Get your sample by choosing part number(s) from the product page or using the search box above.
Future of Datacenter with Ethernet SSDs
With many applications like AI, data analytics, machine learning, and image recognition getting mainstream, the demand to process and transmit massive data securely in the least time is increasing. The present-day data center networks, storage architecture, and servers are under constant improvements to meet the new demands. Ethernet can be considered as the future of data networking compared to other data transferring technologies. It is widely accepted, well documented, established and supports a wide range of applications from consumer PCs to datacenters.
These days as speed and performance requirements are advancing SSDs are now taking the center stage of mass storage media. This is where the most advanced storage technology, Ethernet SSD takes its position. As the name suggests, these are storage drives that can transfer massive data in seconds via Ethernet but not confined to it. Ethernet SSDs make use of the NVMe-oF (Non-Volatile Memory Express- over Fabric)protocol, that combines PCIe® based storage with data transmission through communication fabric like Ethernet or Fibre Channel.
PCIe® SSDs generally use network interface cards(NIC) in transferring data among storage arrays. However, Ethernet SSD will have NIC embedded with the storage drive PCB itself. This helps to reduce the latency making data transfers direct and more efficient. It also simplifies the storage stack, cut down the design complexity and enhance performance.
Ethernet SSD pin out is specified in SNIA SFF9639 and SFF REF TA1012 that uses SFF8639 connector and TA1002 connector, respectively. Amphenol ICC collaborates with our customers to enable Ethernet SSD. We provide SFF8639 connectors to our customer to build 25G ethernet SSD and also Gen 5 SFF8639 connector model for the customer to study 50G ethernet SSD performance.
When traditional PCIe methods are replaced with simple low power Ethernet fibers, it helps in optimizing the data center infrastructure with more efficient storage requirements in a cost-effective way.