Edge Computing

Gaining the digital edge

BullSequana Edge delivers powerful AI inference and real-time analytics at the edge while keeping data safe and secure

Analysts estimate that by the end of 2020, 20 billion IoT devices will be deployed worldwide

Edge computing is becoming a priority in the Internet of Things. The exponential growth of intelligent sensors and devices is generating an unprecedent amount of data.

It is a new IT revolution, meaning a combination of traditional data centers and edge computing are necessary to provide powerful analytics and machine learning inference capabilities where needed.

BullSequana Edge has been designed to meet these challenges, delivering powerful AI inference and streaming analytics capabilities at the edge while ensuring that all data remains safe and secure.

Analysts expect that by 2022, up to 75% of data will be produced and processed right at the edge of the networks

Download the position paper

9 reasons why BullSequana Edge is your best choice for edge computing

Optimized physical and logical security

The BullSequana Edge is equipped with an Intrusion Sensor which disables the system when opened. There is no physical console port to which a local intruder could connect a screen. A secure boot process can be put in place including signed firmware, bootloader, OS and applications, protected by a FIPS140-2 certified TPM and encrypted disks.

Autonomous

No dependence on cloud and datacenter availability and connectivity, ensures that apps are not disrupted in case of limited or intermittent network connectivity from undergrounds to offshore platforms.

Costeffective

Reduced datacenter infrastructure and networking costs.

Low latency and response times for complex AI video inference models

Video & IoT analysis in real-time thanks to powerful AI inference. The server is equipped of 2 Nvidia Tesla 4 GPUs and a powerful 16 core Intel® Xeon® D processor for fast analysis.

Robustness

It features extended temperature from 0° to 48°, shocks and vibration, while keeping its performance.

 Outstanding energy efficiency for AI inference

9 TPLOP/Watt thanks to the careful choice of optimized components.

Wireless networking

The BullSequana Edge can communicate via radio, 4G, LTE or Wi-Fi and can thus be be fully independent from wired networks.

Interactivity

 Both multi-source and multi-format data can be analyzed in real-time. 9 TPLOP/Watt thanks to the careful choice of optimized components.

 Adaptability 

The server is designed to fit perfectly in any type of environment. It can be placed in stand alone, in a rack or on a wall (DIN rail or VESA wall).

Data security and protection

Data is critical at the edge. The effectiveness of IoT applications will rely on data security and data privacy. BullSequana Edge is equipped to avoid the risk of physical intrusion with an Intrusion Sensor which disables the machine in case of physical attacks. A secure boot process can be put in place including signed firmware, bootloader and the OS, protected by a FIPS 140-2 certified TPM and encrypted disks.

Designed to run AI at the edge with optimal performance  

Bull Sequana Edge can host up to two powerful Nvidia Tesla T4 GPUs or optional FPGA’s. This enables the inference of complex AI models right at the edge with lowest possible latency. Together with its powerful 16 core Intel Xeon CPU, BullSequana Edge provides an outstanding compute power-pack for the implementation of most demanding machine learning applications.

Designed to accelerate Edge Data Analytics with hyperconverged infrastructures

BullSequana Edge has been designed to enable open source based hyperconverged infrastructure solutions which enable flexible resource sharing between nodes, centralized management and security hardening.

Streaming Analytics solutions such as Spark and Kafka can flexibly be deployed on this stack. Through its powerful GPUs BullSequana Edge also supports accelerated machine learning algorithms enabled by RAPIDS and similar frameworks.

BullSequana Edge at work

Edge Computer Vision

The Atos Edge Computer Vision is a complete environment running on BullSequana Edge and dedicated to deliver computer vision capabilities.

Read moreDownload the Bullsequana Edge features

Edge Data Analytics

Atos Edge Data Analytics enables organizations to improve their IoT business models with predictive and prescriptive solutions.

Read more

Edge Data Container

Atos Edge Data Container is an all-in-one solution, serving as a decentralized IT system, running at the edge.

Read more

Certifications

Contact us

Mehdi Kasmi – Head of Global Offerings BullSequana Edge

mehdi.kasmi@atos.net

Interested in our BullSequana Edge server?