SlideShare a Scribd company logo
Confidential │ ©2020 VMware, Inc.
Marvell – Qlogic NIC FastLinQ QL41xxx driver issue
David Pasek, dpasek@vmware.com
VMware TAM
2020, Nov 25
Confidential │ ©2020 VMware, Inc. 2
Problem description - symptoms
Network throughput issue in Qlogic FastLinQ QL41xxx
• Containerized application running in OpenShift on top of vSphere has in some circumstances 3x higher network throughput on Intel X710 (1.9.5) in
comparison to Qlogic FastLinQ QL41xxx (3.11.16.0) with reasonable latency
• The issue was observed in production environment – the problem was mitigated by replacing Qlogic by Intel NIC
• VM on ESXi with Intel NIC can achieve total throughput 300 MB/s (~3Gbps) with application latency 20 ms
• VM on ESXi with Qlogic NIC can achieve total throughput 100 MB/s (~1Gbps) with application latency 2000 ms during the issue
• The question is why the same VM running on the identical server hardware (except NIC), can easily handle 300 MB/s (~ 3 Gbps) on Intel NIC and only
100 MB/s (~ 1 GBps) on Qlogic NIC?
Confidential │ ©2020 VMware, Inc.
Test results from HPE Lab
Full test plan with test methodology and network throughput
results is available on demand
Confidential │ ©2020 VMware, Inc. 4
HPE Lab
Lab environment details
2x ESXi hosts
• Server Platform: HPE ProLiant DL560 Gen10
• BIOS: U34 | Date (ISO-8601): 2020-04-08
• NIC1: Intel X710, driver i40en version: 1.9.5, firmware 10.51.5
• NIC2: QLogic QL41xxx, driver qedentv version: 3.11.16.0, firmware mfw 8.52.9.0 storm
• OS/Hypervisor: VMware ESXi 6.7.0 build-16075168 (6.7 U3)
1x Physical Switch
• 10Gb switch ports << network bottleneck by purpose, because customer is using 10Gb switch ports as well
Confidential │ ©2020 VMware, Inc. 5
Test Plan and HPE Lab results – 4.x.1
SUCCESS CRITERIA (Expectation)
• At least 5 Gbps (~500 MB/s) throughput
RESULTS
• Intel:
• Bandwidth/Throughput: 5.91 [Gbps]
• Qlogic:
• Bandwidth/Throughput: 9.39 [Gbps]
COMPARISON
• Qlogic outperform Intel by 59%
VMKernel - TCP (iperf) communication between two ESXi host consoles
iperf -s iperf -c
ESXi Host ESXi Host
tcp
TEST METHOD
• ESXi (DCUI/shell):
• esxcli network firewall set --enabled false
• /usr/lib/vmware/vsan/bin/iperf3.copy -s -B [APP-SERVER-01 vMotion IP]
• ESXi (DCUI/shell):
• esxcli network firewall set --enabled false
• /usr/lib/vmware/vsan/bin/iperf3.copy -t 300 -c [APP-SERVER-01 vMotion IP]
Confidential │ ©2020 VMware, Inc. 6
Test Plan and HPE Lab results – 4.x.2
VM - TCP (nuttcp) communication of 2 VMs across two ESXi hosts
SUCCESS CRITERIA (Expectation)
• At least 8 Gbps (~800 MB/s) throughput
• Note: 20 Gbps should be achievable in pure software stack.
RESULTS
• Intel:
• Bandwidth/Throughput: 9.1 [Gbps]
• Qlogic:
• Bandwidth/Throughput: 8.35 [Gbps]
COMPARISON
• Intel outperform Qlogic by 9%
DRS Rule to keep VMs on ESXi Host
App
Server
nuttcp -S
App
Client
nuttcp -t
DRS Rule to keep VMs on ESXi Host
ESXi Host ESXi Host
tcp
TEST METHOD
• 1x App Server:
• nuttcp -S -P 5000 -N 20
• 1x App Client:
• nuttcp -t -N 4 -P 5000 -T 300 APP-SERVER-01
Confidential │ ©2020 VMware, Inc. 7
Test Plan and HPE Lab results – 4.x.3
VM - TCP (iperf) communication of 2 VMs across two ESXi hosts
SUCCESS CRITERIA (Expectation)
• At least 8 Gbps (~800 MB/s) throughput
• Note: 20 Gbps should be achievable in pure software stack.
RESULTS
• Intel:
• Bandwidth/Throughput: 9.39 [Gbps]
• Qlogic:
• Bandwidth/Throughput: 8.43 [Gbps]
COMPARISON
• Intel outperform Qlogic by 11%
DRS Rule to keep VMs on ESXi Host
App
Server
iperf -s
App
Client
iperf -c
DRS Rule to keep VMs on ESXi Host
ESXi Host ESXi Host
tcp
TEST METHOD
• 1x App Server:
• iperf3 –s
• 1x App Client:
• iperf3 -c [APP-SERVER-01] -t 300 -b 25g -P 4
Confidential │ ©2020 VMware, Inc. 8
Test Plan and HPE Lab results – 4.x.4
VM - UDP (nuttcp – 64 KB) communication of 2 VMs across two ESXi hosts
SUCCESS CRITERIA (Expectation)
• At least 8 Gbps (~800 MB/s) throughput
• Note: 8 Gbps should be achievable in pure software stack.
RESULTS
• Intel:
• Bandwidth/Throughput: 9.1 [Gbps]
• Qlogic:
• Bandwidth/Throughput: 8.68 [Gbps]
COMPARISON
• Intel outperform Qlogic by 5%
TEST METHOD
• 1x App Server:
• nuttcp -S -P 5000 -N 20
• 1x App Client:
• nuttcp -u -Ru -l65507 -N 4 -P 5000 -T 300 -i APP-SERVER-01
DRS Rule to keep VMs on ESXi Host
App
Server
nuttcp -S
App
Client
nuttcp -u
DRS Rule to keep VMs on ESXi Host
ESXi Host ESXi Host
udp
Confidential │ ©2020 VMware, Inc. 9
Test Plan and HPE Lab results – 4.x.5
VM - UDP (iperf 64KB) communication of 2 VMs across two ESXi hosts
SUCCESS CRITERIA (Expectation)
• At least 8 Gbps (~800 MB/s) throughput
• Note: 8 Gbps should be achievable in pure software stack.
RESULTS
• Intel:
• Bandwidth/Throughput: 7.91 [Gbps]
• Qlogic:
• Bandwidth/Throughput: 7.91 [Gbps]
COMPARISON
• Equal Intel and Qlogic results
TEST METHOD
• 1x App Server:
• iperf3 –s
• 1x App Client:
• iperf3 -c [APP-SERVER-01] -u -t 300 -b 25g -P 4 -l 65507
udp
Confidential │ ©2020 VMware, Inc. 10
Test Plan and HPE Lab results – 4.x.6
VM - HTTP (nginx) communication of 2 VMs across two ESXi hosts
SUCCESS CRITERIA (Expectation)
• At least 8 Gbps (~800 MB/s) throughput
• Note: 22 Gbps should be achievable on faster physical network
RESULTS
• Intel:
• Bandwidth/Throughput: 5.15 [Gbps]
• Qlogic:
• Bandwidth/Throughput: 5.93 [Gbps]
COMPARISON
• Qlogic outperform Intel by 15%
FINDINGS
• Expectation is not met. It is strange we cannot
achieve at least 8 Gbps for HTTPS traffic.
DRS Rule to keep VMs on ESXi Host
App
Server
NGINX
App
Client
WRK
DRS Rule to keep VMs on ESXi Host
ESXi Host ESXi Host
http
TEST METHOD
• 1x App Client:
• taskset -c 0-8 /root/wrk -t 8 -c 8 -d 300s http://[APP-LB-01]/1M.txt
• 1x App Server:
• NGINX http server serving 1MB file
Confidential │ ©2020 VMware, Inc. 11
Test Plan - HPE Lab results – 4.x.7
VM - HTTPS (nginx) communication of 2 VMs across two ESXi hosts
SUCCESS CRITERIA (Expectation)
• At least 8 Gbps (~800 MB/s) throughput
• Note: 22 Gbps should be achievable in pure software stack. Intel CPU
Instructions AES-NI accelerates SSL, thus encryption penalty should be
mitigated.
RESULTS
• Intel:
• Bandwidth/Throughput: 8.72 [Gbps]
• Qlogic:
• Bandwidth/Throughput: 8.64 [Gbps]
COMPARISON
• Equal Intel and Qlogic results
FINDINGS
• Expectation is met.
• HTTPS overachieved HTTP results on both NICs, why? Because
on Intel QuickAssist Technology & OpenSSL & NGINX. See. LINK
DRS Rule to keep VMs on ESXi Host
App
Server
NGINX
App
Client
WRK
DRS Rule to keep VMs on ESXi Host
ESXi Host ESXi Host
https
TEST METHOD
• 1x App Client:
• taskset -c 0-8 /root/wrk -t 8 -c 8 -d 300s https://[APP-LB-01]/1M.txt
• 1x App Server:
• NGINX http server serving 1MB file
Confidential │ ©2020 VMware, Inc. 12
Test Plan - HPE Lab results – 4.x.8
VM - HTTP communication across two ESXi hosts via LoadBalancer no RSS
SUCCESS CRITERIA (Expectation)
• At least 4 Gbps (~400 MB/s) throughput
• Note: 4 Gbps should be achievable in pure software stack
RESULTS
• Intel:
• Bandwidth/Throughput: 4.97 [Gbps]
• Qlogic:
• Bandwidth/Throughput: 4.73 [Gbps]
COMPARISON
• Intel outperform Qlogic by 5%
FINDINGS
• Expectation is met.
DRS Rule to keep VMs on ESXi Host
App
Client
WRK
App
Server
NGINX
App
Load Balancer
(Reverse Proxy)
NGINX
App
Server
NGINX
App
Server
NGINX
App
Server
NGINX
ESXi Host ESXi Host
DRS Rule to keep VMs on ESXi Host
http
TEST METHOD
• 1x App Client:
• taskset -c 0-8 /root/wrk -t 8 -c 8 -d 300s http://[APP-LB-01]/1M.txt
• 1x App Load balancer:
• NGINX L7 Load Balancer (reverse http proxy) to 4 members
• RSS not enabled in VM settings (default config)
• 4x App Server:
• NGINX http server serving 1MB file
Confidential │ ©2020 VMware, Inc. 13
Test Plan - HPE Lab results – 4.x.9
VM - HTTP communication across two ESXi hosts via LoadBalancer (RSS)
SUCCESS CRITERIA (Expectation)
• At least 4 Gbps (~400 MB/s) throughput
• Note: 4 Gbps should be achievable in pure software stack
RESULTS
• Intel:
• Bandwidth/Throughput: 5.18 [Gbps]
• Qlogic:
• Bandwidth/Throughput: 6.58 [Gbps]
COMPARISON
• Qlogic outperform Intel by 27%
FINDINGS
• Expectation is met.
• RSS has positive impact on performance in Qlogic
• RSS is not supported on Intel NIC driver, no impact
DRS Rule to keep VMs on ESXi Host
App
Client
WRK
App
Server
NGINX
App
Load Balancer
(Reverse Proxy)
NGINX
App
Server
NGINX
App
Server
NGINX
App
Server
NGINX
ESXi Host ESXi Host
DRS Rule to keep VMs on ESXi Host
http
TEST METHOD
• 1x App Client:
• taskset -c 0-8 /root/wrk -t 8 -c 8 -d 300s http://[APP-LB-01]/1M.txt
• 1x App Load balancer:
• NGINX L7 Load Balancer (reverse http proxy) to 4 members
• End-to-end RSS enabled
• 4x App Server:
• NGINX http server serving 1MB file
Confidential │ ©2020 VMware, Inc.
Thank You

