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 networksDownload the position paper
Designed to operate outside of the data center
Bull Sequana Edge reduces network latency and optimizes network bandwidth. It has been designed to be placed at the edge close to IOT devices. It can function in various locations such as airports, factory floors, shops and under critical conditions, for example partially temperature controled conditions. Thanks to its WIFI, GSM and Lora radio capabilities, BullSequana Edge is independent from traditional network connectivity and can be deployed in remote locations which are not equipped for hosting IT infrastructure. It offers very flexible installment choices such BullSequana Edge as desktop, wall or rack mount options.
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.
- Optimum security and privacy: both the data and the physical server are protected by an advanced chain of security measures with the BullSequana Edge
- Immediate responsiveness: data analysis in real-time
- Autonomous: reduced dependence on cloud and datacenter availability and connectivity, ensures that apps are not disrupted in case of limited or intermittent network connectivity. The BullSequana Edge can communicate via radio, GSM or Wi-Fi.
- Interactivity: both multi-source and multi-format data can be analyzed in real-time
- Cost–effective: reduced datacenter infrastructure and networking costs
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 more
Edge Data Analytics
Atos Edge Data Analytics enables organizations to improve their IoT business models with predictive and prescriptive solutions.Read more
The exponential growth of intelligent sensors and devices is generating an unprecedent amount
of data. This is reshaping IT architectures, as increasingly powerful processing and machine
learning inference capabilities are required at the Edge of the networks to enable next generation,
transformative AI and IoT applications. BullSequana Edge has been designed to meet these
challenges, delivering powerful AI inference and streaming analytics capabilities while ensuring
that all data remains safe and secure.
Edge computing is an important element of the post-cloud era, extending rather than replacing the
cloud. It allows data to be processed rapidly at the edge, close to where devices are generating it.
The decentralized and distributed nature of edge computing avoids unnecessary network
transmission to the cloud and enables the near real-time actuation of connected things.