Supermicro NVMesh Storage Solutions
Low-latency distributed block storage for web-scale data centers
Solution Overview
New application workloads are driving the need for new storage architectures: Cloud, Edge, AI and IoT bring new latency requirements – therefore companies need to rethink their storage strategies.
Supermicro and Excelero have joined forces to deliver a high-performance, low-latency distributed block storage solution leveraging NVMe flash. The Supermicro NVMesh solution enables unmodified applications to utilize pooled NVMe storage devices across a network at local speeds and latencies. It was designed to combine virtually unlimited Supermicro servers with NVMe to scale capacity and performance.


Solution Key Benefits:
- Maximize the utilization of your NVMe SSDs
- Predictable application performance
- Supports standard hardware & protocols
- Use as block storage or with any file system
- Easy to manage & monitor

Product Brief
The Supermicro NVMesh Reference Architecture implements the next-generation of intelligent infrastructure platforms designed to provide applications with all of the raw performance you can expect from local storage — and all of the flexibility, manageability and scalability typically experienced with traditional storage area network-based solutions. Leveraging the latest NVMe and PCIe standards, this reference architecture uses a low-latency, high-bandwidth networking to connect compute and storage together in a flexible way that will fit almost any application’s architecture requirements. Deploy this RA in storage-centric, compute-centric or mixed storage and compute configurations, and integrate it with existing data center application servers as desired.

Supermicro NVMesh Reference Architecture
SuperStorage EDSFF Server | BigTwin Server | Ultra Server | |
---|---|---|---|
Platform Form Factor | 1U 1 node | 2U 4 nodes | 1U 1 node or 2U 1node |
Storage Media | 32x 4TB or 16TB NVMe SSD devices | Up to 24x 15.3TB NVMe SSD devices | Up to 10 or 24x 15.3TB NVMe SSD devices |
CPU (two per node) | Intel® Xeon® Scalable Processors for X11 SuperStorage | 2nd Generation Intel® Xeon® Scalable Processors for X11 BigTwin | Intel® Xeon® processor E5-2600 v4 family for X10 Ultra or Intel® Xeon® Scalable Processors for X11 Ultra |
Memory | suggested 192GB DDR4 2933 per node | suggested 192GB DDR4 2933 per node | suggested 192GB DDR4 2933 per node |
Network | up to 2 Mellanox ConnectX 5 VPI NIC | up to 8 Mellanox ConnectX 5 VPI NIC | up to 2 Mellanox Connect 5 VPI NIC |
RAW Capacity | up to 512 TB per node | up to 92TB per node (367TB per BigTwin) | up to 367TB per node |
Available Data Protection | Fully Distributed RAID, enabling linear scale of performance and capacity | ||
Infrastructure Support | TCP/IP; RoCEv2; Infiniband | ||
IOPs Performance | up to 10M random 4k | up to 20M random 4k | up to 5M random 4k |
Throughput Performance | up to 24GB/s | up to 76GB/s | up to 24GB/s |
IO Latency | Alike the local NVMe SSD Latency | ||
Performance Density | 24GB/s or 10M IOPs per RU | 38GB/s or 10M IOPs per RU | 12GB/s or 2.5M IOPs per RU |
Management | CLI, GUI, REST API | CLI, GUI, REST API | CLI, GUI, REST API |
Operating System Support | Linux | ||
File System Support | POSIX like as ext4 and xfs, plus GPFS, BeeGFS & Lustre | ||
Excelero Software | NVMesh Client | ||
NVMesh Target | |||
Available Excelero Service | NVMesh 3y premium support | ||
NVMesh Install Service, On-site | |||
Available Supermicro Service | 3-year onsite 24x7x4 service / 3-year onsite next business day service | ||
Onsite Integration Services Onsite Survey Onsite logistics Racking & Cabling (with 3rd party switches) Racking & Cabling Engineering Drawing Releasing (Test included) |