More Related Content

What's hot

Introduction to failover clustering with sql server
Introduction to failover clustering with sql serverIntroduction to failover clustering with sql server
Introduction to failover clustering with sql server
Eduardo Castro
 
Introducing Xen Server
Introducing Xen ServerIntroducing Xen Server
Introducing Xen Server
StephenRice86
 
SDN Architecture & Ecosystem
SDN Architecture & EcosystemSDN Architecture & Ecosystem
SDN Architecture & Ecosystem
Kingston Smiler
 
Linux Performance Analysis: New Tools and Old Secrets
Linux Performance Analysis: New Tools and Old SecretsLinux Performance Analysis: New Tools and Old Secrets
Linux Performance Analysis: New Tools and Old Secrets
Brendan Gregg
 
OpenStack Telco Architecture: OpenStack Summit Boston 2017
OpenStack Telco Architecture: OpenStack Summit Boston 2017OpenStack Telco Architecture: OpenStack Summit Boston 2017
OpenStack Telco Architecture: OpenStack Summit Boston 2017
Christian "kiko" Reis
 
VMware vSphere vsan EN.pptx
VMware vSphere vsan EN.pptxVMware vSphere vsan EN.pptx
VMware vSphere vsan EN.pptx
CH431
 
VMware Advance Troubleshooting Workshop - Day 5
VMware Advance Troubleshooting Workshop - Day 5VMware Advance Troubleshooting Workshop - Day 5
VMware Advance Troubleshooting Workshop - Day 5
Vepsun Technologies
 
Volume Encryption In CloudStack
Volume Encryption In CloudStackVolume Encryption In CloudStack
Volume Encryption In CloudStack
ShapeBlue
 
