Accelerating key workloads with 4th Gen Intel® Xeon® - based platforms
4/08/2023
Artificial intelligence (AI), networking, and storage workloads have in common a seemingly insatiable thirst for processor performance. But although the hardware price/ performance curve (Moore’s Law) continues to slake that thirst, it often has resulted in growing energy demands that have thwarted sustainability objectives. Now, though, 4th Gen Intel® Xeon® Scalable processors feature built-in accelerators that enable certain workloads to run faster while using less energy.
Those new processors are enabling Advantech, a leading industrial AI platform and networking solution provider, to deliver energy-efficient server and network appliance solutions that will fulfill high expectations for AI, 5G infrastructure, and edge computing.
Key applications thrive on performance.
Operational technology and communications technologies are converging, accelerated by virtualization, and consolidated onto servers for software-defined everything (SDE) environments and intelligent cloud-native infrastructures.
These forces are spurring innovation
in AI, 5G radio area networks (RANs),
data encryption and security, and
high-performance computing (HPC).
Market analyst firm IDC says that it
has identified more than 150 use cases
for edge computing across industries,
including manufacturing operations,
production asset management, smart
grids, omnichannel operations, public
safety and emergency response,
freight monitoring, and intelligent
transportation systems.
“The early goal of edge computing
was to reduce the bandwidth costs
associated with moving raw data
from where it was created to either an
enterprise data center or the cloud,”
reports Network World. “More recently,
the rise of real-time applications
that require minimal latency, such
as autonomous vehicles and multicamera
video analytics, [is] driving
the concept forward. The ongoing
global deployment of the 5G wireless
standard ties into edge computing because 5G enables faster processing for these cutting-edge, low-latency use cases and applications.”
Market research firm Polaris estimates
that the global edge computing market
will grow at a compound annual
growth rate of more than 38% from
$7.43 billion in 2021 through 2030.
“The growing demand for the edge
computing market is expected to be
driven by the rising adoption of smart
devices and industrial IoT, increasing
expenditure for autonomous vehicles,
and growing demand for deployment
of a 5G network,” according to the
Polaris report.
AI applications in manufacturing,
industrial, and enterprise applications
are often driven by use cases where
video sensors are critical for inference
algorithms used in object detection,
recognition, and analysis applications.
but come with a higher computational cost to deliver high-resolution video. 5G edge application innovations depend on optimizing the transmission path of packets between the processor and the 5G modem from the hardware platform level to eliminate performance barriers between the 5G carrier network, edge applications, and the Data Plane Development System (DPDK) ecosystem.
Industrial and carrier-grade servers
and appliances must be able to
maximize performance and ensure
virtually zero downtime at the edge,
often in harsh environments.
The 4th Gen Intel Xeon Scalable
processor family provides multiple
architecture enhancements to
improve performance over the
previous generation, including the
most built-in accelerators of any
CPU on the market to ensure that
OEMs and device manufacturers
can accelerate performance across
the fastest-growing workloads in AI,
data analytics, networking, storage,
and HPC.
- Up to 60 high-performance cores
- Up to 8x DDR5 channels and 80 lanes of PCIe 5.0
- Built-in accelerators for fast-growing edge AI, network edge, network security, storage, and HPC workloads
OEMs and device manufacturers can take advantage of numerous technological advances in the new processor family, including:
“Intel’s new Xeon processors target
edge applications, especially across
data center environments, that can
benefit from enduring availability and
industrial-grade features across a
wide range of power and performance
envelopes,” says Futurum Research.
“This includes a vast array of
repeatable market-ready solutions
that organizations can deploy and
scale swiftly.”
Some of the notable accelerators are:
- Intel Advanced Matrix Extensions (Intel AMX), a built-in accelerator that improves the performance of deep-learning training and inference, which, according to Intel,
is ideal for workloads such as natural-language processing, recommendation systems, and image recognition.
- Intel Data Streaming Accelerator (Intel DSA), which drives high performance for storage, networking, and data-intensive workloads by improving streaming data movement and transformation operations.
- Intel In-Memory Analytics
Accelerator (Intel IAA), which
improves analytics performance
while offloading tasks from CPU
cores to accelerate database query
throughput and other workloads.
- Intel Dynamic Load Balancer
(Intel DLB), for providing efficient
hardware-based load balancing by
dynamically distributing network
data across multiple CPU cores as
the system load varies.
