A prosperous 2019 of Open Computing in Inspur
08 January 2020
2019 is a year of progress and prosperity for Inspur in the open computing arena. As a key member of OCP, Open19 and ODCC, the world's three largest open computing standards organizations, Inspur leverages decades of experience in hardware design to promote new open architectures, facilitates the construction of a broader and more complete open ecosystem, and drives efficiency upgrades in AI, cloud computing and edge computing, etc.
Here is an overview of Inspur open computing footprint in 2019.
Launched the first high-density cloud-optimized four-socket platform in March
On March 14th, at the OCP Global Summit, Inspur and Intel jointly launched NF8260M5, the first high-density cloud-optimized four-socket platform for cloud service provider market. The platform is a high-density, flexible and powerful 2U server, validated for the Generation Intel® Xeon™ Scalable Processor (Codename Cascade Lake), and optimized with Intel® Optane™ DC persistent memory.
In addition, Inspur showcased three new artificial intelligence (AI) computing solutions and new joint initiatives with OCP.
Inspur attended OCP Global Summit
Continued participation and support to Open19 Summit in May
In May, as a sponsor of Open19 Summit and lead vendor for Open19, Inspur showcased a 1U computing solution and storage solution, integrated with the OCP NIC card in the Open19 Summit.
Inspur attended Open19 summit
Collectively organized OCP China Day in June
The OCP China Day, co-sponsored by the OCP Foundation and OCP platinum member Inspur, kicked off in Beijing on June 25. The event attracted nearly 1,000 executives and engineers in the data center ecosystem, which integrated AI, Edge Computing, and 5G Topics with Open Computing.
OCP China Day
Four open computing products were recognized by OCP as accepted or inspired in Sept.
Inspur's four open computing products were officially certified by OCP AcceptedTM and OCP InspiredTM. They were 2U 4-socket server NF8260M5, 3OU ON5388M5 with support for 8 GPUs, 2OU storage JOBD ON5266M5 and NF8380M5, a 4-socket Olympus server suitable for multi-scene AI applications.
Actively contributed in the OAI standards establishment
Launched the world’s first OAI compliant computing product in Sept.
In Sept., the world's first OAI(Open Accelerator Infrastructure) compliant and liquid cooling rack-scale AI computing product, X-MAN 4.0, was jointly launched by Baidu and Inspur to support deep neural network applications. Furthermore, At the OCP regional summit, Inspur participated a roundtable discussion with leaders from OCP on the in-depth exploration of the development and value proposition of OAI. Inspur is committed to promote standardization for advancing heterogeneous AI computing.
Inspur Attended OCP regional Summit
Revealed the first 21” reference system supporting the OAI Universal BaseBoard specification
In addition, Inspur revealed the first 21” reference system supporting the OAI Universal BaseBoard specification, simplifying inter-module communication and supporting disparate interconnect topologies.
OCP CTO Bill Carter showcased Inspur OAI UBB in his keynote speech
Inspur OAI UBB
Led the OpenRMC project in the community, Released OpenRMC NB API 1.0 and its roadmap in Sept.
Inspur has led the establishment of the OpenRMC project group and promoted the integration of new technologies and new standards. At the OCP regional summit, Inspur released version 1.0 of OpenRMC and roadmap. The new OpenRMC 1.0 is to be contributed in the OCP Hardware Management Working Group for open discussion. The solution is designed to help users, in small and medium-sized data centers especially, slash cost of IT operation and maintenance, unify management of multi-vendor systems and boost efficiency.
OpenRMC Version 2.0—scheduled for release in 2020—intended to realize data center users’ needs for two-way, comprehensive management of entire rack systems with detailed WebUI
Wilson Guo, Inspur Sr. Technical Director, updated OpenRMC at OCP engineering Workshop
I-Flex became open source to the industry in Sept.
Inspur has also announced that I-Flex server design standards will come to be open source in the OCP community. Based on the T-Flex 1.0, collaboration with Tencent, I-Flex is developed by Inspur to satisfy specific demands of Internet businesses. I-Flex enables modularized deployment of servers based on different application scenarios, so as to achieve server upgrades without changing the server architecture.
Inspur’s unique JDM model is the engine of excellence in the intelligent open computing era
Last but not least, Inspur has made product technology innovation and business chain agility the core of its business strategy. In early 2019, Baidu would allocate digital red envelops worth 900 million yuan ($133.5 million) on its app during the CCTV Spring Festival Gala. With the JDM (Joint Design Manufacturing) model, Inspur used L11 rack level integration to deploy 10,000 nodes in 8 hours to the Baidu Beijing data center to guarantee timely and fluent services. It has set a new record for server delivery in the industry, and has taken Inspur’s rack scale delivery capacity to 30,000 nodes per day. JDM model has become the guarantee for Inspur to continue leading in the field of hyper scale data centers.
Inspur JDM best practice
The year has come to a start again. Riding on the achievements, Inspur will continue to explore and contribute to the open computing industry and keep its position as:
1.An active advocate of open computing technology.
2.Increasingly engaged in the open computing community, from contributing specs to developing and establishing standards. Inspur will effectively coordinate with the upstream and downstream industries to provide industrial support for these standards, and to shorten the time from technical standards to industrial applications.
3.Adhere to its Joint Design Manufacturing methodology which will enable clients to embrace open computing.