Securing your cloud with Xen's advanced security features
Securing your cloud with Xen's advanced security featuresSecuring your cloud with Xen's advanced security features
Securing your cloud with Xen's advanced security features
The Linux Foundation
 
Cache coloring Xen Summit 2020
Cache coloring Xen Summit 2020Cache coloring Xen Summit 2020
Cache coloring Xen Summit 2020
Stefano Stabellini
 
[232] 성능어디까지쥐어짜봤니 송태웅
[232] 성능어디까지쥐어짜봤니 송태웅[232] 성능어디까지쥐어짜봤니 송태웅
[232] 성능어디까지쥐어짜봤니 송태웅
NAVER D2
 
OpenStack Architecture and Use Cases
OpenStack Architecture and Use CasesOpenStack Architecture and Use Cases
OpenStack Architecture and Use Cases
Jalal Mostafa
 
컨테이너 기술 ( Container Technology ) 발표 자료 다운로드
컨테이너 기술 ( Container Technology ) 발표 자료 다운로드컨테이너 기술 ( Container Technology ) 발표 자료 다운로드
컨테이너 기술 ( Container Technology ) 발표 자료 다운로드
Opennaru, inc.
 
Kubernetes a comprehensive overview
Kubernetes   a comprehensive overviewKubernetes   a comprehensive overview
Kubernetes a comprehensive overview
Gabriel Carro
 
Cloud, sdn and nfv 기술동향 atto-research-박문기-20171016
Cloud, sdn and nfv 기술동향 atto-research-박문기-20171016Cloud, sdn and nfv 기술동향 atto-research-박문기-20171016
Cloud, sdn and nfv 기술동향 atto-research-박문기-20171016
문기 박
 
Virtualization with KVM (Kernel-based Virtual Machine)
Virtualization with KVM (Kernel-based Virtual Machine)Virtualization with KVM (Kernel-based Virtual Machine)
Virtualization with KVM (Kernel-based Virtual Machine)
Novell
 
Neutron packet logging framework
Neutron packet logging frameworkNeutron packet logging framework
Neutron packet logging framework
Vietnam Open Infrastructure User Group
 
OpenvSwitch Deep Dive
OpenvSwitch Deep DiveOpenvSwitch Deep Dive
OpenvSwitch Deep Dive
rajdeep
 
The Next Generation of Hyperconverged Infrastructure - Cisco
The Next Generation of Hyperconverged Infrastructure - CiscoThe Next Generation of Hyperconverged Infrastructure - Cisco
The Next Generation of Hyperconverged Infrastructure - Cisco
MarcoTechnologies
 
Kvm performance optimization for ubuntu
Kvm performance optimization for ubuntuKvm performance optimization for ubuntu
Kvm performance optimization for ubuntuSim Janghoon
 

What's hot (20)

Introduction to failover clustering with sql server
Introduction to failover clustering with sql serverIntroduction to failover clustering with sql server
Introduction to failover clustering with sql server
 
Introducing Xen Server
Introducing Xen ServerIntroducing Xen Server
Introducing Xen Server
 
SDN Architecture & Ecosystem
SDN Architecture & EcosystemSDN Architecture & Ecosystem
SDN Architecture & Ecosystem
 
Linux Performance Analysis: New Tools and Old Secrets
Linux Performance Analysis: New Tools and Old SecretsLinux Performance Analysis: New Tools and Old Secrets
Linux Performance Analysis: New Tools and Old Secrets
 
OpenStack Telco Architecture: OpenStack Summit Boston 2017
OpenStack Telco Architecture: OpenStack Summit Boston 2017OpenStack Telco Architecture: OpenStack Summit Boston 2017
OpenStack Telco Architecture: OpenStack Summit Boston 2017
 
VMware vSphere vsan EN.pptx
VMware vSphere vsan EN.pptxVMware vSphere vsan EN.pptx
VMware vSphere vsan EN.pptx
 
VMware Advance Troubleshooting Workshop - Day 5
VMware Advance Troubleshooting Workshop - Day 5VMware Advance Troubleshooting Workshop - Day 5
VMware Advance Troubleshooting Workshop - Day 5
 
Volume Encryption In CloudStack
Volume Encryption In CloudStackVolume Encryption In CloudStack
Volume Encryption In CloudStack
 
Securing your cloud with Xen's advanced security features
Securing your cloud with Xen's advanced security featuresSecuring your cloud with Xen's advanced security features
Securing your cloud with Xen's advanced security features
 
Cache coloring Xen Summit 2020
Cache coloring Xen Summit 2020Cache coloring Xen Summit 2020
Cache coloring Xen Summit 2020
 
[232] 성능어디까지쥐어짜봤니 송태웅
[232] 성능어디까지쥐어짜봤니 송태웅[232] 성능어디까지쥐어짜봤니 송태웅
[232] 성능어디까지쥐어짜봤니 송태웅
 
OpenStack Architecture and Use Cases
OpenStack Architecture and Use CasesOpenStack Architecture and Use Cases
OpenStack Architecture and Use Cases
 
컨테이너 기술 ( Container Technology ) 발표 자료 다운로드
컨테이너 기술 ( Container Technology ) 발표 자료 다운로드컨테이너 기술 ( Container Technology ) 발표 자료 다운로드
컨테이너 기술 ( Container Technology ) 발표 자료 다운로드
 
Kubernetes a comprehensive overview
Kubernetes   a comprehensive overviewKubernetes   a comprehensive overview
Kubernetes a comprehensive overview
 
Cloud, sdn and nfv 기술동향 atto-research-박문기-20171016
Cloud, sdn and nfv 기술동향 atto-research-박문기-20171016Cloud, sdn and nfv 기술동향 atto-research-박문기-20171016
Cloud, sdn and nfv 기술동향 atto-research-박문기-20171016
 
Virtualization with KVM (Kernel-based Virtual Machine)
Virtualization with KVM (Kernel-based Virtual Machine)Virtualization with KVM (Kernel-based Virtual Machine)
Virtualization with KVM (Kernel-based Virtual Machine)
 
Neutron packet logging framework
Neutron packet logging frameworkNeutron packet logging framework
Neutron packet logging framework
 
OpenvSwitch Deep Dive
OpenvSwitch Deep DiveOpenvSwitch Deep Dive
OpenvSwitch Deep Dive
 
The Next Generation of Hyperconverged Infrastructure - Cisco
The Next Generation of Hyperconverged Infrastructure - CiscoThe Next Generation of Hyperconverged Infrastructure - Cisco
The Next Generation of Hyperconverged Infrastructure - Cisco
 
Kvm performance optimization for ubuntu
Kvm performance optimization for ubuntuKvm performance optimization for ubuntu
Kvm performance optimization for ubuntu
 