- Intel Advanced Vector Extensions
(Intel AVX) for vRAN, which
increases virtual radio access
network (vRAN) density up to 2x
over the previous generation, with
the same power envelope.
- Intel QuickAssist Technology
(Intel QAT), which accelerates
encryption, decryption, and
data compression, offloading
these tasks from the processor
core to help reduce system
resource consumption.
- Intel Crypto Acceleration, which reduces the penalty of implementing pervasive data encryption and increases the performance of encryption sensitive workloads, such as for Secure Sockets Layer (SSL) web servers, 5G infrastructure, and virtual private networks (VPNs)/ firewalls.
Advantech is leveraging Intel’s latest
processor technology to deliver server
and network appliance products
aimed at enabling an intelligent
planet and accelerating the goals of
industrial intelligence.
4th Gen Intel Xeon processors power
Advantech motherboards to enable
next-generation machine vision,
deep learning, or video analytics
applications for the industrial, medical,
and IoT markets. Combined with
built-in acceleration from Intel Deep
Learning Boost (Intel DL Boost) with
new Intel Advanced Matrix Extensions
(Intel AMX), the new motherboards
provide impressive AI inference
performance compared to previous generation
solutions.
Advantech servers based on the 4th
Gen Intel Xeon Scalable processor
help industrial and communications
equipment vendors and service
providers address increasing AI,
networking, and storage workload
needs, featuring built-in accelerators
to meet performance and sustainability
objectives, including the fastest data
mobility between compute, storage,
and networking layers, thanks to
the hardware-enabled Intel Data
Streaming Accelerator (Intel DSA).
vRAN opens the telecom edge to innovation, enabling new internet of things (IoT) and cloud partnerships and private network services that will enable operators to monetize 5G investment. Advantech is providing an enhanced range of Xeon-based high-performance servers that have been optimized for vRAN and the telecom edge to meet high availability and carrier-grade network needs.
“One of the key markets for the
Advantech SKY server is telecom
carrier-edge deployment servers,”
says Kevin Lin, senior product
manager with Advantech. “With the
built-in forward error correction
(FEC) function in the new processors,
Advantech SKY servers can achieve
twice the performance of an addin
PCIe FEC card. It also reduces
total system power consumption
up to 20%.”
Advantech SKY-8134S-11 and SKY-
820V3 servers provide solution
vendors, system integrators, and
service providers with a powerful tool
to optimize the access network while
meeting 5G (and beyond) traffic, signal
processing, and AI computing needs.
“SKY-8134S-11 also has LAN on
motherboard (LOM), which gives
customers flexibility to add more.
features and functions tailored to their specific needs,” says Lin. The LOM design follows the Intel reference architecture for 25G Intel Ethernet Network Adapter E810 with high precision timing synchronization, which is very important and essential for 5G-NR TDD network success.
For enterprise networking and
security, Advantech network-optimized
appliances utilizing 4th Gen Intel Xeon
Scalable processors improve network
security application throughput,
leveraging built-in Intel Quick Assist
Technology for compression and
encryption acceleration:
- The FWA-5072 is a 1U Rackmount Network Appliance for network functions virtualization (NFV) and SD-WAN that offers up to 32 cores for NFV needs with a single 4th Gen Intel Xeon scalable processor and up to 16 DDR5 RDIMM memory to boost computing capability.
- The FWA-6072 is a 2U Rackmount
Network Appliance with single
4th Gen Intel Xeon Scalable
Processors, support for up to
52 cores, and 16 DDR5 RDIMM,
with eight Network Mezzanine
Card (NMC) slots for network
I/O configuration expansion.
- The FWA-6172 is a 2U Rackmount Network Appliance with dual 4th Gen Intel Xeon Scalable Processors for gateway and next-generation firewall (NGFW) workloads, available with up to 52 cores per socket and up to 16 DDR5 RDIMM memory. It also provides expandable Ethernet capability via the Advantech Network Mezzanine Card (NMC), with as much as 200GbE of bandwidth per slot.
Leveraging its hardware design expertise, Advantech is providing
motherboards, servers, and network appliances based on 4th Gen Intel
Xeon Scalable processors that accelerate industrial, telecom, and
enterprise workloads while meeting challenging density, efficiency, and
environmental deployment needs at the edge. In addition, Advantech
solutions provide high reliability, including advanced platform management
based on in-house BMC development, long life cycles for embedded
industrial and telecom customers, global logistics with local support, and
strong cocreation partnerships for shorter time to market and reduced