Abstract: A system and method for performing enhanced modeling of multi-tiered architectures is presented. The system and method enable selection of a preferred design for a multi-tiered architecture of components based on a set of established criteria, and may employ certain vectors and functions in component attributes, and such attributes may include scalability and scope of fault attributes.

Abstract: A system and method for selecting a preferred design for a multi-tiered architecture of components based on a set of established criteria is provided. The system and method receive a model describing different design constructions and a set of performance and availability requirements and produces a design or set of designs that best meet these requirements. The system and method include a model for representing the infrastructure design space for multi-tier systems and their properties appropriate for judging performance and availability properties of services as well as associated cost. The method and system further include a method for searching the design space in an efficient manner.

Abstract: A method of detecting congestion in a communications network and a network switch are described. The method comprises identifying an output link of a network switch as a congested link on the basis of a packet in a queue of the network switch which is destined for the output link, where the output link has a predetermined state; and identifying a packet in a queue of the network switch as a packet generating congestion if the packet is destined for a congested link.

Abstract: According to the present invention, there is provided a method for controlling congestion in a communications network, the method including receiving congestion feedback data relating to said network, adjusting at least one of a window limit and a rate limit based on said congestion feedback data and injecting data packets onto said network according to said window limit and said rate limit.

Abstract: An embodiment of a method of restoring a communication state of a process includes creating a new socket for a socket saved as part of a checkpoint of the communication state. The new socket is initialized with an adjusted transmission control protocol state saved as part of the checkpoint. The adjusted transmission control protocol state indicates that a send buffer and a receive buffer are empty. Send data saved as part of the checkpoint is written into the new socket. Receive data saved as part of the checkpoint is written into a restart buffer. While at least a portion of the receive data remains in the restart buffer, a socket read system call for the new socket is redirected to read the receive data that remains in the restart buffer.

Abstract: An embodiment of a method of migrating the process domain includes attaching a process-domain interface that includes an internet protocol address to the process domain. The process-domain interface along with the process domain is moved from a first host to a second host.

Abstract: An embodiment of a method of checkpointing parallel processes in execution within a plurality of process domains begins with a step of setting communication rules to stop communication between the process domains. Each process domain comprises an execution environment at a user level for at least one of the parallel processes. The method continues with a step of checkpointing each process domain and any in-transit messages. The method concludes with a step of resetting the communication rules to allow the communication between the process domains.

Abstract: A system and method for performing enhanced modeling of multi-tiered architectures is presented. The system and method enable selection of a preferred design for a multi-tiered architecture of components based on a set of established criteria, and may employ certain vectors and functions in component attributes, and such attributes may include scalability and scope of fault attributes.

Abstract: A system and method for selecting a preferred design for a multi-tiered architecture of components based on a set of established criteria is provided. The system and method receive a model describing different design constructions and a set of performance and availability requirements and produces a design or set of designs that best meet these requirements. The system and method include a model for representing the infrastructure design space for multi-tier systems and their properties appropriate for judging performance and availability properties of services as well as associated cost. The method and system further include a method for searching the design space in an efficient manner.

Abstract: A method of restoring processes within a process domain begins with a step of restoring a tree of processes in which at least two of the processes share at least a resource. The method continues with a step of restoring a checkpoint state of each resource used by the processes after a time when a possible need for a restoration state of the resource exists. According to an embodiment, the restoration state comprises information used by the method during the step of restoring the tree of processes. According to another embodiment, the restoration state comprises information used by the method during the step of restoring the checkpoint state of one or more particular resources. The method concludes with a step of resuming execution of each process after restoration of the checkpoint state of the resources used by the process.

Abstract: An embodiment of a method of checkpointing a virtual memory for a process begins with a step of saving memory resident pages. The method concludes with a step of saving disk swap pages after the step of saving the memory resident pages.

Abstract: A method of checkpointing and restarting processes that share a file that is open begins with a step of assigning a priority to one of the processes that share the file. This identifies a priority process. The method concludes with a step of reopening the file when restoring the priority process.

Abstract: An embodiment of a method of performing a kernel level task upon initial execution of a child process at a user level begins with setting an instruction pointer for an initial child process instruction to an instruction to enter a kernel level. The method continues with beginning the child process which places a return value in a register for the child process and which causes the child process to enter the kernel level. The method concludes with executing a system call having a system call number of the return value. The system call comprises the kernel level task.

Abstract: An embodiment of a method of virtualizing a process identifier begins with creation of a child process from a parent process. The parent process creates the child process by making a system call. If the parent process executes before the child process begins to execute, at a kernel level and without interruption the parent process maps a free virtual process identifier to a kernel process identifier for the child process. If the child process executes before the parent process begins to execute, at the kernel level and without interruption the child process determines the kernel process identifier. The child process then maps the free virtual process identifier to the kernel process identifier.

Abstract: According to the present invention, there is provided a method for controlling congestion in a communications network, the method including receiving congestion feedback data relating to said network, adjusting at least one of a window limit and a rate limit based on said congestion feedback data and injecting data packets onto said network according to said window limit and said rate limit.

Abstract: A method and system providing dynamic allocation of memory through hardware is disclosed. An embodiment provides for a multi-processor system providing for a secure partitioned memory. The system comprises a processor(s), a hardware implemented memory router coupled to the processor(s), and memory coupled to the memory router. The memory router stores memory partition information, which describes the memory allocated to the processor(s). Furthermore, the memory router maps a memory access request from a processor to an address in the memory.

Abstract: A network controls congestion by monitoring how well packets are actually being received at their respective dataflow destinations. The destination nodes are outfitted with a monitor that returns an acknowledgement (ACK) datapacket to the source node for each reception. The return ACK datapackets are marked according to whether congestion was encountered in the delivery to the destination. If so, a rate limiter at the source node is signaled to slow down the data injection rate. If not, the rate limiter is signaled to dial up the injection rate. Several dataflows can be independently and simultaneously controlled this way.

Abstract: A method and apparatus for detecting network congestion including, in some embodiments, detecting congested links and packets associated with a congested link.

Abstract: A network system and method is disclosed that may be useful for addressing congestion issues in network systems. A network system in accordance with the teachings of the invention may provide an acknowledgment packet that may contain information useful to determine, in part, network congestion.

Abstract: A method and system providing dynamic allocation of memory through hardware is disclosed. An embodiment provides for a multi-processor system providing for a secure partitioned memory. The system comprises a processor(s), a hardware implemented memory router coupled to the processor(s), and memory coupled to the memory router. The memory router stores memory partition information, which describes the memory allocated to the processor(s). Furthermore, the memory router maps a memory access request from a processor to an address in the memory.