Here’s the repository of Chelsio White Papers produced prior to 2020
- Azure Stack HCI Configuration using iWARP RDMA (2019)
Quick Start Guide for Windows - SoftiWARP Performance with Chelsio 100GbE (2019)
- Accelerated Inline IPsec Communication with T6 25G (2019)
- We put the iWARP in NVMe-oF (2019)
A Chelsio presentation at NVMe Developer Days Conference 2019 - High Performance 25G S2D for AMD EPYC (2019)
Using AMD EPYC 7551 Platform & Chelsio T6 iWARP RDMA Adapter - 100G NVMe-oF TCP (2019)
Chelsio T6: Bandwidth, IOPS and Latency Performance - Which RDMA? (2019)
- 100G SPDK NVMe over Fabrics (2019)
T6 iWARP RDMA Bandwidth, IOPS and Latency Performance - NVMe-oF with iWARP and NVMe/TCP (2018)
A Chelsio presentation at NVMe Developer Days Conference 2018 - Axellio demos WSSD cluster with Chelsio 100GbE (2018)
Great performance with Ease of Use - 6th Generation CNA Solutions (2018)
High Performance Cost Effective Server Solutions - Windows Guest RDMA Performance at 100Gbps (2018)
Networking to the VM at Bare Metal Speeds & Latencies - Windows FCoE Initiator Performance (2018)
Line Rate Throughput and IOPS - Windows iSCSI Performance at 100Gbps (2018)
Zero Copy Performance using iSCSI Offload - Windows TCP/IP SR-IOV in Virtual Environments (2018)
Line Rate Performance with Chelsio T6 100GbE in Virtual Environments - Windows iSER Performance at 100Gbps (2018)
RDMA Performance and Efficiency without Fabric Overhaul - Windows d.VMMQ Performance (2018)
Line Rate Performance with Chelsio T6 100GbE - WSSD Configuration using iWARP Mesh Topology (2018)
Quick Start Guide for Windows - Microsoft Windows Unified Wire and CNA Solutions (2018)
Lower Latency, More Data, More IOPs, More Protocols, More Options, More Value - Chelsio & IBM Storwize have Partnered (2018)
Chelsio’s R-NIC a key piece in IBM’s storage performance enhancement - XIO demos WSSD cluster at Microsoft Ignite (2018)
XIO demonstrates great performance with Chelsio 10Gb adapters - Dell EMC Storage Center & Chelsio (2018)
Chelsio 25Gb & 100Gb iSCSI adapters,a key piece in Dell EMC’s all flash Storage Center array performance enhancements - Windows NVMe over Fabrics Performance (2018)
Line Rate Performance with Chelsio T6 100GbE iWARP RDMA - Secure Data Replication at 100GbE (2018)
Chelsio T6 iWARP RDMA Solution for Microsoft Storage Replica - 100G Network Performance for AMD EPYC (2018)
Using AMD EPYC 7551 Platform & Chelsio T6 Adapter - 100G NVMe over Fabrics JBOF (2018)
T6 iWARP RDMA Bandwidth, IOPS and Latency Performance - 100G OVS Kernel Datapath Offload for AMD EPYC (2018)
Using AMD EPYC 7551 Platform & Chelsio T6 Adapter - 100G NVMe over Fabrics for AMD EPYC (2018)
Using AMD EPYC 7551 Platform & Chelsio T6 iWARP RDMA Adapter - 100G iSCSI Performance for AMD EPYC (2018)
Using AMD EPYC 7551 Platform & Chelsio T6 Adapter - FreeBSD 100G TOE Performance for AMD EPYC (2018)
Using AMD EPYC 7551 Platform & Chelsio T6 Adapter - The True Cost of Non-Offloaded NICs (2018)
Chelsio T5 & T6 Lines of Offloaded NICs - Migrating to Microsoft Storage Spaces Direct (2018)
King County (WA) Library System upgrades to Storage Spaces Direct and Hyper-V - 100G iSCSI Offload Performance (2018)
Using White Box JBOF Storage Platform & Chelsio T6 Adapter - High Performance NVMe Over 100G iWARP RDMA (2018)
Using White Box JBOF Storage Platform & Chelsio T6 Adapter - Demartek Evaluation: Chelsio Terminator 6 (T6) Unified Wire Adapter iSCSI Offload (2017)
A Demartek, Chelsio Benchmark Report - Chelsio Unified Wire for Arm (64-bit) Platform (2017)
