Around the Storage Block
1745904 Members
4228 Online
108723 Solutions
New Article ๎ฅ‚
StorageExperts

Re: Top 10 storage trends for 2018 and why you should care

 

Whatโ€™s on the horizon for data storage? With the new year comes a new opportunity to make changes such as replacing aging data center infrastructure, incorporating AI into your storage stack, and revisiting your data protection strategy. 

HPE Storage 2018 Predictions_blog.jpgAs we wrap up the first month of a new year, Iโ€™ve had a new crop of prediction blogs filling my inbox. If last year taught us anything, itโ€™s that our world has become increasingly unpredictable.

However, staying ahead of key IT trends can help you build infrastructure thatโ€™s predictive, timeless, and cloud-ready so you can anticipate and prevent issues across the infrastructure stack, support data growth and mobility, and ensure future flexibility.

With the new year comes the opportunity to consider core business investments such as replacing aging data center infrastructure, incorporating AI into your storage stack, or revisiting your data protection strategy. To help guide your investment decisions, in consultation with my storage colleagues I decided to pen my own list of top data storage industry trends for the year and why you should care.

1. Flash storage adoption will become flashier.

One of the most transformative technologies to hit the storage industry, the ripple effects of flash storage adoption will continue this year. Not only will organizations of all sizes continue to replace spinning disk with SSDs for greater performance, energy savings, space efficiency, and reduced management, but we will continue to see an exciting crop of new technologies like integrated data protection, storage federation/automation, policy-based provisioning and public cloud integration building on top of this flash foundation as they disrupt the market.

Why you should care: Flash storage is the new standard, and its benefits are well established. But the massive growth in flash popularity produced supply chain shortages in 2017 that impacted many buyers. With growing demand and dropping prices, itโ€™s likely that similar shortages could occur this year, so you would be wise to begin evaluating platforms and planning your flash storage purchases sooner rather than later.

 2. Artificial intelligence will gain significant traction in the data center.

Vendors who harness the power of big data analytics will continue to differentiate their products and deliver measurable business impact for customers.

Why you should care: AI will result in huge opportunities to radically simplify operations and automate complex manual tasks. Consider the incorporation of AI as a storage purchasing decision criteria for this year.

3. Integrated data protection will do more than protect your data.

Flash storage arrays have revolutionized storage performance, efficiency, and administration. But the flash storage revolution has also cast new light on the need for greater efficiency in protecting data from threats including various types of attack, application corruption, system failures, disasters, and even human errors. This year more organizations will wake up to the fact that they cannot afford the toll of meeting their data protection requirements using disjointed and overlapping technologies and instead opt to invest in integrated approaches to data protection.

Why you should care: Threats to data security and integrity are a major headache for any organization. Using multiple tools and products not only adds complexity, but can degrade performance and result in multiple copies of dataโ€”adding further cost and management overhead. That makes 2018 a good year to consider a well-designed, integrated data protection solution that not only simplifies management but allows you to leverage efficient data copies for activities such as development and testing or other internal processes.

4. Multi-cloud will become a reality.

Remember when people used to talk about bursting data from on-premises to the cloud and back? In 2018 we might still be waiting for the laws of physics to change, but we will start to see more meaningful data mobility between multiple clouds, including on-prem and public clouds from multiple vendors. But itโ€™s not whether the data can move from one place to another quickly, itโ€™s getting the data located where it can be accessed by multiple, application-optimized compute environments, including multiple public clouds.

Why you should care: As IT organizations start to understand the strengths and weaknesses of the cloud service providers they are working with, they are targeting certain types of applications for different cloud services running on different clouds. Maximizing value from data will require new perspectives and skills that include both on-premises systems administration as well as cloud-services brokering. Consequently, itโ€™s important to ensure that your on-prem storage is cloud-ready.

5. The NVMe fan club will grow.

Flash will continue to be a major force in 2018, but the real excitement will come from innovations that have been born out of the flash revolution. Non-volatile memory express (NVMe) is one such innovation. The use of NVMe made waves last year, particularly as an alternative to SCSI-based interfaces to take advantage of not only flash media but next-generation solid state technologies like Storage Class Memory (SCM). This year we are going to see the use of NVMe over SAS- and SATA-based SSDs continue to grow, but more importantly we will see NVMe make the leap from back end to the front end. In particular, initial adoption of NVMe over RDMA-enabled Ethernet networks will take place, also known as NVMe over Fabrics.

Why you should care: NVMe allows you to take your flash storage to the next level, taking advantage of the massive parallelization of SSDs and next-generation SCM technologies while doing away with SCSI overheads. Although across-the-stack NVMe standardization is still some ways away, the use of NVMe over Fabrics lets you extend the benefits of NVMe between host and array, preserving high bandwidth and throughput and delivering the ability to handle massive numbers of queues and commands.

6. Ethernet storage fabric will come into its own.

The uptick in NVMe over fabrics is tangential to another trend we will see throughout 2018โ€”investment in low-latency, high-bandwidth IP storage networks. Ethernet networking has made grand leaps in terms of performance, pricing, and sophistication over the years, and because it supports any flavor of storage, any storage tier, and even compute, itโ€™s a natural choice for companies looking to modernize their storage network infrastructure, not just those looking at NVMe over fabrics.

Why you should care: Fast storage requires fast networking. The latest storage-optimized Ethernet switches are not only designed to deliver consistent performance and ultra-low latency, but they boast zero packet loss. Consider the ability of Ethernet Storage Fabrics to support converged, hyperconverged, software-defined, scale-out, and distributed storage as well as NVMe over Fabrics, and storage network modernization becomes even more attractive. Still not convinced? Switches are now available that support seamless upgrades to 100Gb/s per port once available, meaning your investment will be future proof.

