Related Vendor

Wuxi InfiMotion Technology Co., Ltd.

How independent are the data?

How independent are the data? Cloud and the question of sovereignty.

As public clouds represent one of the main solutions for companies to store their data, reliance on these external providers creates a growing challenge in terms of business continuity and data sovereignty," says Ricardo Mendes, co-founder and CEO of Vawlt, a provider of storage and security software. Other experts on our panel also refer to data sovereignty as a challenge for businesses relying on public clouds. For Cubbit, Signoretti records: "Navigating through complex data sovereignty regulations, such as in the context of GDPR and NIS2, increases complexity for businesses."

Public cloud SaaS services have also introduced additional locations where data needs to be protected. "A big difference today is the number of different places where companies store critical data. This is particularly noticeable when looking at the increased use of SaaS applications. An average mid-size company uses more than 200 SaaS applications, but there are only a few enterprise-class data backup options that can be scaled to protect these applications and allow fast, granular recovery," explains Kim King, Senior Director of Product Marketing at backup software provider HYCU. According to King, over 50 percent of successful ransomware attacks target SaaS applications.

Kerns confirms this view on data protection with SaaS: "Meeting corporate requirements for data protection both in the public cloud and on-premises is a challenge that requires efforts and usually new software solutions." Noting that businesses should learn from the mistakes of others, he adds: "There are cases where some believed that this effort was not necessary for data in a public cloud."

The public cloud - where there's shadow, there's also light.

However, even though public clouds bring challenges, several experts acknowledge that one of the benefits is the democratization of technologies that benefits smaller companies. Norfolk from Bloor Research provides an example of this: "In the past, there was a huge difference between large corporations with real databases and small companies with data stores that didn't support Acid (Atomicity, Consistency, Isolation, and Durability). Cloud technologies have leveled that field."

The current conditions also have an influence on the storage technologies and services offered by providers. According to Serdyuk from Nakivo, the challenges in security are being addressed by the development of even more sophisticated defensive measures against cyber attacks. "Vendors are integrating advanced encryption mechanisms, access controls as well as compliance features into their solutions. Many offer secure enclaves and hardware-based security to meet the evolving threat landscape. However, many storage solutions still lack comprehensive backup and recovery tools," he says.

The need to extract and categorize data from various sources is driving the development of software tools that automate this process. Serdyuk: "Management tools like metadata tagging, version control, and analytic functions are becoming increasingly important."

Multi-Cloud Solutions: Response to Modern Requirements

Companies now also need data management software that supports both hybrid and multicloud infrastructures. "Vendors who recognized this are developing solutions that support easy integration with different cloud providers, on-prem infrastructures, and mixed configurations. They also provide tools for seamless data migration and synchronization across different environments," says Serdyuk.

"We are seeing a pressure for technology consistency across different environments. Some vendors have integrated their technology into the cloud," says Sinclair. An example of such a vendor is Netapp, whose on-prem storage and data management software are now also integrated into public clouds AWS, Microsoft Azure, and Google Cloud. "Others have integrated third-party technologies such as VM Ware or Red Hat Open Shift, which can be used at multiple locations," Sinclair adds.

Regarding the complications arising from the need to maintain data sovereignty and comply with various data regulations that can apply to data storage across national borders in multiple public clouds, Signoretti says, "Providers are prioritizing sovereign solutions for regulated industries such as healthcare and the public sector and are emphasizing compliance in regions like EMEA and APAC. Although they are still subject to the Cloud Act, Microsoft and AWS have recently introduced sovereign cloud storage offerings." The Cloud Act is a US law enacted in 2018, granting US and non-US agencies investigating offenses the right to access company data stored by service providers.

In technical terms, Craig Carlson, a consultant to the Storage Networking Industry Association (Snia) technical council, points to the need to provide AI systems with fast access to data: "AI is currently being addressed by looking into what can be done to drive networks to their peak performance while still being highly scalable. This work is being continued in groups like Ultra Ethernet." A panel called the Ultra Ethernet Consortium is developing an architecture to make Ethernet as fast as modern supercomputing connections, while simultaneously highly scalable, as cost effective as current Ethernet, and backward compatible. Members of the strongly funded consortium include AMD, Arista, Broadcom, Cisco Systems, Huawei, HPE, and Intel.

With the further growth of data volumes, especially in relation to AI and machine learning, entirely new requirements are coming to bear on companies. "More data absolutely leads to ever more complex challenges in relation to storage. Data growth increases requirements in all dimensions," says Tintri's Mowry.

Urgently needed adjustments will likely be available through advancements in data management systems - the metadata tagging, version control, and analytics capabilities mentioned by Guilleaume. Norfolk suspects AI will drive these and other developments: "The biggest challenges today are in the context of AI and its appetite for data - as well as issues of sustainability and resource costs for huge amounts of data. Data quality is also a big problem. Decisions should not be made on outdated, incorrect or biased data. In particular, AI does not handle training well on the basis of biased data."

Data Storage and Sustainability – Where is the Journey Heading?

According to Guilleaume, AI will drive progress in data mobility: "AI/ML will further accelerate the need for data mobility between the layers where the data is stored and those where it is analyzed." These advances in storage management and mobility may not be limited to AI use alone. To this, Carlson of Snia says: "There is always a trickle-down of technology. So, technologies that are being developed for the highest-end AI data centers right now will become mainstream in a few years."

Also, sustainability ranks high on the list of challenges. Roy Illsley, chief analyst at research firm Omdia, emphasizes: "How can we make data storage as green as possible? At some point, we either have to change our lives and the way we do things, or technology will help us. I think it will be a combination of both, which means we need to figure out how to generate less data or be more aware of the data we have."

Arwani cites the impact of data storage on the environment, particularly in terms of CO2 emissions and energy consumption, as one of the current challenges, along with platform interoperability and security. He quotes an estimate from the International Energy Agency (IEA) that data centers' electricity consumption in 2022 accounted for about 1 percent to 1.3 percent of global demand. The IEA predicts that data centers' energy consumption could increase by three to four times by 2026. "These issues are more costly and complex to solve as they require not only technological advances but also awareness and changes in data management itself," Arwani says.

Outlook: Which technology has the potential to solve problems?

On the hardware side, Carlson comments on the fact that the flash technology curve seems to be running out of steam, as it has become much more difficult for flash chip manufacturers to lower costs by packing even more data bits into each flash memory cell. "It will be interesting to see which technology will bring reliable high performance into storage in the next 10 to 20 years," he says. "The long-term use of the current tape-disk-flash model may not be feasible. Hence, we are seeing the development of new and still very experimental technologies, such as DNA storage."

Unsurprisingly, Arwani from Biomemory believes that DNA storage will indeed be the solution: "Vendors are developing more environmentally friendly solutions, such as helium hard drives to reduce energy consumption, or DNA storage technologies, like those being developed by Biomemory and Catalog DNA. These technologies promise a storage density of one exabyte per gram and a durability of several millennia. Moreover, they open up the possibility of new use cases, such as the first space data centers."