Intel is betting that future data center operations will depend on increasingly powerful servers running ASIC-based, programmable processors, and is committed to developing infrastructure processors (IPUs), which are Intel’s programmable network devices designed for reduce costs and increase CPU performance for free.

Intel is among a growing number of vendors – including Nvidia, AWS and AMD – working to build smartNICs and DPUs to support software-defined cloud, computing, networking, storage and security services designed for rapid peripheral, colocation or deployment. network services.

Intel’s original IPU combined Xeon CPU and FPGA, but will eventually become a powerful ASIC that can be customized and controlled with an open source software-based Infrastructure Programmer Development Kit (IPDK). IPDK runs on Linux and uses programming tools such as SPDK, DPDK, and P4 for developers to control network virtualization and storage, as well as workload provisioning.

B is introductory Intel Vision This week’s event in Texas, Intel talks about other new chips and how AI will play in the data center. He outlined a roadmap for the development of his IPU and explained in detail why the device portfolio would be an important part of his data center plans.

Specifically for its IPU roadmap, Intel said it will ship two 200Gb IPUs by the end of the year. One, codenamed Mount Evans, was developed with Alphabet Inc.’s Google Cloud Group. and at this point will be focused on high-end data center servers and hyperscalers.

The ASIC-based Mount Evans IPU can support existing uses such as vSwitch unloading, firewalls, and virtual routing. It implements a hardware-accelerated NVM storage interface scaled by Intel Optane technology to emulate NVMe devices.

Copyright © 2022 IDG Communications, Inc.

Previous article5 Google applications without which there is no Android
Next articleThe massive asteroid 2.5 times larger than the Statue of Unity is taking a “close approach” this week