By Stan Zaffos, Sr. VP, Product Marketing, Infinidat

Data Storage in 2020 - The Move to Smarter Storage

Data storage has come a long way in the 2000s - evolving from storage arrays and archived tape to the AI-powered, software-centric systems that are integral to today's business transformations. What does the first year of the 2020s have in store for storage? Expect more innovation, more "smarts," greater concerns about data protection, and a more strategic role for your organization's storage leader.  

5G will Accelerate the Volume and Velocity of Data Collected 

2020 will be the year 5G enters the mainstream. The new wireless standard will start generating real value for companies deploying IoT projects. It will also start to cause real headaches for procurement leaders who get bills for all the new data that needs to be analyzed and stored.

The projections are eye opening. Gartner predicts 66% of organizations will deploy 5G and 59% will include IoT communications in the use-case for 5G. The number of IoT devices will nearly triple from 2017 to 2025, up to 73 billion, according to IHS Markit. A total of 90 zettabytes (1 billion terabytes) of data will be created on those IoT devices by 2025 - roughly half the 175 zettabytes currently created by all computing worldwide.

Companies deploying IoT projects will need to plan for the data deluge that is coming. They'll need to set up infrastructure and processes to filter the data, pre-analyze it, categorize it, store it and dispose of it.

Today, companies are operating at equilibrium, but 5G will saturate their pipelines. Companies that plan well and allow for some flexibility will execute successful projects at reasonable costs. Those that don't will spend only what they can but be limited in the value they unlock in the limited data they capture and store.

AI Applications Will Usher in the Age of ‘Smarter Storage' 

Artificial intelligence (AI) applications have become so sophisticated, and so ubiquitous, many of us take their complexity for granted. Better known AI-driven applications include autonomous vehicles, online assistants, image recognition in photographs and smart games powered by deep learning (such as Chess or Go).

AI workloads will continue to generate business value in 2020. But, for organizations to increase their reliance on AI, storage vendors will need to make it easier for AI applications to access more data faster, in turn helping the systems learn faster and unlock the value of the data.

AI and deep learning use very large data sets and very fast input/output (I/O) communications. For algorithms to become more accurate, the systems need to perform streams of calculations on large stacks of data. Eventually they spot patterns and get smarter.

As we enter 2020, data sets are getting bigger and demands for instantaneous decision making are becoming more prevalent. This puts stress on the training systems. Expect more demand for smarter storage systems to match the escalating intelligence of the applications themselves. We'll see more investments in tools like software-defined switches to open up more pathways for hardcore analytics; QoS functions to dole out information more strategically; scale-out system architectures; and the ability to deliver data with lower latency.

A National Data Breach Standard in the U.S. Will Put Pressure on Storage Systems

The day of reckoning may not arrive until after 2020, but it's on the way. If the U.S. adopts a national data breach standard that requires companies to encrypt the bulk of their information, many will suddenly find themselves scrambling for more storage space.

The problem is with de-duplication technology. Many storage systems on the market use this technology as a way to reduce their storage needs. The technology looks for patterns and stores a pointer to the first instance of the data. Vendors using the technology can sell low amounts of storage space that expands by several orders of magnitude because the duplicate information doesn't get stored multiple times.

When you encrypt data, the de-duplication advantage goes away. Encrypted data is completely random, so there are no patterns to point to. This means storage systems that use the technology suddenly have much less space on hand. This will force companies to buy more storage to cover the shortfall - or look for alternatives to de-duplication technology.

GDPR in Europe has language recommending encryption as a best practice but no across-the-board rules requiring it. In the U.S. certain states are pursuing GDPR-like privacy rules. If the U.S. passes a national standard mandating encryption, that would change the storage landscape.    

Containers Will Create a More Competitive Storage Environment

Containers and multi-cloud implementations have exploded in recent years. This trend will accelerate in 2020. More enterprises will push to create flexible computing environments where multiple clouds serve specific strategic purposes. They will embrace the flexibility containers promise, creating set-ups where containers can move freely between public cloud, private cloud and on-premises environments.

Aside from making I/O profiles more storage friendly, this flexibility will help improve operational efficiency both on the server side and in storage by simplifying deployment and enabling applications to move between on-premises and various off-premises environments.

Increased use of containers and Kubernetes will help create a more competitive storage environment. Being able to port workloads seamlessly among diverse environments will diminish the strength of vendor lock-in and put pressure on incumbent storage vendors to innovate in areas that improve financial and operational efficiency: lower acquisition and ownership costs; improve staff productivity via more autonomic operation, cloud integration; and new pricing and service offerings.

By 2023 self-Renewing Storage Infrastructures Will Become an On-Premises Reality

Self-renewing storage enables us to provide the illusion of immortal storage by making data migrations and tuning things of the past: the prerequisite to storing massive amounts of immortal data. Self-renewing storage takes the ideas of fault tolerance, self-healing, and virtualization to the next level to create a storage solution that is elastic, intelligent, and always on. Fault tolerance of the storage system to service I/O's in the presence of hardware failures. Self-healing enables storage arrays not disruptively restore full performance prior to hardware failures being non-disruptively repaired.

AI and policy-based orchestration (i.e. data placement) combined with tight integration with the cloud will allow data to reside where it naturally belongs based on the potentially conflicting capacity, performance, and cost demands that exist within any organization without human intervention beyond accepting recommendations. Within the context of self-renewing storage, AI becomes an embedded component of the storage infrastructure rather than an application workload. Self-renewing storage shrinks ownership costs by reducing the time and resources needed to decommission older storage systems which in turn lowers the strength of vendor lock-ins and the cost of infrastructure refreshes: big deals when trying to satisfy insatiable data growth while extending the service lives of installed storage systems.

##

About the Author