AI initiatives, particularly in large language models (LLMs) and generative AI, necessitate a vast array of data sources for training and operational purposes. These data sources are often varied and can include both structured (like databases) and unstructured data (like images and texts). As such, architects faced with implementing AI projects must consider storage solutions ranging from SAN and NAS to object storage. The article discusses the advantages and challenges of different storage types, emphasizing the need for organizations to find the right balance to support AI's extensive data requirements.
"When it comes to training LLMs, the more data sources the better. But, at the same time, enterprises link LLMs to their own data sources, either directly or through retrieval augmented generation (RAG) that improves the accuracy and relevance of results."
"A lot of AI is driven by unstructured data, so applications point at files, images, video, audio - all unstructured data," says Patrick Smith, field chief technology officer EMEA at storage supplier.
Collection
[
|
...
]