Watch on YouTube
Watch on Vimeo
As AI, cyber resilience, and sovereignty converge at the data layer, the siloed storage architectures most enterprises rely on were not designed for that convergence. At AI Infrastructure Field Day, Scality will discuss our recently announced Scality ADI, an autonomous data infrastructure that provides a new operating model that aligns workloads to the right storage media, performance, and data protection. Our speakers will be Giorgio Regni (CTO/founder), Erwan Girard (CPO), and Paul Speciale (CMO), who will discuss market forces, customer AI workload challenges, and provide details of the new offering. Paul Speciale initiated the discussion by highlighting Scality’s 17-year journey in software-defined storage, having built massive-scale object storage solutions that currently manage 12 exabytes of data and serve over a billion users for major cloud service providers and enterprises. He emphasized that the current market is at an inflection point, driven by the increasing demand for AI, escalating cyber threats such as ransomware, and the critical need for data sovereignty.
The presentation underscored significant customer challenges stemming from these forces. Enterprises face issues with fragmented, siloed storage systems that are ill-equipped to handle diverse AI workloads, resulting in low GPU utilization (often 5-10%, while optimized systems reach 70-80%). This inefficiency is compounded by flash scarcity, high prices, long procurement times, and the rapid growth of data, with many expecting 2-3x increases in data volume without proportional staffing. Scality ADI addresses these pain points by offering a unified data infrastructure that eliminates silos and provides a single namespace spanning hot, warm, and cold storage. It delivers the necessary ultra-fast, low-latency performance to keep GPU clusters consistently fed, while also leveraging cost-effective HDD technology for high-throughput capacity. This hardware independence is crucial for flexibility and cost optimization.
Scality ADI is presented as a comprehensive architectural stack designed to automate and manage data at scale, rather than just another layer. A key aspect is its focus on AI at the edge, where data is increasingly created and processed. ADI supports this evolving topology by extending a single, intelligent namespace from the core data center to hundreds of thousands of edge endpoints, enabling bi-directional data flow, analysis, and policy-driven management for applications such as machine vision, agriculture, and autonomous vehicles. The solution also introduces an open-code model, allowing customers to inspect the code to build trust and sovereignty, and even contribute changes that Scality will manage. This, combined with deep NVIDIA ecosystem integration and the autonomous Scality Guardian agent, positions ADI as a flexible, secure, and manageable on-prem alternative to public cloud solutions for critical AI workloads.
Personnel: Paul Speciale
Watch on YouTube
Watch on Vimeo
As AI, cyber resilience, and sovereignty converge at the data layer, the siloed storage architectures most enterprises rely on were not designed for that convergence. At AI Infrastructure Field Day, Scality will discuss our recently announced Scality ADI, an autonomous data infrastructure that provides a new operating model that aligns workloads to the right storage media, performance, and data protection. Our speakers will be Giorgio Regni (CTO/founder), Erwan Girard (CPO), and Paul Speciale (CMO) who will discuss the market forces, customer AI workload challenges and provide details of the new offering.
Giorgio Regni, Scality CTO and co-founder, detailed how the explosion of “tokens” generated by AI models for inference and training creates a demand for storage that traditional S3/HTTP protocols cannot meet due to excessive overhead. Enterprises face challenges in keeping GPUs fully utilized and managing the lifecycle of massive, rapidly growing datasets. Scality ADI addresses this by providing a unified platform that manages data across extreme, hot, warm, and cold storage tiers, ensuring optimal performance for AI workloads while maintaining data integrity and security. This autonomous data infrastructure allows for seamless data movement and access without requiring applications to be rewritten.
To achieve the necessary performance for AI and GPU-centric workloads, Scality has closely collaborated with NVIDIA, contributing to and utilizing the NVIDIA NIXL library and QObject. This enables S3 over RDMA and a “native,” stripped-down object API, bypassing CPU and software stack overhead and allowing direct communication between flash devices and GPUs. This direct data path significantly reduces latency and increases throughput for extreme and hot tiers (TLC flash), while ADI concurrently manages warm data on QLC flash and HDDs, and cold data on tape or cloud storage. Scality’s contributions, including eight pull requests to the NIXL community, highlight their commitment to open standards and optimizing object storage to compete with file and block storage in high-performance AI environments.
Personnel: Giorgio Regni
Watch on YouTube
Watch on Vimeo
As AI, cyber resilience, and sovereignty converge at the data layer, the siloed storage architectures most enterprises rely on were not designed for that convergence. At AI Infrastructure Field Day, Scality discussed their recently announced Scality ADI, an autonomous data infrastructure that provides a new operating model that aligns workloads to the right storage media, performance, and data protection. Their speakers Giorgio Regni (CTO/founder), Erwan Girard (CPO), and Paul Speciale (CMO) discussed the market forces, customer AI workload challenges and provided details of the new offering.
Scality ADI emphasizes simplifying operations and providing actionable insights for sysadmins through its AI-powered intelligence, Scality Guardian. This agent offers augmented troubleshooting, proactive maintenance, and security monitoring by integrating all internal documentation, customer tickets, product bulletins, and accumulated knowledge into a comprehensive “golden data set.” While the autonomy focuses on identifying issues and providing remediation suggestions, Scality maintains a “human in the loop” for critical data-modifying actions to ensure control. These agents interact via chat-based interfaces or an MCP server/API, with their rights managed through fine-grained IM policies, akin to human users. The predictive maintenance intelligence is designed to learn from patterns observed across Scality’s entire customer base, enhancing its effectiveness.
A significant part of the presentation focused on the burgeoning use cases for AI at the edge, spanning retail stores, warehouses, and mining sites. Scality’s Artisca product is purpose-built for these edge environments, addressing challenges such as limited bandwidth and diverse hardware form factors by enabling local data processing. Artisca facilitates the ingestion and interpretation of massive data streams, such as thousands of video feeds, often in conjunction with partners like NVIDIA (using VSS for video analytics), and stores the processed data locally. The Maestro UI provides a unified view for managing a fleet of these distributed edge systems, ensuring high availability and data resiliency even in remote locations. This architecture supports data lifecycling between the edge and core data centers, offering customers optimal sizing based on their specific application needs.
Personnel: Erwan Girard
Thank you for being part of the Tech Field Day community! Our mailing list is a great way to stay up to date on our events and technical content, and we appreciate your signup.
We promise that we’ll never spam you, send ads, or sell your information. This list will only be used to communicate with our community about our events and content. And we’ll limit it to no more than one message per week.
Although we only need your email address, it would be nice if you provided a little more information to help us get to know you better!
