Chelsio Logo

Chelsio Update, March 2018

Chelsio continues to make strides with T6 solution’s multiple design wins:

·         iSCSI, iSER, NVMe-oF, TOE, iWARP RDMA, FCoE offload

·         Traffic Management, Classification, IPsec, TLS/SSL offload

·         Integrated switch (daisy chaining the adapters)

·         DPDK, OVS, VxLAN, SR-IOV Offload

·         Low latency, high IOPs, low power

·         2x25G and 2x100G configurations

·         Single chip CNA (iSCSI, iSER, FCoE, NVMe-oF offload initiator)

·         Lowest cost, highest performance Windows S2D solution

·         Designed into and qualified with several tier-1 OEMs

Recently, King County Library System (KCLS) upgraded to Storage Spaces Direct (S2D) with DataON S2D-3110 solution using Chelsio iWARP RDMA-enabled adapters. KCLS was able to consolidate three Hyper-V clusters into one, with 30% increase in performance and delivering ~2.1M IOPS, and also downsized their data center by 15-20 servers.

Create Advertising has replaced its Mac workstations with ones running Windows 10 and paired them with Windows Server 2016 Storage Spaces Direct using Chelsio iWARP adapters in both Windows 10 workstations and Windows Server 2016 servers. As a result, Create has tripled storage performance and cut costs by 60 percent, while video editors can get to work sooner, collaborate and generate more ideas, wow clients, and beat deadlines.

Chelsio announced that iWARP RDMA has been independently validated by Microsoft as enabling efficient and cost-effective scaling of Microsoft cloud deployments without requiring Data Center Bridging (DCB)  capabilities from network infrastructure switches.  This is the main value of iWARP – not requiring a concurrent upgrade of the switch infrastructure or dedicated IT staff.

Chelsio demonstrated a 100GbE JBOF implementation with NVMe-oF and iSCSI Offload using T6 adapters and Samsung Electronics PM1725 NVMe solid state drives (SSDs).  Only Chelsio’s solution can concurrently address the existing iSCSI market and the emerging NVMe-oF market using the exact same hardware.

Chelsio published an analysis of savings in CPU acquisition and operational costs as a result of using an offloaded NIC. For typical fixed applications, the savings could be in thousands of dollars, while increasing system performance.

Demartek recently published a benchmark of iSCSI offload vs. no offload for a class of real world applications. The results clearly demonstrated the necessity for offload for 100Gb iSCSI.

New Products

  • Chelsio released a complete driver compatible for Microsoft Windows Server 2016 and Windows 10 Client supporting L2 NIC, SR-IOV, SMB Direct, Storage Spaces Direct, Guest RDMA, VxLAN, NVGRE, Packet Direct, Network Direct, VMMQ (vRSS), Dynamic VMMQ (RSSv2), DCB, Storage Replica, Ring Backbone and iSCSI Initiator Offload.
  • Chelsio has now enabled new firmware for its server offload (SO) line of cards to enable hardware offload for initiator CNA.  This enables high value NVMe-oF, iSER, FCoE initiator hardware offload.  iSCSI initiator hardware offload will be added in 2Q18 via firmware update.   Why purchase a stateless offload NIC, if you can have the additional hardware offloaded CNA functionality and benefit from the host CPU savings, for no additional cost?
  • Support for T6 IPsec and TLS encryption offload is now part of Linux & FreeBSD kernels.  Why buy a regular NIC if you can buy an encrypted NIC for the same price, and differentiate the end solution?  You can add data-at-rest or browser encryption support, for no additional power, cost, or performance penalty.

Recent Publications

Chelsio added a number of News reports and white papers to its technical library, including:

  • Chelsio recently demonstrated remote storage access performance capabilities of Chelsio T6 NVMe over 100GbE iWARP fabric solution, with T62100-LP-CR adapter in a White Box JBOF. The Chelsio NVMe solution delivered line-rate throughput performance of 93 Gbps and 2.4M IOPS (at 4K I/O size). 

·         Chelsio demonstrated 100G iSCSI Offload performance in a White Box JBOF with T62100-LP-CR adapter. Numbers were collected for offloaded and non-offloaded traffic, demonstrating line-rate throughput of up to 98 Gbps for both READ and WRITE using ramdisk (nullio), and 95.65 Gbps READ performance using SSDs.  

  • A new Quick Start Guide describes quick steps to configure a 4-node S2D cluster without a ToR switch using Chelsio Switchless Ring Backbone Fabric. Switchless Ring Backbone enables connecting a set of servers, without requiring any discrete external network switch – tremendous savings.

Learn More

·         See a complete list of recent Chelsio newsletters here.

·         See a complete list of recent Chelsio white papers here.

·         See a complete list of Chelsio performance reports here.

·         See a complete list of Competitive Analysis reports here.

·         See a complete list of Case Studies here.

·         See a complete list of Videos here.

·         See a complete list of Webinars here.

In the Wild

Microsoft Recommendation on the RDMA alternatives in Windows – Microsoft recommends iWARP as the safe approach to RDMA.

Fibre Channel vs. iSCSI – An SNIA blog on the capabilities and use cases of these technologies.

Regional Library System Simplifies Their Data Center with a Windows Server 2016 Solution - S2D deployment using Chelsio iWARP RDMA.

Creative Advertising Group Deployment of Windows 10 Workstations and Windows Server 2016 Storage Spaces Direct Using Chelsio iWARP Yields Higher Performance and Lower Cost – Microsoft Success Story

Chelsio Storage over IP Enable Data Infrastructures - Addressing the various application and workload demands by enabling more robust, productive, effective and efficient data infrastructures.

Benefits of RDMA in Accelerating Ethernet Storage -  A Q&A on different RDMA alternatives.

Hyper-Converged Infrastructure Improves Health IT ROI - Providing full support for hyper-converged S2D deployments without requiring Top-of-Rack switches to support DCB capabilities.

Chelsio Ethernet adapters capable of leveraging NVIDIA’s GPUDirect RDMA - Higher throughput using Chelsio Unified Wire adapters compared to RoCE.

The Performance Impact of NVM Express and NVM Express over Fabrics – A NVMe-oF prototype on iWARP.

Twitter Link Remember to follow us on Twitter @chelsio1


1H, 2018 Specials


·         Free Shipping for any online order of $2K or more from Chelsio Online Store

·         30% discount on all new PoCs.

·         5% rebate to end-customer for any new case studies.

·         5% discount if accessories are purchased along with the adapters.

Latest Software and Drivers


·         Unified Wire v3.7.0.1 for Linux, OpenPOWER and ARM

·         Unified Wire v6.0.0.14 for Windows

·         Unified Wire v2.0.0.32 for ESXi 6.5

·         Unified Boot Option ROM v2.0.0.8 for DOS.

Visit Chelsio’s Download Center


Chelsio Products


·         Unified Wire Adapters

·         Product Selector

·         T6 Interop. Matrix

·         Terminator 6 ASIC

·         Terminator 5 ASIC


Chelsio Solutions


·         DPDK

·         OVS Kernel Datapath Offload

·         Traffic Management and QoS

·         Microsoft Windows

·         Storage Target

·         Crypto

·         Classification and Filtering

·         CDN

·         High Performance Computing




o Solutions

·         Video Library

·         Case Studies

·         Webinars

·         Performance Benchmarks

·         White Papers

·         Competitive Analysis Reports

·         T6 Competitive Chart

·         Previous Newsletters

·         Online Store