IBM Research
Cloud & Computing Infrastructure
One consequential feature of Converged Enhanced Ethernet (CEE) is losslessness, achieved through L2 Priority Flow Control (PFC) and Quantized Congestion Notification (QCN). We focus on QCN and its effectiveness in identifying congestive... more
One consequential feature of Converged Enhanced Ethernet (CEE) is losslessness, achieved through L2 Priority Flow Control (PFC) and Quantized Congestion Notification (QCN). We focus on QCN and its effectiveness in identifying congestive flows in input-buffered CEE switches. QCN assumes an idealized, output-queued switch; however, as future switches scale to higher port counts and link speeds, purely output-queued or sharedmemory architectures lead to excessive memory bandwidth requirements; moreover, PFC typically requires dedicated buffers per input. Our objective is to complement PFC’s coarse per-port/priority granularity with QCN’s per-flow control. By detecting buffer overload early, QCN can drastically reduce PFC’s side effects.We install QCN congestion points (CPs) at input buffers with virtual output queues and demonstrate that arrival-based marking cannot correctly discriminate between culprits and victims. Our main contribution is occupancy sampling (QCN-OS), a novel, QCN-c...
- by Mitchell Gusat and +2
- •
The Internet Engineering Task Force (IETF) is defining a set of protocols, also known as the iWARP stack, for Direct Data Placement and Remote Direct Memory Access (RDMA) over IP-based transports. The iWARP stack has the potential to... more
The Internet Engineering Task Force (IETF) is defining a set of protocols, also known as the iWARP stack, for Direct Data Placement and Remote Direct Memory Access (RDMA) over IP-based transports. The iWARP stack has the potential to solve the current problem of high host processing overhead for user data movement to and from the network. Nevertheless, its industry-wide acceptance is strongly dependent on the establishment of an appropriate ecosystem, i.e., of an RDMA host software architecture and a corresponding full-featured open source reference implementation for Linux. This paper evaluates architectural alternatives and relates them to the design and choice of the programming interfaces currently being defined by the industry. Focusing on the recently started OpenRDMA project for Linux, it identifies one architecture as particularly attractive due to its multi-vendor support with centralized, vendor- independent resource management.
One consequential feature of Converged Enhanced Ethernet (CEE) is losslessness, achieved through L2 Priority Flow Control (PFC) and Quantized Congestion Notification (QCN). We focus on QCN and its effectiveness in identifying congestive... more
One consequential feature of Converged Enhanced Ethernet (CEE) is losslessness, achieved through L2 Priority Flow Control (PFC) and Quantized Congestion Notification (QCN). We focus on QCN and its effectiveness in identifying congestive flows in input-buffered CEE switches. QCN assumes an idealized, output-queued switch; however, as future switches scale to higher port counts and link speeds, purely output-queued or sharedmemory architectures lead to excessive memory bandwidth requirements; moreover, PFC typically requires dedicated buffers per input. Our objective is to complement PFC’s coarse per-port/priority granularity with QCN’s per-flow control. By detecting buffer overload early, QCN can drastically reduce PFC’s side effects.We install QCN congestion points (CPs) at input buffers with virtual output queues and demonstrate that arrival-based marking cannot correctly discriminate between culprits and victims. Our main contribution is occupancy sampling (QCN-OS), a novel, QCN-c...
- by Fredy Neeser
- •
The Internet Engineering Task Force (IETF) is defining a set of protocols, also known as the iWARP stack, for Direct Data Placement and Remote Direct Memory Access (RDMA) over IP-based transports. The iWARP stack has the potential to... more
The Internet Engineering Task Force (IETF) is defining a set of protocols, also known as the iWARP stack, for Direct Data Placement and Remote Direct Memory Access (RDMA) over IP-based transports. The iWARP stack has the potential to solve the current problem of high host processing overhead for user data movement to and from the network. Nevertheless, its industry-wide acceptance is strongly dependent on the establishment of an appropriate ecosystem, i.e., of an RDMA host software architecture and a corresponding full-featured open source reference implementation for Linux. This paper evaluates architectural alternatives and relates them to the design and choice of the programming interfaces currently being defined by the industry. Focusing on the recently started OpenRDMA project for Linux, it identifies one architecture as particularly attractive due to its multi-vendor support with centralized, vendor- independent resource management.