Similar to Intel & QLogic NIC performance test results v0.2

Network performance test plan_v0.3
Network performance test plan_v0.3Network performance test plan_v0.3
Network performance test plan_v0.3
David Pasek
 
Deploying flash storage for Ceph without compromising performance
Deploying flash storage for Ceph without compromising performance Deploying flash storage for Ceph without compromising performance
Deploying flash storage for Ceph without compromising performance
Ceph Community
 
DPDK Summit 2015 - HP - Al Sanders
DPDK Summit 2015 - HP - Al SandersDPDK Summit 2015 - HP - Al Sanders
DPDK Summit 2015 - HP - Al Sanders
Jim St. Leger
 
VMworld 2013: Extreme Performance Series: Network Speed Ahead
VMworld 2013: Extreme Performance Series: Network Speed Ahead VMworld 2013: Extreme Performance Series: Network Speed Ahead
VMworld 2013: Extreme Performance Series: Network Speed Ahead
VMworld
 
Boyan Krosnov - Building a software-defined cloud - our experience
Boyan Krosnov - Building a software-defined cloud - our experienceBoyan Krosnov - Building a software-defined cloud - our experience
Boyan Krosnov - Building a software-defined cloud - our experience
ShapeBlue
 
VMworld 2014: Extreme Performance Series
VMworld 2014: Extreme Performance Series VMworld 2014: Extreme Performance Series
VMworld 2014: Extreme Performance Series
VMworld
 
20231211+-+Winning+with+Latency.pdf
20231211+-+Winning+with+Latency.pdf20231211+-+Winning+with+Latency.pdf
20231211+-+Winning+with+Latency.pdf
Nakhoudah
 
Accelerating Cassandra Workloads on Ceph with All-Flash PCIE SSDS
Accelerating Cassandra Workloads on Ceph with All-Flash PCIE SSDSAccelerating Cassandra Workloads on Ceph with All-Flash PCIE SSDS
Accelerating Cassandra Workloads on Ceph with All-Flash PCIE SSDS
Ceph Community
 
Using Agilio SmartNICs for OpenStack Networking Acceleration
Using Agilio SmartNICs for OpenStack Networking AccelerationUsing Agilio SmartNICs for OpenStack Networking Acceleration
Using Agilio SmartNICs for OpenStack Networking Acceleration
Netronome
 
Ceph Day SF 2015 - Deploying flash storage for Ceph without compromising perf...
Ceph Day SF 2015 - Deploying flash storage for Ceph without compromising perf...Ceph Day SF 2015 - Deploying flash storage for Ceph without compromising perf...
Ceph Day SF 2015 - Deploying flash storage for Ceph without compromising perf...
Ceph Community
 
Vsc 71-se-presentation-training
Vsc 71-se-presentation-trainingVsc 71-se-presentation-training
Vsc 71-se-presentation-training
narit_ton
 
Ceph Community Talk on High-Performance Solid Sate Ceph
Ceph Community Talk on High-Performance Solid Sate Ceph Ceph Community Talk on High-Performance Solid Sate Ceph
Ceph Community Talk on High-Performance Solid Sate Ceph
Ceph Community
 
Accelerating Nginx Web Server Performance
Accelerating Nginx Web Server PerformanceAccelerating Nginx Web Server Performance
Accelerating Nginx Web Server Performance
Bruce Tolley
 
F5 Meetup presentation automation 2017
F5 Meetup presentation automation 2017F5 Meetup presentation automation 2017
F5 Meetup presentation automation 2017
Guy Brown
 
Ceph QoS: How to support QoS in distributed storage system - Taewoong Kim
Ceph QoS: How to support QoS in distributed storage system - Taewoong KimCeph QoS: How to support QoS in distributed storage system - Taewoong Kim
Ceph QoS: How to support QoS in distributed storage system - Taewoong Kim
Ceph Community
 
Scaling Kubernetes to Support 50000 Services.pptx
Scaling Kubernetes to Support 50000 Services.pptxScaling Kubernetes to Support 50000 Services.pptx
Scaling Kubernetes to Support 50000 Services.pptx
thaond2
 
What’s New in NGINX Plus R16?
What’s New in NGINX Plus R16?What’s New in NGINX Plus R16?
What’s New in NGINX Plus R16?
NGINX, Inc.
 
VMworld 2013: How SRP Delivers More Than Power to Their Customers
VMworld 2013: How SRP Delivers More Than Power to Their Customers VMworld 2013: How SRP Delivers More Than Power to Their Customers
VMworld 2013: How SRP Delivers More Than Power to Their Customers
VMworld
 
VMworld 2013: How to Replace Websphere Application Server (WAS) with TCserver
VMworld 2013: How to Replace Websphere Application Server (WAS) with TCserver VMworld 2013: How to Replace Websphere Application Server (WAS) with TCserver
VMworld 2013: How to Replace Websphere Application Server (WAS) with TCserver
VMworld
 
Janet-hosted test tools
Janet-hosted test toolsJanet-hosted test tools
Janet-hosted test tools
Jisc
 

Similar to Intel & QLogic NIC performance test results v0.2 (20)

Network performance test plan_v0.3
Network performance test plan_v0.3Network performance test plan_v0.3
Network performance test plan_v0.3
 
Deploying flash storage for Ceph without compromising performance
Deploying flash storage for Ceph without compromising performance Deploying flash storage for Ceph without compromising performance
Deploying flash storage for Ceph without compromising performance
 
DPDK Summit 2015 - HP - Al Sanders
DPDK Summit 2015 - HP - Al SandersDPDK Summit 2015 - HP - Al Sanders
DPDK Summit 2015 - HP - Al Sanders
 
VMworld 2013: Extreme Performance Series: Network Speed Ahead
VMworld 2013: Extreme Performance Series: Network Speed Ahead VMworld 2013: Extreme Performance Series: Network Speed Ahead
VMworld 2013: Extreme Performance Series: Network Speed Ahead
 
Boyan Krosnov - Building a software-defined cloud - our experience
Boyan Krosnov - Building a software-defined cloud - our experienceBoyan Krosnov - Building a software-defined cloud - our experience
Boyan Krosnov - Building a software-defined cloud - our experience
 
VMworld 2014: Extreme Performance Series
VMworld 2014: Extreme Performance Series VMworld 2014: Extreme Performance Series
VMworld 2014: Extreme Performance Series
 
20231211+-+Winning+with+Latency.pdf
20231211+-+Winning+with+Latency.pdf20231211+-+Winning+with+Latency.pdf
20231211+-+Winning+with+Latency.pdf
 
Accelerating Cassandra Workloads on Ceph with All-Flash PCIE SSDS
Accelerating Cassandra Workloads on Ceph with All-Flash PCIE SSDSAccelerating Cassandra Workloads on Ceph with All-Flash PCIE SSDS
Accelerating Cassandra Workloads on Ceph with All-Flash PCIE SSDS
 
Using Agilio SmartNICs for OpenStack Networking Acceleration
Using Agilio SmartNICs for OpenStack Networking AccelerationUsing Agilio SmartNICs for OpenStack Networking Acceleration
Using Agilio SmartNICs for OpenStack Networking Acceleration
 
Ceph Day SF 2015 - Deploying flash storage for Ceph without compromising perf...
Ceph Day SF 2015 - Deploying flash storage for Ceph without compromising perf...Ceph Day SF 2015 - Deploying flash storage for Ceph without compromising perf...
Ceph Day SF 2015 - Deploying flash storage for Ceph without compromising perf...
 
