Harness the Power of NVMe®/TCP to Get Highly Available Storage for Virtualized Applications
Tom Nagelmeyer, Group Product Marketing Manager, VMware
Vivek Sarathy, Director, Product Management, Intel
Kam Eshghi, Chief Strategy Officer, Lightbits
The data center is evolving in response to modern cloud-native applications that need high-performance, scalable capacity, and highly available storage. Organizations can get improved resource utilization, efficient scalability, and rich data services with a composable, disaggregated solution that delivers the high performance and low latency required to run the most demanding cloud-native applications. Lightbits Labs® (Lightbits®) software with a wide range of high-performance Intel® technologies solves customer challenges for VMware® ESXi workloads. Lightbits LightOS® software-defined storage coupled with VMware Sphere® virtualization and Intel’s platform breadth of high-performance technologies for the data center such as Intel Xeon Scalable processors, Intel Optane Persistent Memory, and Intel Ethernet 800 Series Network Adapters address challenges directly, delivering a powerful platform that solves inefficiencies with storage and resource utilization and offers high performance and availability in virtual environments.
Recently VMware announced native support for NVMe® over TCP (NVMe®/TCP) in VMware ESXi 7 Update 3. Concurrently, it was announced that LightOS completed ESXi certification and is listed on the VMware Compatibility Guide as a Day 0 storage partner supporting native NVMe®/TCP with VMware ESXi 7 Update 3. But let us unpack this in greater detail and examine why we are so excited by this important news, given how radically the storage ecosystem is evolving.
Supporting native NVMe®/TCP in VMware ESXi is a major step to deliver an end-to-end NVMe® ecosystem. NVMe®/TCP represents the next-generation iSCSI storage networking standard for linking data storage facilities. One means to overcome the customer challenges of storage resource utilization, workload mobility, recovery times from drive failures, performance, and availability is with NVMe®/TCP.
VMware Support for NVMe®/TCP in VMware ESXi 7 Update 3
A significant challenge for many businesses today is the acceleration of their infrastructure plans. Many organizations cannot afford to wait 18 months to seize new business opportunities. Support for NVMe®/TCP means that organizations can quickly roll out a high-performance, disaggregated storage solution. Organizations can use the TCP/IP infrastructure that they already have in place to scale out to meet shifting business needs better and seize new opportunities.
NVMe® over Fabrics (NVMe-oF™) implementation delivers better performance while lowering deployment costs over existing solutions. VMware clusters can now use NVMe®/TCP on the same network infrastructure and get up to 4x more IOPs and 1/6 the latency for small block operations.
Benefits of NVMe®/TCP:
- Improves storage resource utilization by using simple, efficient NVMe-oF™ to support a hyperscale remote pool of SSDs
- Scales efficiently by enabling composable, redundant storage for scale up and scale out
- Improves performance by supporting disaggregated storage with rich data services at performance and latency that is similar to direct-attached SSDs
More than ever, organizations are focused on ROI. Their big-bet acceleration projects must deliver this as fast as possible. With the speed of digital transformation only going up, investments need to pay for themselves quickly. VMware has a strong roadmap for NVMe-oF™. With the support for NVMe®/TCP, VMware hosts can better support high-performance transactional applications that need high IOPs and low latency block storage solutions.
Simple Storage for Virtualized Applications
Lightbits Labs LightOS is now a fully certified storage partner for VMware’s ESXi 7 Update 3 release, delivering high performance with end-to-end NVMe® over standard Ethernet TCP/IP networks. Lightbits pioneered NVMe®/TCP so LightOS is easy to deploy at scale while delivering performance that is like local flash. For several years, private clouds and cloud service providers have been running cloud-native applications on Lightbits, benefiting from our Kubernetes integration and now there’s a high-performance, highly available storage solution also for VMware users. A single LightOS cluster can deliver over 40M IOPS (random Read) and 10PB user capacity, with less than 200μs latency.
LightOS supercharges VMware environments with these benefits:
- Simple: it’s easy to consume over TCP/IP networks using standard protocols and drivers and has standard-compliant discovery service. Seamlessly integrates into existing infrastructure and VMware environments.
- Flexible: software only, disaggregated architecture allows for independent scaling of storage and compute. Runs on commodity hardware of customer’s choosing for a broad range of use cases.
- High performance: featuring Intel Xeon Scalable processors with Intel DL Boost, Intel Optane PMem, and Intel Ethernet 800 Series Network Adapters with ADQ technology, LightOS delivers high IOPS and low latency with performance that is equivalent to local flash without proprietary hardware and rigid capacity and performance constraints.
- Lower TCO: coupling Intel’s high-performance hardware platform with LightOS optimized for low-cost QLC SSDs, data reduction features dramatically improve efficiency, and disaggregated architecture maximizes resource utilization. Also, no hypervisor is required on storage nodes.
- Highly Available: rich data services such as SSD-level Elastic RAID, per-volume replication, multi-tenancy, snapshots, clones, remote monitoring, and more that result in highly resilient, high availability storage that extends across the entire virtualized platform.
Lightbits software solves customer problems with software-defined storage that is scalable, efficient, and easy to consume. By optimizing for QLC SSDs, LightOS extends SSD endurance by up to 20X. And by scaling storage independently from compute and providing data reduction it dramatically improves efficiency and maximizes resource utilization—all of which reduce TCO. Also, with Lightbits, no hypervisor is required on the storage nodes, further reducing costs.
Accelerating Cloud-Native Applications with High-Performance Storage
Lightbits has been working very closely with our partners Intel and VMware to deliver a completely optimized solution. The combination of LightOS software running on Intel Xeon Scalable processors with Intel DL Boost, Intel Optane PMem, and Intel Ethernet 800 Series Network Adapters with ADQ technology, LightOS delivers high IOPS and optimization for QLC allows us to deliver optimized price-performance. With VMware hosts running on third-generation Intel Xeon processors, they can now accommodate more cloud-native applications on virtual machines and have more scalable, higher-performance storage. Because LightOS uses NVMe®/TCP, companies can rapidly and easily set up high-performance, scalable storage using the datacenter network they already have in place.
VMware with LightOS on storage platforms optimized for Intel’s platform breadth of high-performance technologies enable datacenters to do more with less, allowing for optimal usage of both compute and storage, hence lowering TCO and increasing ROI. Private-cloud infrastructure and cloud service providers’ customers can realize the performance, scalability, and cost-efficiency benefits of this combined solution.
- Webinar: Accelerate Your Cloud-Native Applications with NVMe®/TCP for Virtualized Environments
- VMware Solution Brief
The NVMe®, and NVMe®/TCP word marks are unregistered service marks of NVM Express, Inc. in the United States and other countries. All rights reserved. Unauthorized use strictly prohibited.