desired rate of delivery for the video data packets at a gateway device is determined. Embodiments generate a padded data stream by inserting one or more pause frames in between the video data packets in the received data stream, based on the desired rate of delivery. The padded data stream is transmitted across a link to a network switch, to be transmitted to the gateway device.

Abstract: Systems, methods, and computer-readable media are provided for predictive content pre-fetching and allocation of resources for providing network service access. In some examples, traffic in a network environment is monitored and a related network service to a requested network service is recognized. A UDP probe for the related network service is sent to at least one candidate server of a plurality of candidate servers within the network environment. A candidate server of the plurality of candidate servers is selected for provisioning of the related network service. The candidate server gathers one or more pre-fetched resources for provisioning the related network service. Accordingly, traffic associated with provisioning of the related network service can be steered to the candidate server by a load balancer for provisioning of the related network service using the one or more pre-fetched resources.

Abstract: Systems, methods, and computer-readable media are provided for load balancing requests and controlling object replication based on object popularity. A request for an object can be received at a dispatcher of a storage system from a client. Candidate storage nodes of the storage system for serving the object can be identified by the dispatcher by generating an ordered list of the candidate storage nodes using a two-dimensional consistent hashing function. Distribution of the request for the object through one or more candidate storage nodes for filling the request for the object can be facilitated according to the ordered list of candidate storage nodes. Specifically, the one or more candidate storage nodes can be configured to facilitate distribution of the request by selectively filling the request to the client using cache admission policies formed based on popularity characteristics of requested objects at the one or more candidate storage nodes.

Abstract: Disclosed is a method that modifies a bit indexed explicit replication (BIER) algorithm. The method includes receiving a packet at a node, wherein the packet includes a BIER header identifying a bitstring, the bitstring including a first bit indicating a first destination and a second bit indicating a second destination and forwarding the packet through one or more networks toward the first destination and the second destination based on the bitstring and a predetermined bit selection order. The predetermined bit selection order causes a sequential delivery of the packet to the first destination and the second destination. After the packet arrives at the first destination, the method includes setting the first bit to zero in the bitstring and forwarding the packet through the one or more networks toward the second destination according to the updated bitstring.

Abstract: A message is received which indicates a request for a client-specific service address for service or content provided by a service provider. In response to the request, a client-specific service address may be generated and sent to the client. The address may be used as a destination address in one or more subsequent client requests for service or content. A first portion of the address comprises an IPv6 service prefix assigned to a service network of the service provider. A second portion of the address comprises semantic information having a first portion of encrypted private information and a second portion of unencrypted service information. The encrypted private information may be generated by encrypting private information based on a cryptographic key, where the cryptographic key is derived based on a secret key associated with the service provider and an IP client prefix assigned to the client.

Abstract: The present technology provides a system, method and computer readable medium for steering a content request among plurality of cache servers based on multi-level assessment of content popularity. In some embodiments a three levels of popularity may be determined comprising popular, semi-popular and unpopular designations for the queried content. The processing of the query and delivery of the requested content depends on the aforementioned popularity level designation and comprises a acceptance of the query at the edge cache server to which the query was originally directed, rejection of the query and re-direction to a second edge cache server or redirection of the query to origin server to thereby deliver the requested content. The proposed technology results in higher hit ratio for edge cache clusters by steering requests for semi-popular content to one or more additional cache servers while forwarding request for unpopular content to origin server.

Abstract: Techniques for zero-loss workload mobility with segment routing for virtual machines are presented. The techniques include receiving, by a virtual router, an electronic message destined for a first virtual machine running on a first physical machine and checking a first virtual machine state for the first virtual machine. In response to determining that it is associated with a running state indicating the first physical machine, inserting a segment routing header including an indication of the source virtual machine, the first physical machine, and the first virtual machine. In response to determining that it is associated with a migration state, inserting, by the virtual router, a segment routing header indicating the source virtual machine, an END.S for the first physical machine, the first virtual machine; and an END.SBUF for a second physical machine. The message is then routed based at least in part on the inserted segment routing header.

Abstract: Methods, network devices and computer readable media are disclosed relating to a communication network. In one embodiment, a method includes receiving a message comprising a message bit array. The message bit array identifies egress nodes for the message within a bit indexed explicit replication (BIER) domain of a communications network. A first bit position in the message bit array represents a first one or more of the egress nodes, and a bit value at the first bit position indicates that the message is to be forwarded to one of the first one or more of the egress nodes. The method further includes accessing a first forwarding table entry corresponding to the first bit position, determining whether the first bit position is an anycast bit position having multiple corresponding forwarding table entries, and forwarding the message to one of the first one or more of the egress nodes.

Abstract: Stateless and reliable load balancing using segment routing and an available side-channel may be provided. First, a non-SYN packet associated with a connection may be received. The non-SYN packet may have first data contained in an available side-channel. Next an associated bucket may be retrieved based on a hash of second data in the non-SYN packet. The associated bucket may identify a plurality of servers. Then a one of the plurality of servers may be selected based on the first data contained in the available side-channel.

Abstract: Systems, methods, and computer-readable storage media for multi-destination TCP communications using bit indexed explicit replication (BIER). In some examples, a system can generate a TCP packet associated with a TCP session involving a set of destination devices, and encode an array of bits into the TCP packet to yield a TCP multicast packet. The array of bits can define the destination devices as destinations for the multicast packet. The system can transmit the TCP multicast packet towards the destination devices through a BIER domain. The system can receive acknowledgements from a first subset of the destination devices. Based on the acknowledgements, the system can determine that the first subset of the destination devices received the multicast packet and a second subset of the destination devices did not receive the multicast packet. The system can then retransmit the multicast packet to the second subset of the destination devices.

Abstract: Methods and network devices are disclosed relating to a multi-area communication network employing bit indexed explicit replication (BIER). In one embodiment, a method includes receiving, at a node in a communications network, a message comprising a first message bit array. The method further includes detecting a bit value of a first relative bit position in the first message bit array, where the first relative bit position represents a first group of destination nodes in the communications network, and evaluating the bit value of the first relative bit position in order to determine whether to encapsulate the received message for delivery to the first group of destination nodes. An embodiment of a network device comprises a processor adapted to implement an embodiment of the method.

Abstract: Techniques for zero-loss workload mobility with segment routing for virtual machines are presented. The techniques include receiving, by a virtual router, an electronic message destined for a first virtual machine running on a first physical machine and checking a first virtual machine state for the first virtual machine. In response to determining that it is associated with a running state indicating the first physical machine, inserting a segment routing header including an indication of the source virtual machine, the first physical machine, and the first virtual machine. In response to determining that it is associated with a migration state, inserting, by the virtual router, a segment routing header indicating the source virtual machine, an END.S for the first physical machine, the first virtual machine; and an END.SBUF for a second physical machine. The message is then routed based at least in part on the inserted segment routing header.

Abstract: Aspects of the subject technology provide state-less load-balancing using sequence numbers to identify traffic flows. In some implementations, a process of the technology can include steps for receiving, by a load-balancer, a first packet from a source device including a request to access the service provided by a server coupled to the load-balancer, determining a load for each of the servers, wherein each server is associated with a unique set of sequence numbers, and forwarding the request to a target server selected based on its corresponding load, and wherein the request is configured to cause the target server to issue a reply to the source device. Systems and machine-readable media are also provided.