Chelsio T5/T6 Adapters & Qualcomm Centriq 2400 Processor Family - OVS Kernel Datapath Offload at 100GbE (2017)
Using Qualcomm Centriq 2400 Arm Platform & Chelsio T6 Adapter - Concurrent 100G Offload & Encryption on Arm (2017)
Using Qualcomm Centriq 2400 Arm Platform & Chelsio T6 Adapter - Industry’s First 100G iSCSI Offload for Arm (2017)
Using Qualcomm Centriq 2400 Arm Platform & Chelsio T6 Adapter - NVMe Over 100G iWARP RDMA Fabrics (2017)
Using Qualcomm Centriq 2400 Arm Platform & Chelsio T6 Adapter - Secure Data Replication at 25GbE (2017)
Chelsio T6 iWARP RDMA Solution for Microsoft Storage Replica - High performance 100G iSCSI Storage Solution (2017)
Chelsio T6 iSCSI Throughput, IOPS, Latency and CPU Utilization - Concurrent Offload & Encryption at 100GbE (2017)
Secure TCP/IP processing with T6 Inline-TLS/SSL Crypto Function - iWARP Support in Scalable Xeon Platform (2017)
Frequently Asked Questions - iWARP RDMA – Here and Now (2017)
Intel, Chelsio and Cavium Technology Brief - NVMe over iWARP Fabrics Performance (2017)
T6 Based High Performance and Low Latency NVMe-oF Solution - Apache TLS/SSL Acceleration at 100GbE (2017)
Chelsio T6 vs. Intel AES-NI vs. Software Enabled Encryption - Accelerated IPsec-VPN Communication with T6 (2017)
Chelsio T6 vs. Intel AES-NI - Disk Encryption with 100GbE Crypto Accelerator (2017)
Chelsio T6 vs. Intel AES-NI vs. Software Enabled Encryption - Chelsio 100G DPDK Performance (2017)
High Speed User Space Packet Processing with Terminator 6 - S2D Performance with Network QoS (2017)
Chelsio iWARP RDMA solution for Windows Storage Spaces Direct - S2D Performance with iWARP RDMA (2017)
Chelsio T520-CR vs. Mellanox ConnectX-4 - iWARP RDMA – Best Fit for Storage Spaces Direct (2017)
High Performance, Ease of Use and Seamless Deployment - High Performance S2D with Chelsio 100GbE (2017)
Chelsio T6 iWARP RDMA solution for Windows Storage Spaces Direct - T6 100GbE Crypto Offload Performance (2016)
Superior Throughput with Chelsio Crypto Offload Solution - Demartek Evaluation of the Chelsio T580-CR 40GbE iSCSI Offload Adapter (2016)
A Demartek, Chelsio Benchmark Report - Low Latency Performance at 100GbE (2016)
Chelsio iWARP RDMA, WireDirect (WD-TOE/TCP/UDP) Solution - Industry’s First 100G Offload with FreeBSD (2016)
Zero copy TCP @100Gbps with Less than 1% CPU usage using DDP - Meeting Today’s Data Center Challenges (2016)
A Chelsio, HPCwire, Tabor Report - Quad to Small Form Factor Pluggable Adapter (2016)
QSFP+ to SFP+ Pluggable Adapter Module - Introducing NVMe over 100GbE iWARP Fabrics (2016)
Chelsio T6 Based High Performance and Low Latency NVMe-oF Solution - High Performance iSCSI at 100GbE (2016)
Chelsio T6: Industry’s first 100G iSCSI Offload solution - Windows SMB 3.1.1 Performance at 100Gbps (2016)
Line Rate Performance with Chelsio T6 100GbE iWARP RDMA - Chelsio Terminator 6 ASIC 100GE Crypto Offload (2016)
Enabling Secure Network Interconnects - The Chelsio Terminator 6 ASIC (2016)
Next-Generation Converged Secure Network Interconnects - High Performance Storage Replica at 40GbE (2016)
iWARP RDMA benefits for Windows Server 2016 Storage Replica - Hyper-Converged Scale-Unit with Chelsio 40GbE (2016)
Chelsio T5 iWARP RDMA solution for Windows S2D + Hyper-V - iWARP/RDMA Benefits in Windows 10 (2016)
Chelsio T5 Client RDMA w/ Windows 10 Enterprise - Resilient RoCE: Misconceptions vs. Reality (2016)
- Chelsio Precision Time Protocol (PTP) (2016)
