The info storage panorama is present process a transformational change. AI is altering how methods have to deal with each giant quantities of knowledge and the pace at which that information will be processed. The place enterprise organizations as soon as relied on conventional SAN/NAS architectures, the explosive development of unstructured information, now measured in petabytes, has made one factor clear – object storage has develop into the dominant expertise for enterprise storage wants.
To assist higher perceive how IT leaders are leveraging object storage, MinIO, Inc., the corporate behind the favored open-source cloud storage software program MinIO, has introduced the findings of its Object Storage and AI Report. The report additionally reveals how AI is reshaping adoption and workload patterns.
The findings of the report have been primarily based on a survey of 600+ IT and software program improvement leaders on storage, AI, workloads, and infrastructure. The survey was performed by MinIO in partnership with UserEvidence.
The MinIO report exhibits that greater than 70% of enterprises’ cloud-native information was in object storage, and this proportion is anticipated to develop. Enterprise leaders forecast that 75% of their cloud-native information shall be in goal storage inside two years.
The highest three elements for this explosive development embody the help supplied by AI, efficiency necessities, and scalability. These three causes are interlinked, as AI initiatives demand efficiency at scale, making object storage the best resolution to deal with giant volumes of knowledge effectively.
So how precisely do enterprises use object storage? The survey reveals that the highest three use instances embody superior analytics, AI mannequin coaching, and information lakehouse storage.
“This analysis confirms what MinIO, AWS, Azure, and Google already know – that object storage is the storage expertise of the cloud – public, personal, or edge,” stated Jonathan Symonds, Chief Advertising Officer, MinIO.
“What’s extra fascinating, nevertheless, is the anticipated acceleration of adoption, pushed by GenAI workloads. Fashionable object shops are uniquely certified to satisfy the calls for of those workloads – particularly throughput efficiency, immutability, and exascale. We’re eager to observe this analysis evolve as enterprises construct new storage architectures to help their AI ambitions.”
Whereas AI help has been important to the expansion of object storage, 96% of the respondents shared that they’re going through challenges as a result of AI. These challenges largely stem from the necessity to handle huge volumes of unstructured information and guarantee constant efficiency at scale, each of that are essential for supporting AI workloads successfully.
IT leaders cited safety and privateness (44%), information governance (27%), and cloud-native storage (25%) because the three greatest challenges to AI success of their group.
The report signifies a rising development towards a hybrid cloud strategy for AI and machine studying (ML) workloads. Whereas the general public cloud stays common, 68% of the respondents shared that they’re involved about the price of working AI workloads, and they’re contemplating a hybrid cloud strategy.
This development aligns straight with MinIO’s capabilities, as its object storage is designed to help each private and non-private cloud environments. Nevertheless, organizations might want to rigorously assess tips on how to stability value and efficiency on this evolving panorama.
“This work represents one of the complete seems to be on the AI storage panorama and there are some eye-opening findings”, famous Ray Rhodes, co-founder at UserEvidence. “IT leaders are clear of their choice for object storage – a discovering that transcended firm measurement or geography. We stay up for partnering with MinIO on extra analysis on this key space.”
AB Periasamy, the co-founder and co-CEO of MinIO
Earlier this yr, MinIO shifted its focus from a general-purpose storage resolution to an AI-centric platform with the launch of AIStor. The platform included options equivalent to promptObject API, enabling pure language queries of unstructured information, and high-speed RDMA help for seamless GPU integration.
MinIO was launched in 2014 to deal with the calls for of unstructured information development. From zero strains of code in 2014, to 1.5 billion docker downloads with deployments throughout the globe, it’s now a preferred alternative for open-source object storage. MinIO co-founder and CEO, Anand Babu (AB) Periasamy, is a BigDataWire Individuals to Watch 2018.
The AIStor launch straight aligns with the themes in MinIO’s Object Storage and AI Report, as each emphasize the crucial function of object storage in supporting AI workloads. With AI adoption accelerating and enterprises rethinking storage, the report’s findings counsel object storage shall be essential for assembly the size and efficiency wants of recent AI.
Associated Gadgets
Are Databases Changing into Simply Question Engines for Massive Object Shops?
Object Storage a ‘Complete Cop Out,’ Hammerspace CEO Says. ‘You All Bought Duped’