Oracle’s second-generation Roving Edge Device (RED), the newest product in the company’s e Edge Cloud line, offers exceptional processing power, smooth connection, and integrated security at network edges and in remote areas. Numerous workloads, including corporate applications, AI, and certain OCI services, may be operated at the edge with Roving Edge Device(RED) because to its easy deployment, great price-performance, and better security, which includes the option to run isolated or air-gapped.
What is Oracle Roving Edge Device?
OCI cloud computing and storage are made possible at network edges and in remote areas via the Oracle Roving Edge Device (RED), a portable hardware platform with cloud-integrated services.
The second version of the Roving Edge Device builds upon a strong foundation created to satisfy the requirements of security applications. It not only improves basic capabilities but also adds customized configurations to satisfy business needs in a range of sectors.
Introducing the second generation Oracle RED
With more OCPUs, RAM, storage, and better GPU performance, the second iteration of the Oracle Roving Edge Device offers significant improvements over the first generation.
RED is available in three configurations to suit diverse business needs:
Compute optimized | GPU optimized | Storage optimized | |
Shape Name | RED.2.56 | RED.2.56.GPU | RED.2.56.STG |
CPU | One 4th Gen Intel Xeon 8480+ with 56 cores, 350W | ||
GPU | N/A | Three NVIDIA Corporation AD104GL [L4] | N/A |
Available OCPU | 48 | ||
Usable Memory available for VMs | 384 GB | ||
Storage | Four 15.38 TB SSDs with 61.5 TB raw and 54.4 TB (49.8 TiB) usable | Eight 15.38 SSDs with: 123 TB raw and 108.8 TB (99 TiB) usable |
Where can you deploy Oracle Roving Edge Device?
In your data center or at the edge, RED offers the same OCI development and deployment methods, selectable OCI services, and CPU and GPU forms. This drives development in many industries and technical advancement in today’s fast-paced commercial environment. Due to its unrivaled processing power, smooth connectivity, and steadfast security, the Roving Edge Device is ideal for cutting-edge applications that need speed, dependability, and efficiency.
Improvements in performance with the second-generation RED
Milliseconds matter in the fast-paced world of artificial intelligence. Imagine a future in which your network’s borders are infinite and its edge gets exponentially more intelligent. Customers now have more deployment options with to Oracle Roving Edge Device 2nd Generation (RED), which has a new GPU-optimized configuration with compute- and storage-optimized configuration.
Customers gain from low-latency processing nearer the point of data production and ingestion by utilizing the Intel Xeon 8480+ processor’s capability at the edge, which leads to more timely insights into their data. Oracle and Intel collaborated to perform a number of benchmarks over the first-generation RED in order to test this capability.
It used only Intel Xeon processors to run the Llama 2-7B, Yolov10 model, and Resnet50 convolutional neural network (CNN) for the testing. The Intel Xeon 6230T-based first-generation Roving Edge Device is compared to the Intel Xeon 8480+-based second-generation using the following benchmarks:
Deploying Llama2-7B on RED
An autoregressive, transformer architecture serves as the foundation for the Llama 2 family of pre-trained and optimized text generation models. Three models with seven billion, thirteen billion, and seventy billion parameters are included with Llama 2. Oracle benchmarked the Llama 2 7 billion parameter model for this simulation.
Using the Llama 2-7B model, the second-generation Roving Edge Device may achieve response rates up to 13.6 times faster than RED Gen 1, allowing for lightning-fast performance for edge-based large language model (LLM) inferencing. Enhancement of Throughput Intel Xeon 8480+ Processor.
Using the Llama2-7B paradigm, the RED Gen 2 may achieve up to 12.4 times higher throughput, greatly increasing the edge’s capacity for processing LLM data.
YOLO v10
Real-time object identification and precise, low-latency object category and location prediction in pictures were the goals of the YOLO family of models. Oracle compared using the YOLO v10 model on the two versions of Roving Edge Device in this set of benchmarks.
Up to 60% more performance may be achieved by the new RED generation than by the old one. YOLO v10 increased throughput by 67%.
ResNet-50
A convolutional neural network (CNN) architecture called ResNet-50 is a member of the Residual Networks (ResNet) family, a group of models created to tackle the difficulties involved in deep neural network training. Renowned for its depth and effectiveness in image classification tasks, ResNet-50 was created by researchers at Microsoft Research Asia. There are several levels of ResNet topologies, including ResNet-18 and ResNet-32, with ResNet-50 being a mid-sized version.
Using the ResNet 50 CNN, the second generation achieves a response rate that is up to three times higher than the first.
Why deploy with Oracle Roving Edge Device?
Oracle Roving Edge Device is the best option if you need to deploy application workloads at the edge and need a scalable, secure, and adaptable platform with the advantages of cloud computing and cost-effectiveness. Built to execute time-sensitive, mission-critical applications at the edge in both connected and unconnected areas, it is a powerful cloud-integrated service.
Getting Started
Oracle Roving Edge Device is the perfect infrastructure for anybody seeking a high-security, low-latency data processing and scalable environment at the edge because of its affordable, adaptable configurations and capacity to serve computing, storage, and GPU-intensive applications.