- Resilient RoCEv4: The Experiment Continues (2016)
Comparing RDMA over Ethernet Alternatives Configuration - Concurrent Support of NVMe over RDMA Fabrics and Established Networked Block and File Storage (2016)
A Chelsio presentation at FMS 2016 - Configuring NVMe over iWARP RDMA Fabrics (2016)
Quick Start Guide - High Performance NVMe over 40GbE iWARP (2016)
Peer to peer DMA through PCIe bus – No Host Memory Required - Configuring Chelsio iSCSI Solution (2016)
Quick Start Guide - High Performance S2D with Chelsio 40GbE (2016)
Chelsio T5 iWARP RDMA solution for Windows Storage Spaces Direct - High Performance iSCSI at 40GbE (2016)
Chelsio T5 iSCSI solution delivering 5M IOPS - Configuring Storage Spaces Direct (2016)
Quick Start Guide - Windows Server 2016 Storage Spaces Direct (2016)
A Chelsio Solution Brief - 100G iSCSI – A Bright Future for Ethernet Storage (2016)
A Chelsio Presentation at SNIA DSI 2016 - High-Performance GPU Clustering over 40GbE iWARP (2016)
A Chelsio presentation at NVIDIA GTC 2016 - RoCE Exposed (2015)
The Trials of Deploying Infiniband over Ethernet - RoCE Fails to Scale (2015)
Repetitive Protocol Surgery or the Dangers of Sliding Down the RoCE Path - iSCSI or iSER? (2015)
A Chelsio Presentation at SDC 2015 - Implementing NVMe over Fabrics (2015)
A Chelsio Presentation at SDC 2015 - Ethernet connected JBOD (EBOD) using Chelsio 10GbE Adapters (2015)
A Microsoft presentation at SDC 2015 - High Performance Disaster Recovery at 40GbE (2015)
Chelsio T5 iWARP RDMA solution for Windows Storage Replica - Linux iSER Performance at 40Gbps (2015)
RDMA Performance and Efficiency without Fabric Overhaul. - GPUDirect RDMA over 40Gbps Ethernet (2015)
High Performance CUDA Clustering with Chelsio’s T5 ASIC. - Linux 40GbE DPDK Performance (2015)
High Speed Packet Processing with Terminator 5. - NVMe over 40GbE iWARP RDMA (2015)
Throughput and Latency Benchmark Results. - Building Enterprise-Class Storage Using 40GbE (2015)
Unified Storage Hardware Solution using T5. - ESXi iSCSI Performance at 40Gbps (2015)
Comprehensive Storage Virtualization with Terminator 5. - Complete Connectivity Solution for Microsoft Cloud (2015)
T5 iWARP RDMA and NVGRE Performance. - Chelsio T5 Unified Wire Suite for OpenPOWER (2015)
Accelerate Network & Storage Performance on POWER Architecture. - FreeBSD 40GbE TOE Performance (2015)
High Efficiency with Direct Data Placement (DDP). - The Case Against iWARP (2015)
Addresses myths and allegations disseminated as part of marketing RoCE. - Server-Based Security (2015)
Total Security with Distributed Hardware Firewalls. - GPUDirect over 40GbE iWARP RDMA (2015)
High Performance CUDA Clustering with Chelsio’s T5 ASIC. - iWARP RDMA Performance for Docker (2015)
RDMA Latency and Throughput over 10Gb Ethernet. - T5 Offloaded iSCSI with T10-DIX (2015)
High Performance End-to-End Data Integrity for Ethernet SANs. - Linux 10GbE Latency with Busy Poll Sockets (2015)
Benchmark Study of Chelsio’s T520 and Intel’s X520 Adapters. - Linux OpenStack iSCSI at 10GbE (2015)
Throughput Benchmark Results. - FreeBSD 40GbE Netmap Performance (2015)
Disruptive Performance with Terminator 5. - Accelerating Hadoop with iWARP RDMA (2015)
Execution-Time Benchmark Results for Big Data Offload. - Linux 10GbE NIC/TOE Performance (2015)
Chelsio T520-LL-CR vs. Emulex OneConnect OCE14102-UM. - Linux NIC and iSCSI Performance over 10GbE (2014)
Chelsio T520-LL-CR vs. Intel Fortville XL710. - Windows 10GbE NVGRE Offload Performance (2014)