7. Storage will continue to go software-defined and hyperconverged.

The coupling and decoupling of compute, storage, and networking resources has a long and intricate history. However, in 2018 weโ€™re likely to see the growing taste for bare metal drive a surge in software-defined storage while a yen for simplicity will drive hyperconvergence, which is currently the fastest growing segment of the storage market.

Why you should care: Even though software-defined and hyperconverged models are nearly opposed, increased popularity of both alternative architectures promises you greater choice and greater storage agility in the year to come.

8. Disaggregation of storage from compute will deliver HPC-like performance.

Disaggregation of storage from compute is a new architectural approach that has emerged to address the needs of data-intensive workloads, made possible by low-latency networking and NVMe over Fabrics. As media continues to become more and more durable, the disaggregation of storage controllers from storage enclosures and sharing of data across multiple compute nodes for storage processing has become possible. This architecture takes advantage of NVMe to accelerate workloads at the compute layer

Why you should care: Unlike server-attached flash storage, a disaggregated model allows you greater scalability when it comes to capacity via the use of ultra-dense flash. It also supports high availability features and the ability to manage storage centrally while serving dozens of compute nodes. In sum, disaggregation promises to bring HPC-like performance to the mainstream enterprise, thanks in part to the use of NVMe over Fabrics.

9. Storage class memory will influence buying decisions.

Storage class memory (SCM) is already poised to give the disruptiveness of flash a run for its money. Not exactly memory, not exactly storage, in 2018 we will see SCM start to make big waves. With Intel Optane SSDs now available, the line between RAM and solid state storage has been blurred, and thereโ€™s no going back.

Why you should care: The potential disruptiveness of SCM is undeniable, and you should plan accordingly. In order to leverage the benefits of SCM media, architectural shifts must take place. Although the cost of SCM media will pose a barrier to immediate mass adoption, the pattern here is a familiar one. For this reason, you should strongly consider the architectural flexibility to incorporate SCM technologies as a purchasing criteria for any storage array investments you make this year. Likewise, you should be sure to ask current and prospective vendors about their plans with respect to SCM.   

10. Use of automation frameworks will become the next big skill set.

Infrastructure automation is how modern data centers are managed at cloud scale with greater efficiency and fewer errors. In 2018, automated virtual infrastructures will continue to be strong, but they will face rising competition from containerization and orchestration based on technologies such as Docker and Kubernetes. As a result, in the storage world, automated frameworks for provisioning persistent storage will become the next big skill set for full stack administrators. The readily available number of plug-ins and SDKs will continue to give data center leaders flexible options for completely new platform integrations in 2018.

Why you should care: Containerization can be a key ingredient to building a data center infrastructure capable of keeping up with data and application growth. Automation frameworks reduce operating costs and administration errors by increasing application and workload centricity and also increase staff productivity in containerized environments. With pressure on IT to be more agile and responsive, automation frameworks can be a critical tool in your arsenal.

What do you see on the storage horizon?

There you have the Top 10 trends that HPE sees on the horizon for 2018. To be sure there are many more than these, but the 10 covered here are the ones we see as most important to our customers. Beg to differ? I encourage you to take a moment to weigh in on the trends called out in this post using the form below. What did we miss and what did we get right? What storage market trends do you see taking shape on the near horizon? What new products, features, or technologies do you think HPE should invest in this year? Leave a comment using the form below to share your unique vantage point.

Aligned with the trends: Discover the complete HPE enterprise data storage portfolio.

Vish Mulchand HPE Storage.jpg

 Meet Around the Storage Block blogger Vish Mulchand, Senior Director, Product Management, HPE Storage.

 

 

 

Featured articles:

0 Kudos
About the Author

StorageExperts

Our team of Hewlett Packard Enterprise storage experts helps you to dive deep into relevant infrastructure topics.

Comments
Morten Nyhuus-Eriksen

I think you are forgetting one thing here, and that is NvMe over FC. A lot of us have invested heavely in FC and in my mind that is stil the best way of transporting data at high speed and low latency

Vish Mulchand

Hi Morten

You are absolutely correct. NVME over Fabric for Fibre Channel is a given in our minds. We called out Ethernet because of the magnitude of the change that is coming but I think you are right in that we should have made FC more explicit. Thanks for the feedback

Regards

Vish

Morten, along with what Vish said, I reached out to our Fibre Channel product manager and he had this to say about NVMe over Fibre:

NVMe over Fabrics is designed to work with any suitable storage fabric technology, including Fibre Channel, and HPE is already innovating in this area. For example, our HPE StoreFabric portfolio currently features NVMe-ready switches and directors built to address the requirements of NVMe over Fabrics using Fibre Channel. And our innovations donโ€™t stop there. For example, we already offer NVMe-ready Ethernet switches and have plans to extend Smart SAN technology to the Ethernet fabric via the HPE SmartFabric Automation Suite. This software will manage and orchestrate fabric services across a single data fabric in multi-protocol environments (FC and Ethernet). Smart SAN technology already enables automatic orchestration of fabric zoning without a switch across FC, and Ethernet support is forthcoming. Like our switches, Smart SAN is architected for support of emerging protocols such as those defined by NVMe over Fabrics. If youโ€™re interested in learning more about NVMe, I recommend you download NVMe over Fibre Channel for Dummies.