US 2011/0216651 A1 · Selectively Disabling Reliability Mechanisms on a Network Connection

Abstract

Computerized methods, systems, and computer-storage media for establishing and managing a transmission control protocol (TCP)-based tunnel (“tunnel”) are provided. The tunnel spans between a data center and a private enterprise network and connects endpoints, of a service application, that reside in each location. During communication, the endpoints transmit data packets over one or more channels (e.g., higher-level channel and lower-level channel) that comprise the tunnel. Each of the channels supports reliability mechanisms (e.g., congestion-control mechanism and loss-recovery mechanism) integrally running thereon, for ensuring complete data-packet delivery. To prevent unwarranted performance degradation caused by duplicative efforts of reliability mechanisms, a fabric controller is employed to selectively disable one or more of the reliability mechanisms individually. The selective disablement of the reliability mechanisms is invoked by predefined criteria, such as instructions from a service model or detected identity of a source/destination endpoint, and is invoked on a per network connection basis.

Claims20

Dependent claims

One or more computer-readable media having computer-executable instructions embodied thereon that, when executed, perform a method for communicating across a network connection established between a plurality of endpoints residing in disparate networks, the method comprising:

providing the network connection that spans between a source endpoint and a destination endpoint, wherein the network connection operates as a transport control protocol (TCP)-based tunnel that bridges the disparate networks at which the source endpoint and the destination endpoint reside, respectively;
selectively disabling one or more lower-level reliability mechanisms that are integrally running on the TCP-based tunnel;
and
communicating messages between the first endpoint and the second endpoint through the tunnel without the one or more lower-level reliability mechanisms interfering with a timing at which the messages are sent.

1 more element

A computer system for managing a flow of data between endpoints that reside in individual networks, the computer system comprising:

a data center within a cloud computing platform that hosts a source endpoint, wherein the source endpoint is allocated to an application running on both the cloud computing platform and a private enterprise network;
a resource within the private enterprise network that hosts a destination endpoint that is allocated to the application, wherein the source endpoint and the destination endpoint are connected by a tunnel that transports the data flow directly therebetween, wherein the tunnel enables a higher-level connection running on top of a lower-level connection, and wherein a congestion-control mechanism and a packet-loss mechanism are both built into each of the higher-level connection and the lower-level connection, respectively;
and
a fabric controller running within the data center that establishes the tunnel and that configures the connections within the tunnel, wherein configuring the connections includes selectively disabling the congestion-control mechanism and the packet-loss mechanism built into the lower-level connection and selectively enabling the congestion-control mechanism and the packet-loss mechanism built into the higher-level connection.

1 more element

A computerized method for facilitating communication between a source endpoint and a destination endpoint across a transport control protocol (TCP)-based tunnel, the method comprising:

employing a fabric controller to establish the TCP-based tunnel that communicatively links the source endpoint and the destination endpoint across disparate networks, wherein operation of the source endpoint is supported by a data center and operation of the destination endpoint is supported by a remotely located, resource residing on a private enterprise network managed by a client of the data center;
receiving first internet protocol (IP) packets at the source endpoint that are passed from a virtual machine instantiated within a data center, wherein the first IP packets are carried via a higher-level connection with a first set of reliability mechanisms provisioned thereon;
encapsulating the first IP packets into second IP packets at the source endpoint;
transmitting the second IP packet over the TCP-based tunnel via a lower-level connection that is layered underneath the higher-level connection, wherein the higher-level channel is provisioned with a second set of reliability mechanisms, and wherein the first and second set of reliability mechanisms each include a congestion-control mechanism and a loss-recovery mechanism, respectively;
employing the fabric controller to selectively disable the congestion-control mechanism and the loss-recovery mechanism provisioned on the lower-level connection;
employing the fabric controller to passively allow the congestion-control mechanism and the loss-recovery mechanism provisioned on the higher-level connection to remain enabled;
and
storing a disabled condition of the lower-level connection and an enabled condition of the higher-level connection, wherein the disabled condition represents the disablement of the congestion-control mechanism and the loss-recovery mechanism provisioned on the lower-level connection, and wherein the enabled condition represents the enablement of the congestion-control mechanism and the loss-recovery mechanism provisioned on the higher-level connection.

5 more elements

Dependencies are inferred from the claim text.

Selectively Disabling Reliability Mechanisms on a Network Connection

Life of this patent

Selectively Disabling Reliability Mechanisms on a Network Connection

Life of this patent

Abstract

Claims20