NVMe is disrupting all-flash array (AFA) architectures by providing higher performance, in terms of lower latency, higher IOPS, and higher bandwidth. The benefits of the protocol’s ability to support higher queue depth, commands, and parallel queues, is astounding for scaling virtual machine workloads or achieving near-real time insights (read more about NVMe here) .
There is no doubt that NVMe can address performance at scale, however it also raises a few adoption challenges to maximize the utilization efficiency in existing environments.
New technology demands new learnings.
Top 4 Challenges to NVMe Adoption
As we work with our customers to adopt NVMe-based solutions, including SSD devices, storage platforms, and all-flash storage arrays, here are a few of the biggest challenges they faced:
1. Ecosystem Interoperability: With any new technology or infrastructure there is a risk of interoperability challenges to support existing virtualized workloads like VMware®, Oracle®, Microsoft SQL Server, SAP® or compatibility with existing network infrastructure
2. Multi-protocol support: With workloads distributed across multiple hypervisors/OSs how can you achieve multiple industry standard storage protocols support?
3. Seamless Manageability: By adding another technology to your architecture, there is a management aspect that needs to be addressed. This includes operational challenges like multiple admin interface, long deployment time, support for OSV plug-ins and third-party management tools
4. Silo of NVMe storage: While NVMe addresses performance challenges, it does come at a cost. SAS and hybrid storage arrays or servers are equally important for capacity demand. How do you manage multiple array controllers and multiple interfaces?
Addressing NVMe Challenges
Understanding these challenges helped us design our IntelliFlash™ N-Series arrays . We combined the performance of NVMe with software-defined flash management, and rich data services to help alleviate some of these initial challenges and directly align to the challenges mentioned above. Let me walk you through some of these features.
The N-Series offers plug-and-play capabilities that help shorten the deployment time, while providing full support for existing virtualized workloads, such as VMware, Hyper-V, Oracle, MS-SQL, and SAP. With up to 8x 10GbE, 4x 40GbE Ethernet, 4x 8G and 6x 16 FC data ports, we built this system so that you can integrate the system in your existing networking infrastructure without demanding any specialized connectivity options.
Our N-Series provides block (FC, iSCSI) and File System (NFS and SMB) support. We know that FC and iSCSI are essential to support performance hungry virtualization platforms, like VMware vSphere® and Microsoft Hyper-V and database applications, like Oracle, and MS-SQL workloads. And, NFS is the leading choice to support future AI and ML workloads. That’s why we made sure that a rich OSV and application ecosystem is supported for all protocols, including unified management and plugin options.
As data gets more complex, you’ll need a straightforward way to manage it. We built a simplified, unified, and intuitive Web UI interface that makes configuring, managing, and monitoring very easy. A wizards-based graphical user interface (GUI) is provided for fresh deployment and common storage tasks. Further integration with OSV plug-in tools and REST API options enable seamless manageability of the array for end users. Lastly, a highly optimized dashboard is available to monitor space usage, cache hits, CPU hits, network throughput, and other characteristics of the functionalities supported by the array so you can gain better insights into your data infrastructure.
As data gets more complex, you’ll need a straightforward way to manage it.
Silo of NVMe Storage
Having multiple array controllers and multiple interfaces to manage creates several silos that bring a high operation cost. Our IntelliFlash all-flash, hybrid, and NVMe arrays are all based on the SAME operating environment, with the SAME Feature Set, and SAME user experience. This capability helps reduce the storage silos in the data center and provides a unified solution across multiple workloads and storage mediums.
How to Leverage NVMe-based Storage in Existing FC SAN and iSCSI Environments
Businesses across the globe are grappling with how to not only capture and preserve their data, but more importantly, how to transfer the data into actionable insights. NVMe is going to be a driving force in enhancing demanding workloads, and mission critical applications.
Join me in an upcoming webinar where I’ll discuss how to address the challenges of transitioning to NVMe-based storage and how to leverage NVMe-based solutions in existing FC and iSCSI storage environments. Save your seat below, or watch on demand!
Rajeev Sharma, Sr. Product Manager, Data Center Systems has over 16 years of experience in enterprise storage, focused since 2014 on NVMe.