Abstract: The technique herein optimizes delivery performance from a content delivery network (CDN) edge node, preferably by using knowledge of radio link allocation behavior. According to the technique herein, an IP address of a cellular end user (or equivalent) is recognized by the CDN edge. Then, preferably by tracking idle times both in consecutive downloading bursts in the end user session and in an on/off style radio link allocation pattern, the CDN edge adjusts transport layer network protocol parameters. The approach (which leverages the cross-layer intelligence obtainable from existing CDN infrastructure) thus enables the transport layer protocol implemented at the edge to optimize the delivery performance over the cellular mobile network.

Abstract: Among other things, this document describes systems, methods and devices for performance testing and dynamic placement of computing tasks in a distributed computing environment. In embodiments, a given client request is forwarded up a hierarchy of nodes, or across tiers in the hierarchy. A particular computing node in the system self-determines to perform a computing task to generate (or help generate) particular content for a response to the client. The computing node injects its identifier into the response indicating that it performed those tasks; the identifier is transmitted to the client with particular content. The client runs code that assesses the performance of the system from the client's perspective, e.g., in servicing the request, and beacons this performance data, along with the aforementioned identifier, to a system intelligence component. The performance information may be used to dynamically place and improve the placement of the computing task(s).

Abstract: Among other things, this document describes systems, methods and devices for performance testing and dynamic placement of computing tasks in a distributed computing environment. In embodiments, a given client request is forwarded up a hierarchy of nodes, or across tiers in the hierarchy. A particular computing node in the system self-determines to perform a computing task to generate (or help generate) particular content for a response to the client. The computing node injects its identifier into the response indicating that it performed those tasks; the identifier is transmitted to the client with particular content. The client runs code that assesses the performance of the system from the client's perspective, e.g., in servicing the request, and beacons this performance data, along with the aforementioned identifier, to a system intelligence component. The performance information may be used to dynamically place and improve the placement of the computing task(s).

Abstract: This document describes systems, methods and apparatus for locating an object and/or processed versions of that object in a CDN cache system. When a CDN server needs to send a forward request to an origin server to retrieve an object, the CDN server can append a ‘cache hint’ (sometimes referred to herein as a pointer or as ‘reverse cookie’) to its request. The cache hint preferably includes information that will be stored at the origin server and provided to other CDN servers that subsequently ask for the same object. Preferably the information is a pointer that will enable the object to be located within the CDN and/or enable the location of modified version of the object that have already been created and stored within the CDN.

Abstract: Among other things, this document describes systems, methods and devices for performance testing and dynamic placement of computing tasks in a distributed computing environment. In embodiments, a given client request is forwarded up a hierarchy of nodes, or across tiers in the hierarchy. A particular computing node in the system self-determines to perform a computing task to generate (or help generate) particular content for a response to the client. The computing node injects its identifier into the response indicating that it performed those tasks; the identifier is transmitted to the client with particular content. The client runs code that assesses the performance of the system from the client's perspective, e.g., in servicing the request, and beacons this performance data, along with the aforementioned identifier, to a system intelligence component. The performance information may be used to dynamically place and improve the placement of the computing task(s).

Abstract: Among other things, this document describes systems, methods and devices for performance testing and dynamic placement of computing tasks in a distributed computing environment. In embodiments, a given client request is forwarded up a hierarchy of nodes, or across tiers in the hierarchy. A particular computing node in the system self-determines to perform a computing task to generate (or help generate) particular content for a response to the client. The computing node injects its identifier into the response indicating that it performed those tasks; the identifier is transmitted to the client with particular content. The client runs code that assesses the performance of the system from the client's perspective, e.g., in servicing the request, and beacons this performance data, along with the aforementioned identifier, to a system intelligence component. The performance information may be used to dynamically place and improve the placement of the computing task(s).

Abstract: This disclosure provides for an enhancement to a transport layer switch and, in particular the management of end points. In this approach, a memory space, such as a large logical ring buffer, is shared by incumbent connections to facilitate a space multiplexing end point management scheme. Preferably, memory allocation in the memory space is done packet-by-packet dynamically. Because the memory space is shared by all admitted connections, packets belonging to the same connection are not necessarily located physically consecutive to each other. A packet indexing mechanism that implements a set of pointers ensures that consecutiveness for packets on the same connection is maintained. This approach to end point multiplexing provides significant benefits by improving resource utilization, and enabling a higher number of connections to be served.

