The Rise of Ethernet and Data Networking

The Rise of Ethernet and Data Networking

By Enterprise Networking Mag | Wednesday, March 06, 2019

Data NetworkingData transmission in less time but with a secure environment is putting a huge strain on existing data center network and storage architectures. Also, the demands of the emerging innovations like artificial intelligence, machine learning, and image recognition data analytics have a lot on their plate. Gartner has predicted that by 2021, 50 percent of all data centers would use solid state drives (SSDs) for high-performance workloads.

• Fiber Channels

Fiber channel SANs whose demise had been predicted long ago, though a huge install base, isolation of block storage traffic and ease of management appears to ensure its short-term relevancy. Although Fibre channels provide a regularly updated roadmap offering a well-structured path to upgrade become obsolete now, as the technologies have advanced in terms of future enhancements. Hence, FC no longer holds performance advantage as Ethernet speeds have recently surpassed Fibre channel speeds.

Artificial Intelligence Companies: CapeStartCoalesceCognitivDXiHAL24KLiulishuoPez.AI.

• Ethernet Storage

Ethernet Storage Fabric (ESF) supports unique storage-centric features like delivering smart networking and storage offloads- including Transmission Control Protocol (TCP), Remote direct memory access (RDMA), non-volatile memory express (NVMe) over fabrics to provide highest levels of performance and lowest latency possible. ESF can use RDMA to bypasses the TCP/IP stack and enable direct server to server and server to store data transmission amid application memories without loading the processing overhead onto the CPU. It allows ESF to deliver significant performance and efficiency gains. It can now be integrated with the cloud, storage and enterprise Ethernet networks with RDMA over Converged Ethernet (RoCE) to reduce latency.

Using RDMA alongside the NVMe-oF storage protocol for data transfers between the host and attached storage systems can further bring down the latencies from milliseconds to microseconds; the technology is in its infancy stage and is expecting to be integrated by the Enterprise market. It is envisaged that NVMe interfaces will continue to replace the current technologies such as SAS and SATA to support full line rates demanded by ML and AI applications and further exploit the adoption of RDMA.

Check out: Top Artificial Intelligence Companies.

New Editions