eWEEK content material and product suggestions are editorially impartial. We might earn cash once you click on on hyperlinks to our companions. Learn More.
Storage and archiving know-how deficiencies may thwart AI initiatives, stopping synthetic intelligence know-how from attaining its full potential. While the business is concentrated on different potential roadblocks—together with the inordinate quantities of power consumed by AI knowledge facilities, the dearth of energy availability, the shortage of GPUs and high-powered CPUs, or the dearth of information middle capability in key markets—the inefficiency of many archiving options is being given much less consideration.
This subject is the topic of a latest report from the Active Archive Alliance (AAA), “How Arctive Archives Support Modern AI Strategies.”
“While much of the focus of AI adopters has been on the front end of data processing and analytics, the sustainability of AI workflows must now address the long-term retention and protection of what will be massive and persistent volumes of data,” mentioned Rich Gadomski, Head of Tape Evangelism for FUJIFILM North America and AAA co-chair. “A modern strategy is needed to manage the growth and volume of data, and this can be provided by a sensible active archive implementation coupled with intelligent data management.”
Consider the scale of many giant language fashions (LLMs). All the information they analyze must be saved someplace. Companies with limitless budgets can afford to maintain lots of it in reminiscence and the remainder on solid-state drives (SSDs), however for many organizations, the value makes that out of the query. Even conserving all the information on spinning disks is an costly technique to go when you find yourself coping with the vastness of AI knowledge repositories.
According to Furthur Market Research, storage capability surpassed one zettabyte (ZB) in 2016, 4.eight ZB in 2022, and is predicted to achieve 50 ZB by 2035. Clearly, AI power consumption is destined to change into an actual problem.
That’s the place an lively archive is available in. It gives organizations with an clever knowledge administration layer that may transfer knowledge the place it belongs based mostly on exercise, price, and efficiency. Data wanted by AI purposes is shifted to the place it may be quickly analyzed. Otherwise, it sits in a decrease storage tier comparable to exhausting disks, optical disks, or tape. The knowledge layer ensures there are by no means lengthy delays in ready for knowledge to be made accessible. Automated tiering takes care of information motion. This accommodates prices and gives eco-friendly long-term storage or efficiency storage as wanted.
As organizations create LLMs and originate AI purposes, the necessity for extra storage and environment friendly archiving will solely enhance. The subject of AI and power consumption just isn’t going away, and that can regularly convey storage and archiving dialogue to the fore.
“AI will accelerate demand for active archive storage,” mentioned Mark Pastor Director, Platform Product Management at Western Digital. “Archived data will have more value than ever before and will therefore need to be stored for a long time and actively accessed during its life.”
