Abstract: Methods and apparatus for improved data movement operations through interconnect fabric. In one embodiment, Non-Transparent Bridge (NTB) technology used to perform data movement operations between a host and multiple peer devices using a DMA (direct memory access) engine and at least one descriptor ring having enhanced descriptor entries. In one implementation, descriptor ring entries include source and destination address information, address translation information, and fabric partition information. In one implementation, a DMA engine is configured directly access host memory and generate data packets using the descriptor entry information. In one embodiment, the descriptor ring is a virtual descriptor ring located on DMA hardware, host memory, or elsewhere in the NT fabric address space, and may be accessed by user processes.

Abstract: Methods and apparatus for improved send/receive operations in network interface fabrics. In one exemplary embodiment, mechanisms and protocols for enhanced inter-process (and inter-endpoint) communication, including within very large scale topologies involving e.g., hundreds or even thousands of nodes or endpoints, such as a large-scale high-performance compute or network fabric, are described. In one implementation, the methods and apparatus avoid frequent kernel transitions (and the performance penalties associated therewith) associated with prior approaches through use of UMCs (user message contexts) are created, which contain TX and RX queues that can be read and written directly from user space. A KMC (kernel message context) is also used, in which TX queues are written from the kernel such that access can be arbitrated between unrelated processes. These functions allow for, among other things, significant portions of the foregoing kernel accesses to be obviated.

Abstract: Methods and apparatus for efficient scaling of fabric architectures such as those based on PCIe technology, including up to very large fabrics and numbers of hosts/devices for use in ultra-high performance applications such as for example data centers and computing clusters. In one aspect, methods and apparatus for using Non-Transparent Bridge (NTB) technology to export Message Signaled Interrupts (MSIs) to external hosts are described. In a further aspect, an IO Virtual Address (IOVA) space is created is used as a method of sharing an address space between hosts, including across the foregoing NTB(s). Additionally, a Fabric Manager (FM) entity is disclosed and utilized for programming e.g., PCIe switch hardware to effect a desired host/fabric configuration.

Abstract: Methods and apparatus for improved send/receive operations in network interface fabrics. In one exemplary embodiment, mechanisms and protocols for enhanced inter-process (and inter-endpoint) communication, including within very large scale topologies involving e.g., hundreds or even thousands of nodes or endpoints, such as a large-scale high-performance compute or network fabric, are described. In one implementation, the methods and apparatus avoid frequent kernel transitions (and the performance penalties associated therewith) associated with prior approaches through use of UMCs (user message contexts) are created, which contain TX and RX queues that can be read and written directly from user space. A KMC (kernel message context) is also used, in which TX queues are written from the kernel such that access can be arbitrated between unrelated processes. These functions allow for, among other things, significant portions of the foregoing kernel accesses to be obviated.

Abstract: A network switch, based on the PCI Express protocol, is disclosed. The switch is in communication with a processor, local memory and includes a plurality of non-transparent bridges and, optionally transparent bridges, leading to PCI Express endpoints. By configuring the non-transparent bridges appropriately, the network switch can facilitate simultaneous communication between any two sets of servers without needing to store any data in the local memory or FIFO resources of the switch. For example, the network switch may configure the non-transparent bridges so as to have access to the physical memory of every server attached to it. It can then move data from the memory of any server to the memory of any other server.

Abstract: A network switch, based on the PCI Express protocol, is disclosed. The switch includes a processor, local memory and a plurality of non-transparent bridges. By configuring the non-transparent bridges appropriately, the network switch can facilitate a number of different communication mechanisms, including TCP/IP communication between servers, server clusters, and virtualized I/O device utilization. For example, the network switch may configure the non-transparent bridges so as to have access to the physical memory of every server attached to it. It can then move data from the memory of any server to the memory of any other server. In another embodiment, the network switch is connected to an I/O device, and multiple servers are given access to that I/O device via virtualized connections.