COVID-19 has accelerated many digital adoption trends. With businesses and tech teams working from home, tools and applications are being rapidly developed to survive and thrive in this changing world. In 2021, technology acceleration will permeate the industry and continue to impact developers, data scientists and business users across the board.
GPUs in mainstream
NVIDIA’s acquisition of ARM for $40 billion in 2020 was a significant event in the world of GPU compute. ARM’s CPUs are in billions of devices worldwide, providing NVIDIA with many targets for its deep learning technologies in edge scenarios. This will accelerate data analytics in the cloud and edge devices in 2021, including computer vision systems, mobile and embedded device applications.
However, this will not slow down the intense cloud storage and compute trend, where Intel and AMD processors are crunching through big data workloads, often side-by-side with NVIDIA GPUs. In fact, the movement of big data from on-prem data centers and Hadoop data lakes to cheap cloud storage is one of the most rapid trends I’ve seen in data management and analytics-driven by the lower cost of cheap cloud storage. Technologists can expect expanded adoption of hybrid GPU/CPU compute in the cloud.
Specifically, high-value examples include visual analytics and data science on huge data e.g. geospatial use cases such as transportation and logistics, natural resource and energy sector operations. Innovative technologies involving LLVM compilation of parameterized SQL, PostgreSQL with GPU, and native GPU Python functionality will flourish.
I see the worlds of self-service BI and visual analytics melding twofold in 2021. BI and analytics vendors will provide seamless experiences for extending their graphics palettes and embedded functionality with native capabilities. Additionally, marketplaces will share extensions across broad communities of practice.
Visual analysis in cloud-to-edge environments
2021 will bring many technologies to the edge – from data management to visual analysis and data science. Developing, managing and deploying cloud-to-edge machine learning models for computer vision applications requires sophisticated preprocessing, analysis and machine learning capabilities. In addition to event-driven design and integration for model inference, management and operations. Such analytics pipelines have a wide range of applications from customer analytics and demand forecasting in retail, to asset health management, product development and operations scenarios. Image and video feeds from both fixed-site connected cameras and in-motion drones will feed these applications. The advent of new developer environments like AWS Panorama; and cloud services like AWS Lookout for Vision, will provide new tools for developers and fuel growth in this area.
Tech professionals need to keep a 360-degree view on industry changes in 2021. The facets of the developer, programmer, software vendor and analyst roles are varied – leaving many areas of their job to be disrupted. To stay afloat in 2021, professionals should keep an eye on up-and-coming programming languages, and cloud cloud-to-edge services advancements.