Abstract: Among other things, this document describes systems, methods and devices for performance testing and dynamic placement of computing tasks in a distributed computing environment. In embodiments, a given client request is forwarded up a hierarchy of nodes, or across tiers in the hierarchy. A particular computing node in the system self-determines to perform a computing task to generate (or help generate) particular content for a response to the client. The computing node injects its identifier into the response indicating that it performed those tasks; the identifier is transmitted to the client with particular content. The client runs code that assesses the performance of the system from the client's perspective, e.g., in servicing the request, and beacons this performance data, along with the aforementioned identifier, to a system intelligence component. The performance information may be used to dynamically place and improve the placement of the computing task(s).

Abstract: This document describes systems, methods and apparatus for locating an object and/or processed versions of that object in a CDN cache system. When a CDN server needs to send a forward request to an origin server to retrieve an object, the CDN server can append a ‘cache hint’ (sometimes referred to herein as a pointer or as ‘reverse cookie’) to its request. The cache hint preferably includes information that will be stored at the origin server and provided to other CDN servers that subsequently ask for the same object. Preferably the information is a pointer that will enable the object to be located within the CDN and/or enable the location of modified version of the object that have already been created and stored within the CDN.

Abstract: Among other things, this document describes systems, methods and devices for performance testing and dynamic placement of computing tasks in a distributed computing environment. In embodiments, a given client request is forwarded up a hierarchy of nodes, or across tiers in the hierarchy. A particular computing node in the system self-determines to perform a computing task to generate (or help generate) particular content for a response to the client. The computing node injects its identifier into the response indicating that it performed those tasks; the identifier is transmitted to the client with particular content. The client runs code that assesses the performance of the system from the client's perspective, e.g., in servicing the request, and beacons this performance data, along with the aforementioned identifier, to a system intelligence component. The performance information may be used to dynamically place and improve the placement of the computing task(s).

Abstract: This document describes systems, methods and apparatus for locating an object and/or processed versions of that object in a CDN cache system. When a CDN server needs to send a forward request to an origin server to retrieve an object, the CDN server can append a ‘cache hint’ (sometimes referred to herein as a pointer or as ‘reverse cookie’) to its request. The cache hint preferably includes information that will be stored at the origin server and provided to other CDN servers that subsequently ask for the same object. Preferably the information is a pointer that will enable the object to be located within the CDN and/or enable the location of modified version of the object that have already been created and stored within the CDN.

Abstract: The technique herein optimizes delivery performance from a content delivery network (CDN) edge node, preferably by using knowledge of radio link allocation behavior. According to the technique herein, an IP address of a cellular end user (or equivalent) is recognized by the CDN edge. Then, preferably by tracking idle times both in consecutive downloading bursts in the end user session and in an on/off style radio link allocation pattern, the CDN edge adjusts transport layer network protocol parameters. The approach (which leverages the cross-layer intelligence obtainable from existing CDN infrastructure) thus enables the transport layer protocol implemented at the edge to optimize the delivery performance over the cellular mobile network.

Abstract: This document describes, among other things, improved methods, systems, and apparatus for relaying packets on computer networks. Preferably, the relay function is accelerated at a host by implementing selected forwarding functions in hardware, such as the host's network interface card, while upper software layers at the host retain at least some access to the packet flow to handle more complex operations and/or monitoring. In a so-called “split TCP” arrangement, for example, a relay host terminates a first TCP connection from a given host and forwards packets on that connection to another given host on a second TCP connection. The relay host has a TCP forwarding table implemented at the device level, configurable by a relay management application running in the kernel or user-space. Special forwarding table modes may be used to enable full-TCP protocol support while also taking advantage of hardware acceleration.

Abstract: The technique herein optimizes delivery performance from a content delivery network (CDN) edge node, preferably by using knowledge of radio link allocation behavior. According to the technique herein, an IP address of a cellular end user (or equivalent) is recognized by the CDN edge. Then, preferably by tracking idle times both in consecutive downloading bursts in the end user session and in an on/off style radio link allocation pattern, the CDN edge adjusts transport layer network protocol parameters. The approach (which leverages the cross-layer intelligence obtainable from existing CDN infrastructure) thus enables the transport layer protocol implemented at the edge to optimize the delivery performance over the cellular mobile network.

