Chelsio Demonstrates 100Gb/s iSCSI Performance

Chelsio Communications, Inc. announced performance Ethernet storage network capabilities of its line of T6 100GbE Unified Wire controller.

Click to enlarge

The demonstration shows company’s T6 silicon enabling 100Gb/s line-rate iSCSI performance for a cost-effective enterprise storage target solution built with volume, off-the-shelf hardware and software components.

Showcasing the iSCSI hardware offload capabilities of the T6 Unified Wire controller, the demonstration shows how all-flash and hybrid storage array OEMs can enable such arrays with performance iSCSI target capability that is compatible with the existing iSCSI ecosystem and supports the routability, scalability and reach advantages of iSCSI. The complete T6 iSCSI benchmark paper detailing the hardware/software configuration used and results achieved is available here. A video of the actual demonstration is here.

Click to enlarge

iSCSI offload is especially suited for enterprise deployments:
It requires only software peers (that are supported in all volume OSs) and as such can be non-disruptively deployed in existing environments.

The second source consists of OS built-in software-only stacks. The enterprise customer will never be in a line-down situation because of the adapter hardware. It is therefore much easier to address dual source requirements.

It is supported in Linux (LIO Target, Open-iSCSI Initiator), FreeBSD (Target and Initiator), Windows (Initiator – the popular iSCSI initiator in the industry), VMware, and other OSs and distributions.
It is built on the proven TCP/IP protocol and as such can scale, route and handle congestion and network loss resiliently and robustly without needing special switch features.
It allows the same redline performance as other storage networking protocols such as NVMe over Fabrics or iSER, and has built-in RDMA, but it does so over the proven TCP/IP protocol (removing any challenges pertaining to TCP/IP ‘overhead’).
It does not suffer from interoperability requirements of fabrics-based storage protocols such as iSER or NVMe over Fabrics.
It is a proven $3.6 billion market today without any technology adoption risks.
It runs at 100Gb/s and beyond, and will scale consistently with Ethernet evolution.
The CPU savings resulting from iSCSI offload results in an immediate bill of materials reduction in storage applications – same or more performance at lower cost.
iSCSI is mature, robust, tested, multi-vendor, deterministic, and ubiquitous. It has a much higher performance than FC which it replaces.
It allows enterprise OEMs to consider consolidating their adapter vendor product line and hence truly leverages the economies of scale of Ethernet.

Click to enlarge

T6 100G Unified Wire controllers enable enterprise storage systems that are purpose-built to deliver optimized storage performance for various application workloads in mission-critical virtualized, private cloud environments. The T6-enabled storage systems lets IT departments balance performance, capacity and features with game-changing storage economics. Firm’s enabled storage systems also integrate into diverse enterprise IT environments by enabling SAN and NAS deployments from the same array through iSCSI and TCP/IP protocol offload for iSCSI block-level protocol, as well as NFS and CIFS file-level protocols.

“The extreme performance benefits of All-Flash storage arrays can be unrealized if using inefficient networking” said Seamus Crehan, president, Crehan Research, Inc. “Chelsio’s T6 100G Unified Wire solution with iSCSI offload enables consistent very high-throughput, low-latency access to Flash storage for datacenter applications.“

“Our T6 Unified Wire controller establishes a groundbreaking performance bar for Ethernet storage by enabling a 100Gb/s line-rate iSCSI storage target solution with very high efficiency,” said Kianoosh Naghshineh, CEO, Chelsio. “Our iSCSI offload technology allows 100GbE to completely supersede proprietary storage fabrics, without requiring the deployment of new protocols that force an infrastructure forklift upgrade. With transition to all-flash arrays, all storage systems are essentially HPC systems and require offload technology to realize the Flash performance capabilities and value.“