Bringing Intelligence to the Edge: Edge AI-as-a-Service to Accelerate Distributed Intelligence

The new edge-optimized flagship server ThinkEdge SE455 V3 is powered by AMD’s 8004 EPYC Processor that brings 4th Generation Zen4c processor technology to the edge. The processor enables more cores in a smaller footprint, resulting in high energy efficiency and performance. The ThinkEdge SE455 V3 edge server also works in rugged environments, with NEBS Level 3 certification for operations under harsh conditions.

The introduction of an as-a-Service consumption model indicates a demand for a subscription-based pay-per-use model when it comes to deploying edge Artificial Intelligence (AI), which fits the requirements of organizations that have limited resources and funding for edge AI deployments or are looking to build use cases/pilot projects. Increasingly, infrastructure vendors are offering subscription-based offerings as organizations look for support from vendors that can help design, manage, and support edge deployments, leaving organizations to focus on core business initiatives.

The introduction of TruScale for the edge and AI follows the announcement from Lenovo on its intent to double down on innovation investments, including an additional US$1 billion over 3 years to accelerate AI deployments worldwide, including AI devices, infrastructure, and software.

Distributed intelligence, or edge AI, allows organizations to process and analyze data collected at gateway servers or Internet of Things (IoT) devices deployed at the edge of the network, reducing the need for data to travel between devices at the edge and a public/private cloud. By eliminating the need for data transfer to a centralized cloud, time-sensitive applications that require real-time performance such as autonomous vehicles or Automated Guided Vehicles (AGVs) can respond to dynamic changes in the environment, ensure high safety levels, and provide higher performance efficiency and adaptability.

Technology providers recognize the potential of distributed intelligence, with many offering edge AI solutions. Increasingly, ABI Research is seeing the introduction of Edge AI-as-a-Service types of deployment model. Infrastructure vendors such as Dell Technologies with APEX and HPE with GreenLake are all offering pay-per-use solutions, targeted to help organizations that are looking to start deploying edge AI solutions without investing heavily upfront. This will lower the entry barrier for edge AI deployments, and ultimately accelerate the growth of edge AI.

The introduction of Edge AI-as-a-Service will have a significant impact on galvanizing the acceleration of edge AI use cases. Beyond lowering initial Capital Expenditure (CAPEX), organizations deploying Edge AI-as-a-Service can scale AI capabilities up or down based on varying workloads and business needs. Another benefit of consuming Edge AI-as-a Service is lowering technology management complexity. Organizations can focus on executing information obtained from their edge AI deployment and leave the development and management of their edge AI environment to technology providers.

Lenovo’s announcement on TruScale edge and AI services allows organizations to facilitate faster deployment of edge AI solutions. This is crucial for industries such as retail, where retailers are looking at developing edge AI solutions for inventory management, customer insights and personalization, in-store customer traffic, and checkout automation to increase customer experience and gain a competitive advantage.

Moving forward, ABI Research expects Lenovo to continue adding more edge solutions as part of its TruScale portfolio, further democratizing and simplifying AI deployments by providing organizations with access to 150+ turnkey AI solutions, deployed through an as-a-Service model.

Potential use cases for Edge AI-as-a-Service would include organizations that deploy many smart IoT devices in their environment. An example would be mining, which uses cameras, sensor tracking, and drones for daily mining activities. Deploying Edge AI-as-a-Service here will help mining companies focus on core mining operations and leave technology management to the technology providers.

Another example would be a process manufacturing plant, with edge AI enabling plant engineers to have a better understanding of production levels through data analytics. An Edge AI-as-a-Service accelerated server will be able to leverage AI/Machine Learning (ML) tools to handle the heavy AI inferencing workload without the need to transfer data collected at the edge to a centralized data center or public cloud.

As the demand for hybrid/multi-cloud and edge-to-cloud orchestration intensifies, Edge AI-as-a-Service will continue to be an important strategy in any organization's digital transformation journey. Traditional infrastructure providers will continue to transform their business from hardware to providing services. As these infrastructure providers shift to offering pay-per-use models, organizations will need to also be aware of the implications of deploying as-a-Service solutions. Details such as usage metrics, pricing structure, contract terms, and data privacy will need to be scrutinized to avoid future complications.