How AI can help enterprises reduce data storage costs


We’re excited to bring Transform 2022 back in person on July 19 and pretty much July 20-28. Join AI and data leaders for insightful conversations and exciting networking opportunities. Register today!

The amount of data managed by enterprises around the world is growing. According to one source, the total amount of data created, captured, copied and consumed worldwide in 2020 was approximately 64.2 zettabytes, equivalent to one trillion gigabytes. Unsurprisingly, companies are reporting that the cost of storing their data is also rising. In a 2018 Enterprise Storage Forum questionnairebusiness executives said high operating costs, lack of storage capacity and aging equipment were among their top concerns.

The rising cost of storage has prompted many companies to adopt cloud options, which offer the advantage of low entry costs. But with costs rising as more businesses go online – a Pepperdata report found that more than a third of companies have budget overruns for cloud services of up to 40% — IT leaders are exploring alternatives.

On the cloud side, an emerging group of startups is applying AI to the problem of managing cloud spend. Vendors such as Densify and Cast AI claim that their AI-powered platforms can recommend the best storage configuration for a company’s workloads by considering different requirements. Other technology vendors have turned their attention to on-premises systems and created algorithms that they claim can lower storage costs, either with hardware suggestions or with new file compression techniques.

“Data storage today faces several challenges: storage implementations often consist of a variety of different storage media, such as memory, flash, disk drives, and tapes. In addition, organizations use multiple storage arrays based on access protocols…or based on the criticality of the workloads,” Gartner research VP Arun Chandrasekaran told VentureBeat via email. “Using AI has the potential to streamline data lifecycle management based on data criticality, performance, security and cost requirements.”

Cloud Optimization

During the pandemic, pressures to digitize operations have led a record number of companies to move to the cloud. According to a recent survey by O’Reilly, 90% of organizations were using some form of cloud computing by 2021, while Flexera’s State of the Cloud report shows that 35% of companies will have spent more than $12 million on cloud activities by 2021.

The adoption trend has led to startups developing AI-powered platforms designed to customize usage to control spending. One is Densify, which analyzes workloads in private data centers, Amazon Web Services, Microsoft Azure, Google Cloud Platform, and IBM’s cloud offerings to determine how much CPU, RAM, and storage they need — then suggests ways to save. Densify can use already available log data to start optimizing right away. After that, the platform will continue to assess cloud provider pricing changes, application needs, and new products to see where customers can further reduce costs.

“Usually you have 50% of the savings in two to four weeks,” CEO Gerry Smith told VentureBeat in an earlier interview. “Depending on where the savings are located, within another two to four months, [you’ll get] 100% of the savings.”

Cast AI, a competitor to Densify, similarly uses AI to optimize cloud spend. The platform supports major cloud service providers, connects to existing clouds and generates a report to identify cost-saving opportunities.

“We have other models that use global datasets for predictions of market characteristics,” CEO Yuri Frayman told VentureBeat in October 2021. “For example, we are training a global model to predict instance preemptions by machine type, region, availability zone, and seasonality. shared autonomously with all customers and all data is used to continuously train the model.”

On-premises and compression

For businesses that haven’t yet made the move to the cloud — or who have spread their data across cloud and on-premises environments — there are solutions like Accenture’s Storage Optimization Analytics, which combines search and AI to understand business content and automate data classification. .

Accenture claims it reduces storage costs by detecting duplicate or near duplicate content, enabling customers to move or archive the right data at the right time. Storage Optimization Analytics also automates migration to lower cost storage and tracks storage savings, calculating overall return on investment (ROI).

IT provider Rahi Systems offers a similar service called Pure1 Meta, which uses AI models to predict capacity and performance and advise on workload deployment and optimization. Pure1 Meta can run simulations for specific workloads, generate answers to capacity planning questions, and ostensibly help increase resource utilization.

An Nvidia AI model that compresses videos.

AI is also increasingly playing a role in file compression. For videos, music, and images, AI-based compression can provide the same or nearly the same level of visual quality with fewer bits. Another advantage is that it is easier to upgrade, standardize and deploy new AI codecs compared to standard codecs, because the models can be trained in a relatively short time and – most importantly – do not require special hardware.

Websites like Compression.ai and VanceAI use models to compress images without sacrificing quality or resolution. Qualcomm and Google have been experimenting with AI-driven codecs for both audio and video. And DeepMind, owned by Alphabet, has created an AI system to compress videos on YouTube, reducing the average amount of data YouTube has to stream to users by 4% without noticeable loss of video quality.

Looking to the future

Gartner’s Chandrasekaran notes that adoption of AI data management technologies, which fall under the “AIops” category, remains quite low. (AIops platforms aim to improve IT by using AI to analyze data in the organization of tools and devices). But he adds that the pandemic has been a catalyst for adoption as organizations strive to automate faster to respond to “rapidly changing” circumstances.

Recent studies agree. According to Emergn, 87% of companies expect their investments in automation skills to increase in the next 12 to 26 months. And in a 2020 K2 poll92% of business leaders said they view process automation as essential to success in the modern workplace.

“There is a lot of ‘AI washing’ in the industry today. Therefore, it can be frustrating to monitor supplier claims and implement a solution that delivers ROI. AIops requires a lot of integration,” Chandrasekaran said. “For teams that are not proficient in designing and maintaining complex data environments, a robust AIops implementation can become a dream. There must also be a culture shift, in which organizations are prepared to make data-driven decisions.”

Looking ahead, Chandrasekaran expects to see more “versatile” AI-powered storage management solutions than the products already on the market. These solutions could enable more intelligent automation and recovery workflows through the use of AI, he said.

“AI techniques can help optimize the placement of data on the right storage tiers, balancing performance and cost. In addition, AI can help with better data infrastructure availability, enabling companies to access data faster and create a reliable infrastructure,” Chandrasekaran added.

VentureBeat’s mission is to be a digital city square for tech decision makers to learn about transformative business technology and transactions. Learn more about membership.

Leave a Reply

Your email address will not be published.