Chelsio T520-LL-CR vs. Intel Fortville XL710. - Linux iSER Performance (2014)
iWARP RDMA over 40Gb Ethernet vs. IB-FDR. - Windows SMB 3.0 Performance at 40Gbps (2014)
RDMA vs. NIC Throughput, IOPS and CPU Utilization. - Linux NIC and iSCSI Performance over 40GbE (2014)
Chelsio T580-CR vs. Intel Fortville XL710. - iSER: Frequently Asked Questions (2014)
A Chelsio Technical Paper. - NFS/RDMA over 40Gbps iWARP (2014)
A Chelsio Presentation at SDC 2014 - Next Generation iSCSI Enterprise Grade Data Integrity and Performance (2014)
A Chelsio Presentation at SDC 2014 - iSCSI Heritage and Future (2014)
iSCSI Built-in RDMA Compared to iSER over iWARP and IB. - Linux 10GbE NIC/iSCSI Performance (2014)
Chelsio T520-SO-CR vs. Intel X520 Latency and CPU Efficiency. - Lustre over iWARP RDMA at 40Gbps (2014)
Comparing Performance over Ethernet and IB-FDR. - SMBDirect Latency on Windows Server 2012 R2 (2014)
Realize Full SSD Storage Performance with T5 RDMA over Ethernet. - FCoE at 40Gbps with FC-BB-6 (2014)
Line Rate Throughput and 4M IOPS over Terminator 5. - NVM Express over Fabrics (2014)
High Performance SSD Interconnect with NVMe over Chelsio iWARP RDMA. - NFS/RDMA over 40Gbps Ethernet (2014)
Boosting NFS with iWARP RDMA Performance and Efficiency. - OpenIndiana 40Gbps TOE vs. NIC Performance (2014)
Throughput Benchmark Results. - iWARP: Ready for Data Center and Cloud Applications (2014)
RDMA Over 40Gb Targets Data Center and Cloud Applications. - NVGRE Offload for Windows Server 2012 R2 (2014)
High Performance Network Virtualization with Terminator 5. - T5 netmap Performance on FreeBSD (2014)
Disruptive Networking I/O with Terminator 5. - Windows Server 2012 R2 40G NIC Performance (2014)
Chelsio T580-SO-CR vs. Mellanox ConnectX-3 VPI. - iWARP Speed for Virtual Machines (2014)
High Efficiency VM Networking with T5 RDMA over Ethernet. - Offload Traffic Failover (2014)
Fault Tolerance across Ports and Adapters. - iWARP: From Clusters to Cloud RDMA (2014)
RDMA over Ethernet Goes Mainstream. - RoCE – Plug and Debug (2014)
In Search of a User Friendly RDMA over Ethernet Solution. - SMBDirect 40 GbE iWARP vs 56G Infiniband (2014)
Chelsio T5 40 Gb/sec Ethernet vs Mellanox ConnectX-3 56Gb/sec Infiniband Adapters on Windows Server 2012 R2. - Ideal solution for Big Data applications – Apache Hadoop (2014)
Chelsio Terminator 5 (T5) 40/10/1GbE Unified Wire Adapters. - NFS with iWARP at 40GbE vs. IB-FDR (2014)
Throughput and IOPS Benchmark Results. - iSCSI at 40Gbps (2014)
Line Rate Throughput and 3M+ IOPS with Terminator 5. - FreeBSD NIC at 40Gbps (2014)
Extreme Performance with Terminator 5. - High Performance iSCSI for Virtual Machines (2014)
Software Defined Storage with Terminator 5. - Solaris/OpenIndiana at 40Gbps (2014)
Line Rate Throughput at Small I/O Sizes with Terminator 5. - Storage Fabric Unification with Ethernet (2014)
Realizing the Benefits of Network Convergence with Terminator 5. - Windows Server 2012 R2 SMB Performance (2014)
RDMA vs. NIC Throughput, IOPS and CPB. - FCoE at 40Gbps with FC-BB-6 (2014)
Line Rate Throughput and 2M+ IOPS over Terminator 5. - RoCE is Dead, Long Live RoIP? (2014)
A Chelsio Technical Paper. - Network Direct Chelsio 40GbE vs Mellanox 56G IB (2013)
Latency and Bandwidth Benchmark Results on Windows Server 2012 R2. - 40Gb TOE vs NIC Performance (2013)
Bandwidth, Connection/Request/Response, Apache Bench and SCP Results. - 10Gb TOE vs NIC Performance (2013)
