Vmware, Inc.

Rajagopal Sreenivasan

Sudarshana Kandachar Sridhara Rao

Raghav Kempanna

Vipin Padmam Ramesh

Kumara Parameshwaran

Sreeram Iyer

Srinivasa Srikanth Podila

Tathagat Priyadarshi

Varun Lakkur Ambaji Rao

Prasad Rao

Hiral Mehta

Kumara Parameshwaran Rathnavel

Rakesh Kumar R

Aravindhan K

Samar Yadav

Some embodiments provide a novel method for dynamically processing data message flows using different non-uniform memory access (NUMA) nodes of a processing system. Each NUMA node includes a memory and processors that can access data other memories of other NUMA nodes. A load balancing application associated with a first NUMA node receives flows destined for an endpoint application. The flows are assigned to the first NUMA node to be forwarded to the endpoint application. The load balancing application monitors a central processing (CPU) usage of the first NUMA node to determine whether the CPU usage of the first NUMA node exceeds a particular threshold. When the CPU usage of the first NUMA node exceeds the particular threshold, the load balancing application reassigns at least a subset of the flows to the second NUMA node for processing.

Dynamically processing data message flows using different NUMA nodes of a processing system

Some embodiments of the invention provide a method of deploying a tenant deployable element to one public cloud. The method identifies first and second candidate resource elements respectively of first and second resource element sub-types to deploy in a public cloud to implement the tenant deployable element. The method identifies, for the first and second candidate resource elements respectively first and second sets of performance metric values to evaluate. The method evaluates the identified first and second sets of metrics to select one candidate resource element to implement the tenant deployable element in the public cloud. The method uses the selected resource element to implement the tenant deployable element in the public cloud.

Method and apparatus for deploying tenant deployable elements across public clouds based on harvested performance metrics of sub-types of resource elements in the public clouds

Some embodiments of the invention provide a method of deploying first and second tenant deployable elements to a set of one or more public clouds, the first and second tenant deployable elements being different types of elements. The method identifies first and second sets of performance metrics respectively for first and second sets of candidate resource elements to use to deploy the first and second tenant deployable elements, the two sets of performance metrics being different sets of metrics because the first and second tenant deployable elements being different types of elements, the first set of performance metrics having at least one metric that is not included in the second set of performance metrics. The method uses the different sets of metrics evaluate the first and second sets of candidate resource elements, in order to select one of the first set of candidate resource elements for the first tenant deployable element and to select one of the second set of candidate resource elements for the second tenant deployable element. The method deploys the first and second tenant deployable elements in the set of PCDs by using the selected candidate resource elements.

Method and apparatus for deploying tenant deployable elements across public clouds based on harvested performance metrics of types of resource elements in the public clouds

An example method of file transfer between a client and a server includes: initiating, by the client, a front-end control connection between the client and a horizontally scaled proxy service; creating, by a first proxy instance of a plurality of proxy instances of the horizontally scaled proxy service, a back-end control connection between the first proxy instance and the server; returning, to the client from the first proxy instance, a unique client parameter associated with the front-end connection as a destination port for a front-end data connection; initiating, by the client, the front-end data connection between the client and the horizontally scaled proxy service, the front-end data connection using the destination port as returned by the first proxy instance; and creating, by the first proxy instance, a back-end data connection between the first proxy instance and the server.

Port management in a horizontally scaled file transfer system

Example methods and systems for application security enforcement are described. In one example, a computer system may detect, from a client device, a packet requiring processing by a first server pool; and determine whether the packet is associated with a security attack. In response to determination that the packet is not associated with the security attack, the packet may be steered towards the first server pool to cause processing of the packet by one of multiple first application servers. Otherwise, the packet may be steered towards a second server pool to cause processing of the packet by one of multiple second application servers and to learn attack information associated with the security attack. The multiple second application servers in the second server pool may be capable of mimicking behavior of the multiple first application servers in the first server pool.

Application security enforcement

Some embodiments provide a method that, at a first domain name system (DNS) cluster of a set of DNS clusters, receives a DNS request from a client. The first DNS cluster identifies, based on an identifier of the client in the DNS request, a home DNS cluster of the client. The method forwards the DNS request to the home DNS cluster. The home DNS cluster supplies a DNS response to the client. Identifying the home DNS cluster, in some embodiments, includes performing a hash on the identifier of the client. Supplying the DNS response, in some embodiments, includes receiving a virtual IP (VIP) address associated with one of a plurality of sets of application servers to the client and providing the received VIP address to the client in the DNS response.

Method and apparatus for enhanced client persistence in multi-site GSLB deployments

Some embodiments of the invention provide a method of deploying tenant deployable elements on resource elements in public cloud datacenters. The method receives a request to deploy a particular tenant deployable element in one of a first public cloud datacenter (PCD) and a second PCD, the first public cloud datacenter belonging to a first cloud service provider (CSP) and the second public cloud datacenter belonging to a second CSP. For each candidate resource element in the first PCD, the method identifies a first set of performance metrics associated with the candidate resource element. For each candidate resource element in the second PCD, the method identifies a second set of performance metrics associated with the candidate resource element. The method evaluates the identified first and second sets of metrics to select a resource element to implement the particular tenant deployable element in one of the first and second PCDs. The method uses the selected resource element to implement the particular tenant deployable element in one of the first and second PCDs.

Method and apparatus to evaluate resource elements and public clouds for deploying tenant deployable elements based on harvested performance metrics

Example methods and systems for capability-aware service request distribution are described. In one example, a network element may obtain a set of capability information associated with multiple load balancers capable of service request handling and identify varying capability levels among the multiple load balancers. The network element may assign unequal path cost information to the multiple load balancers to indicate the varying capability levels. The unequal path cost information may include (a) a first path cost that is assigned to the first load balancer; and (b) a second path cost that is assigned to the second load balancer and different to the first path cost to indicate the lower capability level of the second load balancer compared to the first load balancer. In response to detecting a service request, the service request may be steered towards one of the multiple load balancers based on the unequal path cost information.

Capability-aware service request distribution to load balancers

Some embodiments provide a method for generating a test traffic configuration for testing a first network. From a second network, the method receives a set of data streams representing data traffic observed in the second network. The method uses a machine learning engine to analyze the set of data streams in order to determine traffic patterns in the second network. The method generates the test traffic configuration for testing the first network by replicating the traffic patterns of the second network in the first network.

Generation of test traffic configuration based on real-world traffic

Described herein are systems, methods, and software to dynamically set a keepalive status for a client-server connection. In one implementation, a load balancer, providing load balancing operations for a plurality of servers, identifies a first FIN packet from a client to a server of the plurality of servers and sets a Transmission Control Protocol (TCP) keepalive status in association the connection between the client and the load balancer. The proxy further identifies a second FIN packet from the server to the client and removes the TCP keepalive status in association with the connection in response to the second FIN packet.

Growing community of inventors

Bengaluru, India

Rajagopal Sreenivasan