|A Performance Study to Guide RDMA Programming Decisions||Patrick MacArthur and Dr. Robert Russell||
This paper describes a performance study of Remote Direct Memory Access (RDMA) programming techniques. Its goal is to use these results as a guide for making "best practice" RDMA programming decisions. Infiniband RDMA is widely used in scientific high performance computing (HPC) clusters as a low-latency, high-bandwidth, reliable interconnect accessed via MPI. Recently it is gaining adherents outside scientific HPC as high-speed clusters appear in other application areas for which MPI is not suitable. RDMA enables user applications to move data
|January 18, 2013||White Papers|
The OPEN (One-Pair Ether-Net) Alliance is designed to encourage wide scale adoption of Ethernet-based, single pair unshielded networks as the standard in automotive applications. The jointly developed OPEN Alliance SIG (Special Interest Group) will address industry requirements for improving in-vehicle safety, comfort, and infotainment, while significantly reducing network complexity and cabling costs.
|January 2, 2013||External Links|
|UNH-IOL, OPEN Alliance target 100 Mbps Ethernet for cars||
The University of New Hampshire InterOperability Laboratory (UNH-IOL) in partnership with the OPEN Alliance Special Interest Group (SIG) on Monday unveiled the Automotive Ethernet Consortium.
This consortium is charged with testing and promoting BroadR-Reach, a 100 Mbps Ethernet connectivity specification for the automotive industry. UNH-IOL will be the first lab to conduct tests on the emerging standard.
|January 2, 2013||External Links|
|10 Gigabit Ethernet Clause 50 Tutorial||Andy Baldman||
|December 13, 2012||Tutorials|
|A General-Purpose API for iWARP and InfiniBand||Dr. Robert Russell||
Remote Direct Memory Access (RDMA) allows data to be transferred over a network directly from the memory of one computer to the memory of another computer without CPU intervention. There are two major types of RDMA hardware on the market today: InfiniBand, and RDMA over IP, also known as iWARP. This hardware is supported by open software that was developed by the OpenFabrics Alliance (OFA) and that is known as
The UNH-EXS interface is a multi-threaded implementation of the ES-API plus additional extensions, which enables programmers to utilize RDMA hardware via the OFED stack in a convenient, relatively familiar manner. The UNH-EXS interface is implemented entirely in user space on the Linux operating system. This provides easy porting, modification and adoption of UNH-EXS, since it requires no changes to existing Linux kernels. We present results on the performance of some benchmark applications using the UNH-EXS interface on both iWARP and InfiniBand hardware.
|December 13, 2012||White Papers|
|Fairness in a data center||Mikkel Hagen||
Existing data centers utilize several networking technologies in order to handle the performance requirements of different workloads. Maintaining diverse networking technologies increases complexity and is not cost effective. This results in the current trend to converge all traffic into a single networking fabric. Ethernet is both cost-effective and ubiquitous, and as such it has been chosen as the technology of choice for the converged fabric. However, traditional Ethernet does not satisfy the needs of all traffic workloads, for the most part, due to its lossy nature and, therefore, has to be enhanced to allow for full convergence. The resulting technology, Data Center Bridging (DCB), is a new set of standards defined by the IEEE to make Ethernet lossless even in the presence of congestion. As with any new networking technology, it is critical to analyze how the different protocols within DCB interact with each other as well as how each protocol interacts with existing technologies in other layers of the protocol stack.
This dissertation presents two novel schemes that address critical issues in DCB networks: fairness with respect to packet lengths and fairness with respect to flow control and bandwidth utilization. The Deficit Round Robin with Adaptive Weight Control (DRR-AWC) algorithm actively monitors the incoming streams and adjusts the scheduling weights of the outbound port. The algorithm was implemented on a real DCB switch and shown to increase fairness for traffic consisting of mixed-length packets. Targeted Priority-based Flow Control (TPFC) provides a hop-by-hop flow control mechanism that restricts the flow of aggressor streams while allowing victim streams to continue unimpeded. Two variants of the targeting mechanism within TPFC are presented and their performance evaluated through simulation.
|December 7, 2012||Theses|
|The Extended Sockets Interface for Accessing RDMA Hardware||Dr. Robert Russell||
The Extended Sockets API (ES-API) is a specification published by the OpenGroup that defines extensions to the
|December 7, 2012||White Papers|
|Space-Time Diagrams (PPT)||Jeff Laird||
used at UNH-IOL for training and discussion
|December 6, 2012||White Papers|
|Advantages and Testing Considerations of a Converged DCB Network Presentation||Peter J. Scruton||
Peter participated in the 2012 Ethernet Technology Summit in San Jose, California as a panelist in one of the Sessions. The Session focus was on the Continuing Role of Ethernet in Storage.
|December 4, 2012||External Links|
|Comparison of DCBX Versions||Daniel Shea||
A comparison of the baseline versus the IEEE standard versions of Data Center Bridging Capabilities Exchange.
|September 11, 2012||Tutorials|