Abstract: Disclosed are methods, systems, and apparatus for load-balancing image-processing jobs based on the number of pixels in the images and/or the nature of the processing that is requested on those pixels. For example, a set of machines may run software to provide various types of image processing services, such as format conversion, recompression, resizing, cropping, among others. These are referred to as image servers. In accordance with the teachings hereof, the load on each image server can be calculated based on the number of pixels in the images that are waiting to be processed in the image server's processing queue, adjusted by the type of processing that is requested on each image. The adjustment typically reflects and adjusts for the relative time needed to perform various types of processing. Load scores can be further adjusted based on the processing capabilities of each image server, in some embodiments.

Abstract: This disclosure provides for a Layer 4 switching approach wherein a set of L4 switches are organized into a cluster so as to act as a single (or “big”) Layer 4 switch. Connections between the L4 switches are carried out, e.g., using Layer 2 switches. To this end, an intra-cluster routing entity of the switch maintains mapping information (e.g., in a database, or set of data structures) about connections that have been established by the individual switches within the cluster. In this approach, each host (itself a switch) preferably acts like a group of ports of the larger (big) switch. This obviates having each member host from having to maintain connections to many possible destinations. Rather, the intra-cluster routing entity maintains the information about which hosts (and its ports) are connected to which destinations, and the connections are re-used as necessary, even if connections on one side of the “big” switch ceased being used.

Abstract: This disclosure provides for an enhancement to a transport layer switch and, in particular the management of end points. In this approach, a memory space, such as a large logical ring buffer, is shared by incumbent connections to facilitate a space multiplexing end point management scheme. Preferably, memory allocation in the memory space is done packet-by-packet dynamically. Because the memory space is shared by all admitted connections, packets belonging to the same connection are not necessarily located physically consecutive to each other. A packet indexing mechanism that implements a set of pointers ensures that consecutiveness for packets on the same connection is maintained. This approach to end point multiplexing provides significant benefits by improving resource utilization, and enabling a higher number of connections to be served.

Abstract: This disclosure provides for an enhancement to a transport layer switch and, in particular the management of end points. In this approach, a memory space, such as a large logical ring buffer, is shared by incumbent connections to facilitate a space multiplexing end point management scheme. Preferably, memory allocation in the memory space is done packet-by-packet dynamically. Because the memory space is shared by all admitted connections, packets belonging to the same connection are not necessarily located physically consecutive to each other. A packet indexing mechanism that implements a set of pointers ensures that consecutiveness for packets on the same connection is maintained. This approach to end point multiplexing provides significant benefits by improving resource utilization, and enabling a higher number of connections to be served.

Abstract: This disclosure provides for a Layer 4 switching approach wherein a set of L4 switches are organized into a cluster so as to act as a single (or “big”) Layer 4 switch. Connections between the L4 switches are carried out, e.g., using Layer 2 switches. To this end, an intra-cluster routing entity of the switch maintains mapping information (e.g., in a database, or set of data structures) about connections that have been established by the individual switches within the cluster. In this approach, each host (itself a switch) preferably acts like a group of ports of the larger (big) switch. This obviates having each member host from having to maintain connections to many possible destinations. Rather, the intra-cluster routing entity maintains the information about which hosts (and its ports) are connected to which destinations, and the connections are re-used as necessary, even if connections on one side of the “big” switch ceased being used.

Abstract: This document describes, among other things, improved methods, systems, and apparatus for relaying packets on computer networks. Preferably, the relay function is accelerated at a host by implementing selected forwarding functions in hardware, such as the host's network interface card, while upper software layers at the host retain at least some access to the packet flow to handle more complex operations and/or monitoring. In a so-called “split TCP” arrangement, for example, a relay host terminates a first TCP connection from a given host and forwards packets on that connection to another given host on a second TCP connection. The relay host has a TCP forwarding table implemented at the device level, configurable by a relay management application running in the kernel or user-space. Special forwarding table modes may be used to enable full-TCP protocol support while also taking advantage of hardware acceleration.