Vsc 71-se-presentation-training
Vsc 71-se-presentation-trainingVsc 71-se-presentation-training
Vsc 71-se-presentation-training
 
Ceph Community Talk on High-Performance Solid Sate Ceph
Ceph Community Talk on High-Performance Solid Sate Ceph Ceph Community Talk on High-Performance Solid Sate Ceph
Ceph Community Talk on High-Performance Solid Sate Ceph
 
Accelerating Nginx Web Server Performance
Accelerating Nginx Web Server PerformanceAccelerating Nginx Web Server Performance
Accelerating Nginx Web Server Performance
 
F5 Meetup presentation automation 2017
F5 Meetup presentation automation 2017F5 Meetup presentation automation 2017
F5 Meetup presentation automation 2017
 
Ceph QoS: How to support QoS in distributed storage system - Taewoong Kim
Ceph QoS: How to support QoS in distributed storage system - Taewoong KimCeph QoS: How to support QoS in distributed storage system - Taewoong Kim
Ceph QoS: How to support QoS in distributed storage system - Taewoong Kim
 
Scaling Kubernetes to Support 50000 Services.pptx
Scaling Kubernetes to Support 50000 Services.pptxScaling Kubernetes to Support 50000 Services.pptx
Scaling Kubernetes to Support 50000 Services.pptx
 
What’s New in NGINX Plus R16?
What’s New in NGINX Plus R16?What’s New in NGINX Plus R16?
What’s New in NGINX Plus R16?
 
VMworld 2013: How SRP Delivers More Than Power to Their Customers
VMworld 2013: How SRP Delivers More Than Power to Their Customers VMworld 2013: How SRP Delivers More Than Power to Their Customers
VMworld 2013: How SRP Delivers More Than Power to Their Customers
 
VMworld 2013: How to Replace Websphere Application Server (WAS) with TCserver
VMworld 2013: How to Replace Websphere Application Server (WAS) with TCserver VMworld 2013: How to Replace Websphere Application Server (WAS) with TCserver
VMworld 2013: How to Replace Websphere Application Server (WAS) with TCserver
 
Janet-hosted test tools
Janet-hosted test toolsJanet-hosted test tools
Janet-hosted test tools
 

More from David Pasek

FlexBook Software - Conceptual Architecture
FlexBook Software - Conceptual ArchitectureFlexBook Software - Conceptual Architecture
FlexBook Software - Conceptual Architecture
David Pasek
 
Flex Cloud - Conceptual Design - ver 0.2
Flex Cloud - Conceptual Design - ver 0.2Flex Cloud - Conceptual Design - ver 0.2
Flex Cloud - Conceptual Design - ver 0.2
David Pasek
 
E tourism v oblasti cestovního ruchu
E tourism v oblasti cestovního ruchuE tourism v oblasti cestovního ruchu
E tourism v oblasti cestovního ruchu
David Pasek
 
Architektura a implementace digitálních knihoven v prostředí sítě Internet
Architektura a implementace digitálních knihoven v prostředí sítě InternetArchitektura a implementace digitálních knihoven v prostředí sítě Internet
Architektura a implementace digitálních knihoven v prostředí sítě Internet
David Pasek
 
Exchange office 3.0 - Stanovisko Státní banky československé
Exchange office 3.0 - Stanovisko Státní banky československéExchange office 3.0 - Stanovisko Státní banky československé
Exchange office 3.0 - Stanovisko Státní banky československé
David Pasek
 
vSAN architecture components
vSAN architecture componentsvSAN architecture components
vSAN architecture components
David Pasek
 
FlexBook overview - v2.4
FlexBook overview - v2.4FlexBook overview - v2.4
FlexBook overview - v2.4
David Pasek
 
VMware HCI solutions - 2020-01-16
VMware HCI solutions - 2020-01-16VMware HCI solutions - 2020-01-16
VMware HCI solutions - 2020-01-16
David Pasek
 
Hybrid cloud overview and VCF on VxRAIL
Hybrid cloud overview and VCF on VxRAILHybrid cloud overview and VCF on VxRAIL
Hybrid cloud overview and VCF on VxRAIL
David Pasek
 
Private IaaS Cloud Provider
Private IaaS Cloud ProviderPrivate IaaS Cloud Provider
Private IaaS Cloud Provider
David Pasek
 
Spectre/Meltdown security vulnerabilities FAQ
Spectre/Meltdown security vulnerabilities FAQSpectre/Meltdown security vulnerabilities FAQ
Spectre/Meltdown security vulnerabilities FAQ
David Pasek
 
FlexBook Basic Overview - v2.0
FlexBook Basic Overview - v2.0FlexBook Basic Overview - v2.0
FlexBook Basic Overview - v2.0
David Pasek
 
Spectre meltdown performance_tests - v0.3
Spectre meltdown performance_tests - v0.3Spectre meltdown performance_tests - v0.3
Spectre meltdown performance_tests - v0.3
David Pasek
 
FlexBook basic overview v2.0
FlexBook basic overview v2.0FlexBook basic overview v2.0
FlexBook basic overview v2.0
David Pasek
 
FlexBook - reservation system basic overview v1.1
FlexBook - reservation system basic overview v1.1FlexBook - reservation system basic overview v1.1
FlexBook - reservation system basic overview v1.1
David Pasek
 
CLI for VMware Distributed Switch (Community project)
CLI for VMware Distributed Switch (Community project)CLI for VMware Distributed Switch (Community project)
CLI for VMware Distributed Switch (Community project)
David Pasek
 
Dell VLT reference architecture v2 0
Dell VLT reference architecture v2 0Dell VLT reference architecture v2 0
Dell VLT reference architecture v2 0
David Pasek
 
Metro Cluster High Availability or SRM Disaster Recovery?
Metro Cluster High Availability or SRM Disaster Recovery?Metro Cluster High Availability or SRM Disaster Recovery?
Metro Cluster High Availability or SRM Disaster Recovery?
David Pasek
 
