Everpure Unveils Data Management Features for Large-Scale AI Operations
IT DAILY ·
✦ Resumen de IA
Everpure announced on the 1st the release of new platform features that simplify enterprise data management and support data-centric large-scale AI operations.
The new features focus on reducing fragmented data context, unpredictable inference costs, and slow, complex deployment processes.
Everpure plans to offer the new features, including Data Intelligence, Native MCP Integration, Turn-Key Deployment, Privacy-First File Intelligence, KVA, DeepReduce, and token optimization capabilities, from October.
Everpure announced the release of new platform features on the 1st. The launch is aimed at simplifying enterprise data management and supporting data-centric large-scale AI operations. In unveiling the new features, Everpure focused on removing bottlenecks to large-scale AI adoption in enterprises.
Everpure identified fragmented data context, unpredictable inference costs, and slow, complex deployment processes as the bottlenecks enterprises face when adopting AI at scale. The feature update is a measure designed to reduce these data, cost, and deployment hurdles.
Everpure presented its vision of "data primacy" at the Pure//Accelerate event in June. The core of that vision is that enterprise architecture in the AI era needs to be designed around data. This feature update extends that vision and strengthens enterprise data management capabilities by putting data at the center rather than applications.
Everpure simplified enterprise data management and strengthened security capabilities based on Everpure Data Intelligence. This responds to the difficulties companies adopting AI agents face in applying secure access and governance controls to distributed data.
Everpure Data Intelligence covers enterprise information across its own platform, public cloud, SaaS applications, and third-party storage. It also searches, classifies, and contextualizes information at the source, with the goal of addressing enterprises' challenges in accessing distributed data.
Through Native MCP Integration, Everpure supports the open Model Context Protocol (MCP). Based on this, it has enabled AI agents and security tools to query the data catalog in real time using natural language and to identify data sensitivity.
The new feature first simplifies adoption by supporting Turn-Key Deployment using the existing Pure1 console. As a result, no separate management server needs to be built, and deployment is possible without complex specialized support.
It then enhances data management visibility by applying Privacy-First File Intelligence. This makes it possible to identify access permissions without directly opening files, and to determine periods of inactivity without directly opening files. The goal is to prevent data exposure risks in advance and efficiently secure storage capacity.
It also adds a function that supports high-performance AI execution without moving data from the original data storage system. Along with that, performance and cost optimization features have also been added.
In addition, it provides Pure KVA (Key-Value Accelerator) based on FlashBlade. This feature places context in GPU memory to cut time to first token (TTFT) by up to 20 times, reduces GPU idle time in multi-tenant environments, and lowers response latency.
From a storage operations perspective, the new feature applies Always-On DeepReduce data compression based on whole-system storage block analysis and sub-block-level similarity detection. This delivers no write-performance degradation and requires no manual work, while helping expand usable storage and reduce hardware and cloud costs.
For AI cost management, it provides an open-weight model-based "intelligent token optimization reference architecture." This helps reduce API token usage from external AI service providers and supports predictable management of AI costs.
Everpure said the new features address the combined challenges of simplification, security, performance, and cost. It also said they support the expansion of agentic workflows on a unified enterprise foundation, and help improve cyber resilience through sensitive data classification and access-rights identification, while contributing to data protection and recovery prioritization.
Prakash Darji, Everpure's executive, said the limitation of enterprise AI is not insufficient model performance, but rather that data is not ready to support real-time autonomous agents. He added that Everpure removes these barriers by continuously applying governance to enterprise data, automating it, and enabling immediate access.
Everpure's latest announcement is about new data management features. These new data management features are scheduled to be available from October.
Source: IT DAILY · Kwon Young-seok
Original: https://www.itdaily.kr/news/articleView.html?idxno=241956
References
This article was produced with the help of an automated content generation algorithm.
Source: IT DAILY
Ver originalThis article was summarized and organized by BizCrush based on the original article from IT DAILY. For exact quotations and full details, please refer to the original article.