Industry executives and experts share their predictions for 2023.  Read them in this 15th annual VMblog.com series exclusive.

Storage in the Cloud vs. On Premise - The Growing Divide

By Jeff Whitaker, VP, Marketing and Product, Panasas

We know that all organizations today rely on data to drive growth and innovation. But when it comes to data storage needs, every organization is different. Each has its own unique types of use case applications that must be supported. There is no one-size-fits-all storage model, and different performance needs for different applications require purpose-built architectures. While most companies are interested in a path to the cloud, striking the right balance between successful data management and optimal application performance remains a common struggle in that journey. Because of this, we're going to continue to see a growing divide between applications that are suitable for the cloud versus those that are not, and that divide will be centered around storage performance.

Here are the top five high-performance data storage trends I see emerging in 2023.

High performance computing (HPC) environments will consolidate. Organizations will re-evaluate their storage infrastructure and realize that one storage platform supporting one specific workload is simply not practical and will not scale to support the many different application performances that enterprises, universities, and research institutes demand. More IT teams will see the opportunity in consolidating these environments and will voice the realities to stakeholders that the former approach is too costly and complex to manage.

The Storage as a Service trend will be accompanied by on-premises storage. Storage as a Service (STaaS) will continue to grow in 2023, but the industry - and organizations - will realize that STaaS means different things: a public cloud offering, an individual company delivering it, or a storage provider offering customers a service-model delivery option. We expect to see STaaS as a public cloud offering gain greater traction in 2023, as it is a viable option for many applications, in particular activities associated with developing, testing, science, and algorithms. However, heavy CPU and GPU resources will continue to drive higher and higher performance needs from storage, which is effectively impossible to achieve from the cloud. As a result, there will also continue to be a trend of organizations keeping an element of compute and storage infrastructure on premises. Enterprises are interested in a ‘payment over time' model and as such we anticipate this financial model taking off over the next one to three years.

Hard disk drives are here to stay. The inflection point where flash overtakes hard disk drives (HDDs) continues to be pushed out with supply chain needs and costs. The business model is simply not there to replace hard drives. Hybrid data environments with both flash and HDDs will become more prominent as they address the high-cost challenges with flash at the same time as taking the inefficient data movement to hard drives. As such, we expect to continue to see HDDs as a significant element of the datacenter for the next few years.

Applications' need for speed will drive purpose-built, high-performance storage: In 2023, the race will continue between compute, storage, and network. The thirst for optimal performance will drive the need for parallel architectures that make traditional enterprise storage platforms irrelevant due to the sheer capacity requirements of data-intensive, high-performance applications. GPUs will continue to outgrow the performance capabilities of traditional storage systems and drive the need for purpose-built HPC storage. As such, we'll see organizations leverage a parallel high-performance architecture that will give them the abilities they need from their storage system. The emergence of quantum computing will only intensify the need for IT infrastructure innovation.

The CFD market will grow to solve energy challenges. High-performance data-driven workloads will be increasingly applied to the energy sector over the next one to three years. Computational Fluid Dynamics (CFD) simulation, modeling, and analysis techniques are widely adopted in the automotive, aerospace and defense, electrical and electronics, and energy industries. We expect to see significant CFD research in the clean energy sector, as it will play an increasingly important role in shifting how energy is produced. This research will be heavily driven by the price of gas and net zero targets and by the organizations and governments exploring and optimizing wind-based, ocean-based, and other renewable forms of energy. As the power of data and high-performance applications is realized in industries such as energy with CFD workloads, we expect to see enterprise IT teams pay closer attention to these applications and ensure they are supported by scalable, efficient, and cost-effective infrastructure solutions.

##

ABOUT THE AUTHOR

Jeff Whitaker, former VP of Product at Excelero and a founding member of the cloud team at NetApp, leads the Panasas Marketing and Product Management teams. He is responsible for making inroads into new markets and expanding Panasas’ revenue growth.

Jeff is a veteran technology leader with more than 25 years’ experience developing innovative storage, semiconductor, and networking solutions. For over a decade, he designed products that helped customers around the world transition their complex IT environments to run on the top three public cloud providers.

Jeff holds a Bachelor of Science in electrical engineering from California State University at Chico.