Rezervační systém Flexbook - stručný přehled v.0.8
Rezervační systém Flexbook - stručný přehled v.0.8Rezervační systém Flexbook - stručný přehled v.0.8
Rezervační systém Flexbook - stručný přehled v.0.8
David Pasek
 
Creating content packs in VMware LogInsight
Creating content packs in VMware LogInsightCreating content packs in VMware LogInsight
Creating content packs in VMware LogInsight
David Pasek
 

More from David Pasek (20)

FlexBook Software - Conceptual Architecture
FlexBook Software - Conceptual ArchitectureFlexBook Software - Conceptual Architecture
FlexBook Software - Conceptual Architecture
 
Flex Cloud - Conceptual Design - ver 0.2
Flex Cloud - Conceptual Design - ver 0.2Flex Cloud - Conceptual Design - ver 0.2
Flex Cloud - Conceptual Design - ver 0.2
 
E tourism v oblasti cestovního ruchu
E tourism v oblasti cestovního ruchuE tourism v oblasti cestovního ruchu
E tourism v oblasti cestovního ruchu
 
Architektura a implementace digitálních knihoven v prostředí sítě Internet
Architektura a implementace digitálních knihoven v prostředí sítě InternetArchitektura a implementace digitálních knihoven v prostředí sítě Internet
Architektura a implementace digitálních knihoven v prostředí sítě Internet
 
Exchange office 3.0 - Stanovisko Státní banky československé
Exchange office 3.0 - Stanovisko Státní banky československéExchange office 3.0 - Stanovisko Státní banky československé
Exchange office 3.0 - Stanovisko Státní banky československé
 
vSAN architecture components
vSAN architecture componentsvSAN architecture components
vSAN architecture components
 
FlexBook overview - v2.4
FlexBook overview - v2.4FlexBook overview - v2.4
FlexBook overview - v2.4
 
VMware HCI solutions - 2020-01-16
VMware HCI solutions - 2020-01-16VMware HCI solutions - 2020-01-16
VMware HCI solutions - 2020-01-16
 
Hybrid cloud overview and VCF on VxRAIL
Hybrid cloud overview and VCF on VxRAILHybrid cloud overview and VCF on VxRAIL
Hybrid cloud overview and VCF on VxRAIL
 
Private IaaS Cloud Provider
Private IaaS Cloud ProviderPrivate IaaS Cloud Provider
Private IaaS Cloud Provider
 
Spectre/Meltdown security vulnerabilities FAQ
Spectre/Meltdown security vulnerabilities FAQSpectre/Meltdown security vulnerabilities FAQ
Spectre/Meltdown security vulnerabilities FAQ
 
FlexBook Basic Overview - v2.0
FlexBook Basic Overview - v2.0FlexBook Basic Overview - v2.0
FlexBook Basic Overview - v2.0
 
Spectre meltdown performance_tests - v0.3
Spectre meltdown performance_tests - v0.3Spectre meltdown performance_tests - v0.3
Spectre meltdown performance_tests - v0.3
 
FlexBook basic overview v2.0
FlexBook basic overview v2.0FlexBook basic overview v2.0
FlexBook basic overview v2.0
 
FlexBook - reservation system basic overview v1.1
FlexBook - reservation system basic overview v1.1FlexBook - reservation system basic overview v1.1
FlexBook - reservation system basic overview v1.1
 
CLI for VMware Distributed Switch (Community project)
CLI for VMware Distributed Switch (Community project)CLI for VMware Distributed Switch (Community project)
CLI for VMware Distributed Switch (Community project)
 
Dell VLT reference architecture v2 0
Dell VLT reference architecture v2 0Dell VLT reference architecture v2 0
Dell VLT reference architecture v2 0
 
Metro Cluster High Availability or SRM Disaster Recovery?
Metro Cluster High Availability or SRM Disaster Recovery?Metro Cluster High Availability or SRM Disaster Recovery?
Metro Cluster High Availability or SRM Disaster Recovery?
 
Rezervační systém Flexbook - stručný přehled v.0.8
Rezervační systém Flexbook - stručný přehled v.0.8Rezervační systém Flexbook - stručný přehled v.0.8
Rezervační systém Flexbook - stručný přehled v.0.8
 
Creating content packs in VMware LogInsight
Creating content packs in VMware LogInsightCreating content packs in VMware LogInsight
Creating content packs in VMware LogInsight
 

Recently uploaded

A tale of scale & speed: How the US Navy is enabling software delivery from l...
A tale of scale & speed: How the US Navy is enabling software delivery from l...A tale of scale & speed: How the US Navy is enabling software delivery from l...
A tale of scale & speed: How the US Navy is enabling software delivery from l...
sonjaschweigert1
 
Why You Should Replace Windows 11 with Nitrux Linux 3.5.0 for enhanced perfor...
Why You Should Replace Windows 11 with Nitrux Linux 3.5.0 for enhanced perfor...Why You Should Replace Windows 11 with Nitrux Linux 3.5.0 for enhanced perfor...
Why You Should Replace Windows 11 with Nitrux Linux 3.5.0 for enhanced perfor...
SOFTTECHHUB
 
Microsoft - Power Platform_G.Aspiotis.pdf
Microsoft - Power Platform_G.Aspiotis.pdfMicrosoft - Power Platform_G.Aspiotis.pdf
Microsoft - Power Platform_G.Aspiotis.pdf
Uni Systems S.M.S.A.
 
GraphSummit Singapore | Graphing Success: Revolutionising Organisational Stru...
GraphSummit Singapore | Graphing Success: Revolutionising Organisational Stru...GraphSummit Singapore | Graphing Success: Revolutionising Organisational Stru...
GraphSummit Singapore | Graphing Success: Revolutionising Organisational Stru...
Neo4j
 
Securing your Kubernetes cluster_ a step-by-step guide to success !
Securing your Kubernetes cluster_ a step-by-step guide to success !Securing your Kubernetes cluster_ a step-by-step guide to success !
Securing your Kubernetes cluster_ a step-by-step guide to success !
KatiaHIMEUR1
 
Large Language Model (LLM) and it’s Geospatial Applications
Large Language Model (LLM) and it’s Geospatial ApplicationsLarge Language Model (LLM) and it’s Geospatial Applications
Large Language Model (LLM) and it’s Geospatial Applications
Rohit Gautam
 
GraphSummit Singapore | The Future of Agility: Supercharging Digital Transfor...
GraphSummit Singapore | The Future of Agility: Supercharging Digital Transfor...GraphSummit Singapore | The Future of Agility: Supercharging Digital Transfor...
GraphSummit Singapore | The Future of Agility: Supercharging Digital Transfor...
Neo4j
 
20240605 QFM017 Machine Intelligence Reading List May 2024
20240605 QFM017 Machine Intelligence Reading List May 202420240605 QFM017 Machine Intelligence Reading List May 2024
20240605 QFM017 Machine Intelligence Reading List May 2024
Matthew Sinclair
 
