AI PaaS for Edge Operators

Participate in AI/ML ecosystem to build the next generation models

Overview: 

  

With an estimated 80 ZB of data generation at the Edge in 2021 and longer term projections of one trillion nodes to be deployed at the Edge by 2035, enterprise-wide digital transformation efforts to gain meaningful predictive insights, using AI/ML modeling brings unique opportunities.

Edge ecosystems are constantly evolving and today many comprise of collation of multiple organizations offering different services to provide an end to end solution. This ecosystem comprises of diverse set of companies viz. 5G National carriers, rural telco's, edge infrastructure orchestrators, low latency Internet providers, IOT solution providers, AI/ ML service providers and  data center & Edge operators. Edge computing creates a sprawling data footprint across a distributed architecture and would need Cloud Native solutions that support heterogeneity of location, remote management and autonomy at scale; enables developers; and integrates well with core processing (in the cloud or data center). IOT devices are generating enormous amounts of data that are required to be processed in real time by AI-inference engines at the edge to enable quick decision making. With the advent of 5G networks, autonomous edge infrastructure orchestration and availability of low-latency internet backbone services the need for integrating AI and analytics Platform is more immediate.  

Intelligent Edge Platform:

 

Zeblok Computational and NodeWeaver are partnering to offer an “Intelligent Edge” Platform, which integrates NodeWeaver’s Edge infrastructure orchestration, and Zeblok’s turnkey AI Platform-as-a-Service, a comprehensive enterprise solution for AI/ML model development, training and deployment at the Edge.

 

Enterprises can now benefit from the following features

  • Autonomously provisioned edge infrastructure

  • Secure networking, with zero trust at Edge locations

  • Turnkey high-performance computing (HPC) orchestration for AI model development

  • Marketplace for curated AI algorithms​

This solution enables data scientists to be productive and IT operations to be scalable and efficient, saving time and money.

Solution Brief

Challenges

Edge Ecosystem 

  • Managing large heterogeneous datasets    

  • Orchestration of infrastructure spread across geographically dispersed Edge locations    

  • Provisioning of secure and scalable networking    

  • Limitations of the current compute infrastructure to process data    

  • Skillsets, capital and time needed to build high performance computing (HPC) orchestration to power AI platforms

Data Scientists

  • Flexibility to push AI-Inferences to any Edge location

  • Need data flow efficiency from Edge to Cloud for training

  • AI/ML Platform with support for Edge-Machine Learning

  • Need better data comprehension

  • Need better AI algorithms

Solution Implementation for an Edge Operator

Zeblok AI + NodeWeaver Stack

NodeWeaver Nano Cloud

NodeWeaver is a software-defined operating platform that enables the deployment of highly resilient, agile and scalable compute clusters capable of running multiple virtual machines and container-based workloads, optimized for running workloads at the edge fully autonomously, integrating self-management, self-optimizing, self-healing features that dramatically reduces cost of ownership. NodeWeaver integrates orchestration, software-defined storage, software-defined networking, multiple hypervisors all managed by the intelligent autonomous system.

  • Hardware Vendor Agnostic

  • Low-Touch Fleet Installation

  • Lightweight

  • Reliability and Scalability

  • Maintenance Simplicity

  • Management Flexibility

  • 3rd Party Monitoring/Management tools

Zeblok AI PaaS

The Zeblok AI platform is an enterprise-ready turnkey AI Platform-as-a-Service (AI PaaS) that enables data scientists to develop, customize and deploy AI projects quickly and efficiently, generate new insights and enhance decision-making capabilities.

The Zeblok platform has been built ground up to provide the simplest end to end user experience including collaboration across teams. Its discrete composable foundational components make it easy to deploy and satisfy a multitude of challenges that enterprises face

  • AI & HPC WorkStation

  • AI-Data Lake 

  • Intelligent marketplace

  • Multi-Cloud

  • Secure AI Runtime

  • Explainable AI driven insights

Benefits

  • Agility: The solution simplifies deployment of highly resilient, agile and scalable compute clusters capable of running multiple virtual machines and container-based workloads, optimized for running workloads at the edge fully autonomously, integrating self-management, self-optimizing and self-healing. IT operations managers will be able to efficiently orchestrate infrastructure and build reliable and secure networking at the edge for single or thousands of edge locations spanned across continents.

   

  • Large Dataset Capable: Data engineers and data scientists can ingest large volumes of data in accelerated AI Data Lake and use CUDA optimized Jupyter Notebooks to train their AI models on GPU-based infrastructure at the Edge.

 

  • Savings: Enterprises will save iteration time during AI model training cycles and will save millions of dollars by eliminating the need to transport large volumes of data to the core data center. Data scientists will be more productive and efficient and will dramatically reduce the time required for AI model development and deployment.

Zeblok logo original-01 (1).jpg
LINKS
ABOUT

Email: zeblok@zeblok.com

Tel: +1 (631) 223-8233

HQ Office:

1500 Stony Brook Road

Stony Brook, NY 11794

Office:

51 JFK Parkway

First Floor West

Short Hills, NJ 07078

    © 2020 Zeblok Computational, Inc.