Bandwidth, Connection/Request/Response, Apache Bench and SCP Results. - 40Gb Ethernet: A Competitive Alternative to InfiniBand (2013)
An IBM Benchmark Report. - Chelsio T520-LL-CR vs. Solarflare SFN7122F (2013)
Microburst, Latency and Message Rate Competitive Benchmark Results. - RoCE at a Crossroads (2013)
iWARP is the no-risk high performance solution for 40Gb Ethernet clustering, leveraging TCP/IP’s mature and proven design, with the required congestion control, scalability and routability, preserving existing hardware and requiring no new protocols, interoperability, or long maturity period. RoCE in contrast is an expensive journey of experimentation, rediscovery and frustration that would take years to complete, if it ever does. - Chelsio T4 iSCSI vs QLogic (2013)
Using Chelsio’s 10Gb Ethernet Unified Wire Network Adapters with the T4 ASIC, Chelsio delivers superior iSCSI SAN performance vs QLogic’s competing product. - Chelsio T4 iSCSI vs Emulex (2013)
Using Chelsio’s 10Gb Ethernet Unified Wire Network Adapters with the T4 ASIC, Chelsio delivers superior iSCSI SAN performance vs Emulex’s competing product. - Chelsio T4 iSCSI vs Intel (2013)
Using Chelsio’s 10Gb Ethernet Unified Wire Network Adapters with the T4 ASIC, Chelsio delivers superior iSCSI SAN performance vs Intel’s competing product. - iSCSI over 40Gb Ethernet (2013)
Using Chelsio’s Unified Wire Network Adapters with the T5 ASIC, Chelsio delivers superior iSCSI SAN performance unparalleled in industry today. - STAC-N1 Benchmark For TCP Traffic (2013)
Showing the lowest latency and jitter for TCP traffic. - STAC-N1 Benchmark For UDP Traffic (2013)
Showing the lowest latency and jitter for UDP traffic. - SMBDirect over Ethernet using iWARP on Windows Server 2012 R2 (2013)
Delivering Superior Performance with SMB over Ethernet on Windows Server 2012 R2. - TCP Offload at 40Gbps (2013)
Chelsio’s Terminator TCP Offload Engine (TOE) is the first and currently only engine capable of full TCP/IP at 40Gbps. - Packet Rate Performance Report (2013)
A Chelsio Benchmark Report - High Frequency Trading Report (2013)
A Chelsio Benchmark Report - Preliminary Ultra Low Latency Report (2013)
A Chelsio Benchmark Report - Delivering Superior Performance with SMB over Ethernet on Windows Server 2012 (2013)
A Chelsio Technology Brief - The Chelsio Terminator 5 ASIC (2013)
An Architecture White Paper - LAMMPS and WRF on iWARP vs. InfiniBand FDR (2013)
A Chelsio Technical Paper - Scaling SMB for Server 2012 With RDMA over Ethernet (2012)
A Chelsio White Paper - RoCE : FAQ Read More… (2012)
A Chelsio White Paper - RoCE: The Missing Fine Print (2012)
A Chelsio White Paper - Highly Available Scale-Out File Server on IBM System x3650 M4 (2012)
An IBM Paper - Chelsio Demonstrates 2 Million+ IOPs Over a Single 10Gb Ethernet Port (2012)
A Chelsio Technical Paper - Low Latency NIC/Adapters Benchmarks (2012)
A Chelsio Technical Paper - RoCE: Autopsy of an Experiment (2012)
A Chelsio White Paper - Migrating Legacy IB Networks to Ethernet (2012)
A Chelsio White Paper - Vivavo Selects Chelsio For Hybrid Cloud Storage Infrastructure (2012)
A Chelsio Customer Case Study - RDMA for Ethernet over Windows Server 2012 (2012)
A Chelsio White Paper - Windows Server 2012 Unified Wire Offering (2012)
A Chelsio White Paper - InfiniBand’s Fifteen Minutes (2012)
A Chelsio White Paper - Unplanned Obsolescence: EOL for FC? (2012)
A Chelsio White Paper - Delivering Superior Performance with SMB Direct over Ethernet on Windows Server 2012 (2012)
A Chelsio Techical Brief - Delivering More Than One Million IOPS (2012)