FIDO Alliance Osaka Seminar: The WebAuthn API and Discoverable Credentials.pdf
FIDO Alliance Osaka Seminar: The WebAuthn API and Discoverable Credentials.pdfFIDO Alliance Osaka Seminar: The WebAuthn API and Discoverable Credentials.pdf
FIDO Alliance Osaka Seminar: The WebAuthn API and Discoverable Credentials.pdf
FIDO Alliance
 
How to Get CNIC Information System with Paksim Ga.pptx
How to Get CNIC Information System with Paksim Ga.pptxHow to Get CNIC Information System with Paksim Ga.pptx
How to Get CNIC Information System with Paksim Ga.pptx
danishmna97
 
Artificial Intelligence for XMLDevelopment
Artificial Intelligence for XMLDevelopmentArtificial Intelligence for XMLDevelopment
Artificial Intelligence for XMLDevelopment
Octavian Nadolu
 
Video Streaming: Then, Now, and in the Future
Video Streaming: Then, Now, and in the FutureVideo Streaming: Then, Now, and in the Future
Video Streaming: Then, Now, and in the Future
Alpen-Adria-Universität
 
Pushing the limits of ePRTC: 100ns holdover for 100 days
Pushing the limits of ePRTC: 100ns holdover for 100 daysPushing the limits of ePRTC: 100ns holdover for 100 days
Pushing the limits of ePRTC: 100ns holdover for 100 days
Adtran
 
Introduction to CHERI technology - Cybersecurity
Introduction to CHERI technology - CybersecurityIntroduction to CHERI technology - Cybersecurity
Introduction to CHERI technology - Cybersecurity
mikeeftimakis1
 
UiPath Test Automation using UiPath Test Suite series, part 5
UiPath Test Automation using UiPath Test Suite series, part 5UiPath Test Automation using UiPath Test Suite series, part 5
UiPath Test Automation using UiPath Test Suite series, part 5
DianaGray10
 
Secstrike : Reverse Engineering & Pwnable tools for CTF.pptx
Secstrike : Reverse Engineering & Pwnable tools for CTF.pptxSecstrike : Reverse Engineering & Pwnable tools for CTF.pptx
Secstrike : Reverse Engineering & Pwnable tools for CTF.pptx
nkrafacyberclub
 
20240609 QFM020 Irresponsible AI Reading List May 2024
20240609 QFM020 Irresponsible AI Reading List May 202420240609 QFM020 Irresponsible AI Reading List May 2024
20240609 QFM020 Irresponsible AI Reading List May 2024
Matthew Sinclair
 
Removing Uninteresting Bytes in Software Fuzzing
Removing Uninteresting Bytes in Software FuzzingRemoving Uninteresting Bytes in Software Fuzzing
Removing Uninteresting Bytes in Software Fuzzing
Aftab Hussain
 
GridMate - End to end testing is a critical piece to ensure quality and avoid...
GridMate - End to end testing is a critical piece to ensure quality and avoid...GridMate - End to end testing is a critical piece to ensure quality and avoid...
GridMate - End to end testing is a critical piece to ensure quality and avoid...
ThomasParaiso2
 
GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...
GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...
GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...
James Anderson
 

Recently uploaded (20)

A tale of scale & speed: How the US Navy is enabling software delivery from l...
A tale of scale & speed: How the US Navy is enabling software delivery from l...A tale of scale & speed: How the US Navy is enabling software delivery from l...
A tale of scale & speed: How the US Navy is enabling software delivery from l...
 
Why You Should Replace Windows 11 with Nitrux Linux 3.5.0 for enhanced perfor...
Why You Should Replace Windows 11 with Nitrux Linux 3.5.0 for enhanced perfor...Why You Should Replace Windows 11 with Nitrux Linux 3.5.0 for enhanced perfor...
Why You Should Replace Windows 11 with Nitrux Linux 3.5.0 for enhanced perfor...
 
Microsoft - Power Platform_G.Aspiotis.pdf
Microsoft - Power Platform_G.Aspiotis.pdfMicrosoft - Power Platform_G.Aspiotis.pdf
Microsoft - Power Platform_G.Aspiotis.pdf
 
GraphSummit Singapore | Graphing Success: Revolutionising Organisational Stru...
GraphSummit Singapore | Graphing Success: Revolutionising Organisational Stru...GraphSummit Singapore | Graphing Success: Revolutionising Organisational Stru...
GraphSummit Singapore | Graphing Success: Revolutionising Organisational Stru...
 
Securing your Kubernetes cluster_ a step-by-step guide to success !
Securing your Kubernetes cluster_ a step-by-step guide to success !Securing your Kubernetes cluster_ a step-by-step guide to success !
Securing your Kubernetes cluster_ a step-by-step guide to success !
 
Large Language Model (LLM) and it’s Geospatial Applications
Large Language Model (LLM) and it’s Geospatial ApplicationsLarge Language Model (LLM) and it’s Geospatial Applications
Large Language Model (LLM) and it’s Geospatial Applications
 
GraphSummit Singapore | The Future of Agility: Supercharging Digital Transfor...
GraphSummit Singapore | The Future of Agility: Supercharging Digital Transfor...GraphSummit Singapore | The Future of Agility: Supercharging Digital Transfor...
GraphSummit Singapore | The Future of Agility: Supercharging Digital Transfor...
 
20240605 QFM017 Machine Intelligence Reading List May 2024
20240605 QFM017 Machine Intelligence Reading List May 202420240605 QFM017 Machine Intelligence Reading List May 2024
20240605 QFM017 Machine Intelligence Reading List May 2024
 
FIDO Alliance Osaka Seminar: The WebAuthn API and Discoverable Credentials.pdf
FIDO Alliance Osaka Seminar: The WebAuthn API and Discoverable Credentials.pdfFIDO Alliance Osaka Seminar: The WebAuthn API and Discoverable Credentials.pdf
FIDO Alliance Osaka Seminar: The WebAuthn API and Discoverable Credentials.pdf
 
How to Get CNIC Information System with Paksim Ga.pptx
How to Get CNIC Information System with Paksim Ga.pptxHow to Get CNIC Information System with Paksim Ga.pptx
How to Get CNIC Information System with Paksim Ga.pptx
 
Artificial Intelligence for XMLDevelopment
Artificial Intelligence for XMLDevelopmentArtificial Intelligence for XMLDevelopment
Artificial Intelligence for XMLDevelopment
 
Video Streaming: Then, Now, and in the Future
Video Streaming: Then, Now, and in the FutureVideo Streaming: Then, Now, and in the Future
Video Streaming: Then, Now, and in the Future
 
