While over the past few years there has been a massive shift for companies to move data to the cloud for ease of use, cost savings, scalability, and more, there are still advantages to storing certain data workloads on-premises. Why? This could be for better performance, to meet regulations and compliance laws, to enhance security, or to have more local control over data for day-to-day operations.
We could apply the old saying here that "everything old is new again." After years of moving data center workloads and applications to cloud environments, some organizations are repatriating parts of their workloads back to local data center infrastructure.
But this doesn't mean organizations are falling back to the same storage environment they once managed years ago. Demands have changed. New workloads have evolved. And new regulatory and compliance initiatives have altered data management requirements. In addition, performance, reliability, and data security needs have evolved in today's digital landscape where on-premises, edge, and public and private cloud infrastructure work together to provide flexible, dynamic, and scalable storage on demand.
The Need to Hug Your Servers, or Your Data
The cloud is and will remain a highly valuable resource for meeting business objectives. According to a 2023 study by KPMG and HFS Research, companies are increasingly adopting software-as-a-service (SaaS) throughout the enterprise, and within the next two years artificial intelligence (AI) and analytics (47%), infrastructure and middleware (47%), productivity and collaboration (44%) and security (43%) will be the highest workload application use cases in the cloud.
However, some use cases are providing enterprises and IT managers with reasons to keep data on-prem or repatriate their data, or some of their workloads, back onsite. This gives IT managers the ability to physically hug their servers — if they really wanted to. In other words, data is stored within reach where latency is reduced, and they have better control over and ability to secure critical data.
An Enterprise Strategy Group report says that the following applications may not be good cloud candidates: those with specific or significant networking requirements related to latency or bandwidth (cited by 32% of IT professionals surveyed); high-value applications/data that have compliance, regulatory, or data sovereignty significance (29%); AI and/or machine learning (ML) workloads (29%); applications or data with strong security requirements (28%); and data analytics workloads (25%).
There's no shortage of opinions and resources out there telling you why you should or should not move your workloads to the cloud or vice versa. And there is not a one-size-fits-all approach. Security, compliance, scalability, cost, performance, location, and talent all must be carefully considered before deciding which architecture will work best for the business or specific applications.
With that said, let's talk about a couple of scenarios that might work best on-prem.
Real-time, low-latency applications: 5G and IoT are propelling the need for faster, real-time decision-making at the edge where data is generated. Advanced workloads around generative AI (GenAI) and the IoT require lightning-fast latency that can only be delivered by moving compute closer to the data. Moving data to a cloud data center hundreds of miles away may not provide the speed necessary to take full advantage of these systems.
Consider the healthcare industry where faster diagnosis, decision-making, and responsiveness are critical for optimal patient treatment and care. 5G's low latency has paved the way for AI-edge-supported medical devices, such as laparoscopes, revolutionizing medical surgeries by offering real-time data and insights. Or envision a scenario where a surgeon equipped with AR glasses, connected to a 5G network, gains access to real-time patient data and medical imaging during surgery. These technological advancements not only aid in diagnosis and help to identify anomalies but also enhance surgical decision-making both before and after procedures.
The proliferation of IoT devices, ranging from wearable health monitors to smart medical equipment, also generates tons of patient data and brings a wealth of opportunities in their fields of computer- and image-guided assisted surgery. Most of this data will not live on-prem only. Eventually certain data will move to the cloud, whether public or private, delivering a hybrid architecture approach that is continuing to reshape the future of healthcare.
Data compliance and regulations: Consumer data protection laws can also move organizations to the point where they feel a need to "hug their servers," bringing data back on-premises where they have better control over and ability to secure critical data that may be governed by data compliance and regulatory laws. This could include controls on who has access to data, location of storage, or other data sovereignty requirements.
One of the most prominently known regulations that covers data privacy is the EU General Data Protection Regulation (GDPR), which includes guidelines for businesses that collect personal data. There are others, like the Health Insurance Portability and Accountability Act (HIPAA), the Family Educational Rights and Privacy Act (FERPA), and more. For businesses bound by these regulations, establishing robust data governance policies, encryption protocols, and access controls may be paramount. On-premises deployment can help ensure thorough compliance.
Security: The intensifying need for robust or zero-trust data security initiatives is compelling companies to reconsider their data storage strategies, also prompting a trend toward on-premises solutions. Think about defense companies, financial institutions, or government entities. Some of their data can be stored on a public cloud platform, while other sensitive data cannot even risk a slight hint of potential exposure. In the latter scenario, a local, air-tight on-premises system may provide better security. By retaining data on-premises, companies have greater control over their security infrastructure, enabling them to tailor security defenses to specific needs and implement strict access controls. Unlike the cloud, where security measures are often managed by third-party providers, on-premises solutions empower organizations to directly oversee their ecosystem and customize their security protocols, allowing a more responsive and adaptable defense against emerging threats. For some organizations, on-premises data security and management helps fortify data architectures and can instill confidence in the protection of sensitive data and information.
Rethinking Storage Architecture
The ongoing transformation fueled by IoT, 5G, and AI and the dynamic interplay between endpoint, edge, and cloud have led data center managers to reassess and rearchitect their storage infrastructure. The decision between using on-premises and/or cloud storage involves careful and thoughtful consideration of each. While on-premises solutions offer granular control over security and data management, cloud solutions provide scalability and flexibility. The growth of edge computing and the increasing reliance on 5G further complicates the landscape, introducing new dimensions to data storage decisions.
As organizations contend with these choices, it becomes evident that there is no one-size-fits-all approach. The optimal strategy relies on a careful evaluation of specific business needs, security considerations, and the dynamic technology landscape.
The future of data storage ultimately lies in a strategic blend of on-premises and cloud solutions that give an organization a flexible and adaptive approach to meet the evolving demands of the digital era. Put workloads where they make sense, both technically and economically, to meet the performance, cost, scale, and security needs in the most efficient and effective way possible.
Brad Warbiany is director of HDD product marketing at Western Digital.