A Chelsio Technical Brief - Interoperability Testing and iWARP Performance (2012)
A Chelsio/Arista White Paper - Integrating iSCSI Deployments with Existing SAS-Based SANs (2012)
A Chelsio White Paper - Migrating to iSCSI SANs While Leveraging Existing Fibre Channel Infrastructure (2012)
A Chelsio White Paper - Low Latency for HFT (2012)
A Chelsio Technical Brief - Cisco 10Gb Ethernet for ECLIPSE Reservoir Simulation (2012)
A Cisco-Schlumberger White Paper - A Rocky Road for RoCE (2012)
A Chelsio White Paper - 10 Gigabit Ethernet Matches QDR InfiniBand for Clustered HPC (2011)
A Cisco Performance Brief - Delivering HPC Applications With Juniper Networks (2011)
A Juniper/Chelsio white paper - High Performance Networking for Optimized Hadoop Deployments (2011)
A Chelsio White Paper - QoS and Traffic Management (2011)
A Chelsio White Paper - Low Latency Market Data Messaging (2011)
A Chelsio Technical Brief - HPC Converging on Low-Latency iWARP (2011)
A Chelsio Technical Brief - Delivering HPC Applications (2010)
A Joint White Paper From Juniper and Chelsio on HPC Solutions - Performance of HPC Applications over InfiniBand, 10 Gb Ethernet and 1 Gb Ethernet (2010)
IBM researchers compared several applications in the HPC arena using 4x InfiniBand DDR and 10Gb Ethernet using iWARP - Chelsio 10GbE iSCSI: Enabling ‘No-Compromise’ IP SANs For Enterprise Networks (2010)
White paper that reviews 10GbE-based network convergence and how it helps enterprises address storage infrastructure challenges. - Chelsio for Cloud Computing Solutions (2010)
This presentation shows how Chelsio’s Unified Wire adapters are designed to be the ideal Cloud Computing adapter, enabling consolidation, convergence and virtualization of the data center - The Chelsio Terminator 4 ASIC (2010)
This paper examines Chelsio’s Terminator ASICs, focusing on the fourth-generation device known as Terminator 4 (T4). - Chelsio Unified Storage: A Cost-Effective iSCSI Solution (2010)
A product profile by the Taneja Group - Purdue University’s Coates Cluster – World’s Largest 10Gb Ethernet High Performance Cluster (2009)
- iWARP Brings Low-Latency Fabric Technology to Ethernet (2009)
This Ethernet Alliance white paper describes how iWARP RDMA is gaining momentum in the industry as a low-latency HPC solution - Find out why 10Gb Ethernet is a better choice for your HPC environment (2009)
This report highlights performance benefits of 10GbE over DDR IB for real life applications - HP and Halliburton use Chelsio 10GbE adapters in performance benchmark test (2009)
This paper compares the performance of the Landmark® reservoir simulator on HP servers - I/O performance of Hitachi NAS Platform using Chelsio 10Gb Ethernet Adapters (2009)
This Tolly Report shows the high performance NAS Platform 3200 Series from Hitachi with 10GbE - STAC reports highlight Chelsio TOE benefits (2008)
STAC measured performance of a market data stack based on RMDS6 to measure latency and throughput and demonstrated benefits of Chelsio’s TOE. - NCSU tests highlight Chelsio TOE performance (2008)
In recent tests conducted Chelsio’s adapters sustained line rate throughput for various combinations of workload. - Chelsio performance highly suited for financial services (2008)
This report illustrates the performance of Chelsio’s 10GbE solution and its benefits in HPC environments. Chelsio’s performance is compared with that of InfiniBand… - 10GbE shows more than 2x FC performance (2007)
