The Intel Tiber Edge Platform version 24.11 is now accessible. Businesses can create, implement, protect, and oversee Edge and AI solutions from pipeline to production with ease and scalability with to the Intel Tiber Edge Platform. It contains the ability to create edge-native, optimized AI apps as well as the ability to implement distributed, managed, Kubernetes-based edge node infrastructure.
Businesses may take advantage of cloud-like simplicity and operational efficiency across their edge infrastructure and deployments by utilising Kubernetes to orchestrate these apps across the edge infrastructure at scale.
New Features
Along with multi-tenancy support and streamlined concurrent onboarding of many edge nodes, this edition supports the onboarding of a new set of optimized apps and software.
Here are some of the main features of the latest software release, 24.11, among other updates:
Sector-Specific Solutions
New Skills in Spatial Intelligence
- Large-scale tracking across several objects is now possible because to recent improvements in vertical tracker performance and scalability.
- Multiple scene controllers can collaborate on bigger aggregated scenes to advancements in distributed scene hierarchy.
- For simpler scaling of reusable microservices inside Intel Tiber Edge platform, Scenescape has recently incorporated the EVAM visual analytics and Data Store microservices.
Intel Edge Insights System for Video and Time Series Analytics with a unified user interface
- Keycloak-based identity and access management combined with a new user interface provide improved access control for various personas.
- Streamlined AI task implementation with data intake, inference, publication, and visualization solutions.
Applications of AI
Optimizing Large Language Models (LLMs) with OpenVINO
- Intel Arc Graphics Optimizations using KV Cache compression to reduce memory and model load time to enhance LLMs’ first token latency.
- Using data-aware techniques without datasets, a new capability for creating synthetic text data will improve the accuracy of LLM compression.
- Sneak peek: Multimodal AI implementation using the GenAI API for pipelines for enhanced contextual awareness, quicker text-to-visual conversions, and audio-to-text conversions.
- Using a tiny draft model that is occasionally corrected by the full-size model, speculative decoding in the GenAI API improves speed and generates text more efficiently.
Software for Infrastructure
Support for Multiple Tenancies
- For businesses supporting major infrastructure installations like smart cities or industrial IOT, multitenancy support offers safe, scalable deployment options for optimal cost effectiveness and resource optimization.
- Large-scale deployments may be safely managed with ease with to centralized platform features like identity and access management and observability.
Streamlined Edge Node Concurrent Onboarding
- To make the onboarding process easier, pre-register edge nodes and set up OS profiles.
- Up to 50 edge nodes may be onboarded simultaneously, which speeds up setup and improves scalability.
Support for Third-Party Applications and AI Workflow Deployment
- RAG framework for LLM pipelines, deployment and lifecycle management allows for computer vision model development process, and verified third-party application
Concerning the Intel Tiber Edge Platform
Intel’s cutting-edge, industry-leading foundational hardware supports these three core levels of the modular platform:
- Industry Solutions: AI-powered edge endpoints for diverse industries and applications.
- Services for edge-specific AI and app development, deployment, and administration.
- Orchestration, fleet/infrastructure management, and edge node software.
Accessibility
Intel Tiber Edge Platform became generally available in Q1 2024.
The platform will evolve to assist partners and enterprises using Edge and AI applications. Contact local Intel representative to learn more about utilising the platform in your firm or to access its vast technical documentation.
Summary
In Summary to introduction of Intel Tiber Edge Platform, a software program that lets companies create and administer edge computing and artificial intelligence applications, on version 24.11. Large language model (LLM) optimizations utilising OpenVINO, a single user interface for video and time-series analytics, and improved spatial intelligence capabilities are some of the main enhancements.
For easier administration of big installations and faster onboarding for edge nodes, the platform supports several tenancies. It is composed of three levels, each powered by Intel hardware: edge infrastructure, AI & apps, and industry solutions. The platform is accessible anywhere.
