New Resource on the Advantages of RoCE for Next Generation Data Centers
A new RoCE Initiative whitepaper, titled RoCE Accelerates Data Center Performance, Cost Efficiency, and Scalability, explains the many benefits of deploying RoCE and how it can address the ever-changing needs of next generation data centers. The document features recent case studies and outlines the following RoCE advantages:
- Freeing the CPU from Performing Storage Access
- Altering Design Optimization
- Future-Proofing the Data Center
- Presenting New Opportunities for the Changing Data Center
Learn more about RoCE’s impact on the evolving data center by downloading the whitepaper from the RoCE Initiative’s Resource Library.
Knowledge of RDMA Is Key to Understanding the Modern Data Center
The data storage industry is dramatically and rapidly transforming with the emergence of solid-state, software-defined, open source and cloud applications. Matching performance to these changes in the industry is becoming increasingly more important. A recent Network Computing article explains how familiarity with Solid-State Drives (SSDs), cloud-based backup, hyperconverged infrastructure, NVMe, RDMA, and more are key to staying ahead of the game in the storage industry.
Check out the full article for more information on what data storage skills will be in demand in the years ahead.
The Impact of HDR InfiniBand on the HPC and Machine Learning Platforms
The recently announced 200Gb/s HDR InfiniBand solutions have ushered in revolutionary In-Network Computing and In-Network Memory capabilities, furthering the new archetype of Data-Centric data centers – for High-Performance Computing, Machine Learning, Cloud, Web2.0, Big Data, Financial Services and more. These capabilities radically increase network scalability and introduce new accelerations for storage platforms and data center security, propelling HPC and machine learning into the next generation.
For a detailed overview on HDR InfiniBand’s impact on HPC and machine learning, read the complete insideHPC article.
Persistent Memory is Making Its Mark on Computing
This past January, the Storage Networking Industry Association (SNIA) hosted a Persistent Memory Summit in San Jose to explore the implications and implementation of Persistent memory (PM), a high performance solid-state memory. Tom Coughlin’s recent article in Forbes highlighted a selection of presentations from the event that addressed the need for next gen storage technologies and how such advancements can support changing computer architectures, improved performance and lower power requirements. Within his report, Tom highlighted a session from OpenFabrics Alliance Vice Chair, Paul Grun, that explained how the combination of RDMA and PM technologies can enable new network storage architectures.
Read the Forbes article for additional analysis of the Persistent Memory Summit, as well as PM and RDMA trends in computing.
InfiniBand Holds Its Ground in Evolving Enterprise and HPC Industry
In a recent The Next Platform article, IBTA Marketing Work Group co-chair, Bill Lee, discusses how the organization’s performance roadmap and steady latency and scalability improvements will continue to give InfiniBand technology an edge over other interconnects within the HPC industry. Bill goes into detail about the IBTA’s work to develop faster InfiniBand data rates such as HDR, NDR and XDR, and how these improvements will support a world where 1,000 Gb/s server connectivity is the norm. Additionally, he highlights other recent improvements to the InfiniBand architecture specification, such as the Virtualization Annex, which will enable the technology to expand into enterprise data centers and cloud deployments.
The Next Platform piece dives into InfiniBand’s roadmap and how the technology will stand its ground in the changing industry landscape.
HPC Makes Its Way to the Cloud
The cloud computing industry has grown exponentially over the past few years and expanded across a variety of business sectors. However, only recently have cloud providers begun to adopt HPC technologies to cater to users with more intensive computing demands. HPC often requires substantially more expensive hardware and cloud providers are seeing an opportunity to service this new group of demanding users that are unable to invest in their own internal HPC infrastructure.
For details on how cloud providers are expanding their HPC offerings, read InsideHPC's article.
Albert Einstein Institute Leverages InfiniBand for Black Hole and Neutron Stars Collision Research
A new research team at the Albert Einstein Institute (AEI) is utilizing InfiniBand networking to help study the collision of black holes and neutron stars. Through sophisticated computer calculations and simulations, this research will explore the fundamental questions in the new field of gravitational-wave astronomy over the next five years. The large-scale simulations will be processed by the AEI’s “Holodeck” supercomputer, which consists of 630 CPU cores connected by high-speed InfiniBand interconnects.
To learn more about the Holodeck and InfiniBand’s role in gravitational-wave research, read Science Daily’s article.
Mellanox Announces: Baidu's Machine Learning Platforms to Leverage RoCE Solutions
Mellanox recently announced that the company’s Spectrum™ Ethernet switches and ConnectX®-4 100Gb/s Ethernet adapters have been selected by Baidu, the leading Chinese language Internet search provider, for their Machine Learning platforms. The RoCE-enabled solutions are key to providing higher data speeds and more efficient data movement for Baidu’s machine learning technology. The resulting partnership will enable Baidu to demonstrate a notable 200 percent improvement in machine learning training times, resulting in faster decision making.
View Mellanox’s official announcement to learn more.
ISC High Performance 2017
June 18-22, 2017
ISC High Performance 2017 will focus on HPC technological development, application in scientific fields, and adoption in commercial environments. The conference will offer 12 HPC topics in three categories: systems, applications, and emerging topics. All topics will be addressed in multiple power-packed sessions
The following IBTA members will be attending ISC High Performance 2017:
- Bull (D-1126)
- HPE (E-1030)
- Intel (F-930)
- Cray (E-291)
- IBM (D-1140)
- Mellanox (E-911)
- Fujitsu (G-820)
For more information, visit the event website.
OpenFabrics Alliance (OFA) Workshop 2017 – Session Abstracts Published
The 13th Annual OpenFabrics Alliance (OFA) Workshop will be taking place on March 27-31, 2017, at the Hyatt Regency in Austin, Texas. Each year, the workshop generates lively exchanges among Alliance members, developers, and users who all share a vested interest in high performance networks. OFA hosts its workshop annually as a joint effort among open source networking community members. The objective is to create an event driven by the needs and interests of its attendees.
To view the official list of session abstracts, visit the OFA Workshop 2017 Abstracts and Agenda page.
If you are unable to attend this year’s workshop, session recordings will be published on both the OFA website and insideHPC.
For more information on the 2017 workshop program as well as a few sessions we recommend checking out, read the recent recent blog.
IBTA Plugfest #31
April 17-28, 2017
Join the IBTA for its 31st InfiniBand Compliance and Interoperability Plugfest, taking place April 17-28 at the University of New Hampshire (UNH) Interoperability Lab (IOL). Attendees will have the opportunity to measure their products for compliance with the InfiniBand Architecture Specification as well as interoperability with other InfiniBand and RoCE products.
Details and registration information are available on IBTA Plugfest page.
For questions or comments, please contact email@example.com