This is a study by NetApp & IBM that shows 10GbE has more than 2x the performance of 4Gb FC for running Oracle RAC and Microsoft Exchange applications. The results are now being done again with TOE. - Sandia Labs compares 10GbE & IB (2007)
This tri-labs study describes the performance of InfiniBand versus 10GbE for cluster sizes of up to 128 nodes. It is shown that for node counts of 14 or more, that 10GbE has 5x the bandwidth of SDR-IB. This is because of a lack of built-in congestion control and adaptive routing in IB and is not a function of the link speed. Therefore DDR IB will behave the same. - Osaka University installs largest 10GbE cluster using Chelsio NICs (2007)
This is a summary of the 700-node TOE cluster at Osaka University . The largest 10Gb installation to date and in the top 50 supercomputer clusters in the HPC space. - NFS over RDMA (2007)
This is a study by Sandia that shows that 10GbE has 4x the performance of DDR IB for NFS applications - Chelsio Unified Wire (2007)
There are several approaches to a unified wire and not all fabric convergence schemes produce the same results. This document describes a top down approach and why Chelsio’s is the right one. - Building High Performance IP SANs with Chelsio’s 10GbE iSCSI Solution (2005)
This white paper, authored by industry analyst Tom Burniece, summarizes the requirements for a complete storage target processing solution for iSCSI at 10Gbps, providing maximum performance over the full range of storage applications. - The Value Proposition of TCP Offload for 10Gb Ethernet (2005)
TCP offload has been a controversial technology, partly because past implementations showed little performance or cost benefits. With the recent advent of 10 Gbps Ethernet, and the resulting CPU-to-network performance gap, the value proposition of a properly architected TCP offload engine is demonstrated both in terms of performance and cost savings. - Jumbo Ethernet Frames: Advantages and Disadvantages (2005)
A decade has passed since Jumbo Ethernet frames were first proposed, and an IEEE standard or significant deployment has yet to be seen. This paper synthesizes the reasons behind the lack of acceptance of Jumbo frames, and further articulates how a TCP Offload Engine using standard 1500-byte Ethernet frames obviates the need for using non-standard Jumbo frames. - Independent Performance Benchmarks of 10Gb Ethernet with TCP Offload (2005)
This whitepaper presents the results of three independent performance studies involving Chelsio’s Terminator TCP Offload Engine. The studies include detailed comparisons of 10Gb Ethernet with TOE, Infiniband, Myrinet and 10GbE without TOE, using both micro-benchmarks and real-world cluster applications. The results consistently show that 10GbE with TOE delivers superior performance compared to the other technologies. - 10 Gigabit Ethernet iSCSI in Action (2004)
This white paper describes how the standardization of iSCSI, combined with the emergence of 10 GbE solutions, has finally made Ethernet a viable high-performance connectivity solution for storage networks. - The Terminator Architecture (2004)
This white paper covers Chelsio’s unique Terminator Architecture, building a case for a VLIW processor vs. multi-processor SOC to terminate TCP and process L5-L7 protocols. - 10 Gigabit Ethernet in HPC (2004)
This white paper describes the evolution of Ethernet and the growing need for 10 Gigabit-per-second solutions in high-performance computing environments.