Pushing the limits of ePRTC: 100ns holdover for 100 days
Pushing the limits of ePRTC: 100ns holdover for 100 daysPushing the limits of ePRTC: 100ns holdover for 100 days
Pushing the limits of ePRTC: 100ns holdover for 100 days
 
Introduction to CHERI technology - Cybersecurity
Introduction to CHERI technology - CybersecurityIntroduction to CHERI technology - Cybersecurity
Introduction to CHERI technology - Cybersecurity
 
UiPath Test Automation using UiPath Test Suite series, part 5
UiPath Test Automation using UiPath Test Suite series, part 5UiPath Test Automation using UiPath Test Suite series, part 5
UiPath Test Automation using UiPath Test Suite series, part 5
 
Secstrike : Reverse Engineering & Pwnable tools for CTF.pptx
Secstrike : Reverse Engineering & Pwnable tools for CTF.pptxSecstrike : Reverse Engineering & Pwnable tools for CTF.pptx
Secstrike : Reverse Engineering & Pwnable tools for CTF.pptx
 
20240609 QFM020 Irresponsible AI Reading List May 2024
20240609 QFM020 Irresponsible AI Reading List May 202420240609 QFM020 Irresponsible AI Reading List May 2024
20240609 QFM020 Irresponsible AI Reading List May 2024
 
Removing Uninteresting Bytes in Software Fuzzing
Removing Uninteresting Bytes in Software FuzzingRemoving Uninteresting Bytes in Software Fuzzing
Removing Uninteresting Bytes in Software Fuzzing
 
GridMate - End to end testing is a critical piece to ensure quality and avoid...
GridMate - End to end testing is a critical piece to ensure quality and avoid...GridMate - End to end testing is a critical piece to ensure quality and avoid...
GridMate - End to end testing is a critical piece to ensure quality and avoid...
 
GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...
GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...
GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...
 

Intel & QLogic NIC performance test results v0.2

  • 1. Confidential │ ©2020 VMware, Inc. Marvell – Qlogic NIC FastLinQ QL41xxx driver issue David Pasek, dpasek@vmware.com VMware TAM 2020, Nov 25
  • 2. Confidential │ ©2020 VMware, Inc. 2 Problem description - symptoms Network throughput issue in Qlogic FastLinQ QL41xxx • Containerized application running in OpenShift on top of vSphere has in some circumstances 3x higher network throughput on Intel X710 (1.9.5) in comparison to Qlogic FastLinQ QL41xxx (3.11.16.0) with reasonable latency • The issue was observed in production environment – the problem was mitigated by replacing Qlogic by Intel NIC • VM on ESXi with Intel NIC can achieve total throughput 300 MB/s (~3Gbps) with application latency 20 ms • VM on ESXi with Qlogic NIC can achieve total throughput 100 MB/s (~1Gbps) with application latency 2000 ms during the issue • The question is why the same VM running on the identical server hardware (except NIC), can easily handle 300 MB/s (~ 3 Gbps) on Intel NIC and only 100 MB/s (~ 1 GBps) on Qlogic NIC?
  • 3. Confidential │ ©2020 VMware, Inc. Test results from HPE Lab Full test plan with test methodology and network throughput results is available on demand
  • 4. Confidential │ ©2020 VMware, Inc. 4 HPE Lab Lab environment details 2x ESXi hosts • Server Platform: HPE ProLiant DL560 Gen10 • BIOS: U34 | Date (ISO-8601): 2020-04-08 • NIC1: Intel X710, driver i40en version: 1.9.5, firmware 10.51.5 • NIC2: QLogic QL41xxx, driver qedentv version: 3.11.16.0, firmware mfw 8.52.9.0 storm • OS/Hypervisor: VMware ESXi 6.7.0 build-16075168 (6.7 U3) 1x Physical Switch • 10Gb switch ports << network bottleneck by purpose, because customer is using 10Gb switch ports as well
  • 5. Confidential │ ©2020 VMware, Inc. 5 Test Plan and HPE Lab results – 4.x.1 SUCCESS CRITERIA (Expectation) • At least 5 Gbps (~500 MB/s) throughput RESULTS • Intel: • Bandwidth/Throughput: 5.91 [Gbps] • Qlogic: • Bandwidth/Throughput: 9.39 [Gbps] COMPARISON • Qlogic outperform Intel by 59% VMKernel - TCP (iperf) communication between two ESXi host consoles iperf -s iperf -c ESXi Host ESXi Host tcp TEST METHOD • ESXi (DCUI/shell): • esxcli network firewall set --enabled false • /usr/lib/vmware/vsan/bin/iperf3.copy -s -B [APP-SERVER-01 vMotion IP] • ESXi (DCUI/shell): • esxcli network firewall set --enabled false • /usr/lib/vmware/vsan/bin/iperf3.copy -t 300 -c [APP-SERVER-01 vMotion IP]
  • 6. Confidential │ ©2020 VMware, Inc. 6 Test Plan and HPE Lab results – 4.x.2 VM - TCP (nuttcp) communication of 2 VMs across two ESXi hosts SUCCESS CRITERIA (Expectation) • At least 8 Gbps (~800 MB/s) throughput • Note: 20 Gbps should be achievable in pure software stack. RESULTS • Intel: • Bandwidth/Throughput: 9.1 [Gbps] • Qlogic: • Bandwidth/Throughput: 8.35 [Gbps] COMPARISON • Intel outperform Qlogic by 9% DRS Rule to keep VMs on ESXi Host App Server nuttcp -S App Client nuttcp -t DRS Rule to keep VMs on ESXi Host ESXi Host ESXi Host tcp TEST METHOD • 1x App Server: • nuttcp -S -P 5000 -N 20 • 1x App Client: • nuttcp -t -N 4 -P 5000 -T 300 APP-SERVER-01
  • 7. Confidential │ ©2020 VMware, Inc. 7 Test Plan and HPE Lab results – 4.x.3 VM - TCP (iperf) communication of 2 VMs across two ESXi hosts SUCCESS CRITERIA (Expectation) • At least 8 Gbps (~800 MB/s) throughput • Note: 20 Gbps should be achievable in pure software stack. RESULTS • Intel: • Bandwidth/Throughput: 9.39 [Gbps] • Qlogic: • Bandwidth/Throughput: 8.43 [Gbps] COMPARISON • Intel outperform Qlogic by 11% DRS Rule to keep VMs on ESXi Host App Server iperf -s App Client iperf -c DRS Rule to keep VMs on ESXi Host ESXi Host ESXi Host tcp TEST METHOD • 1x App Server: • iperf3 –s • 1x App Client: • iperf3 -c [APP-SERVER-01] -t 300 -b 25g -P 4
  • 8. Confidential │ ©2020 VMware, Inc. 8 Test Plan and HPE Lab results – 4.x.4 VM - UDP (nuttcp – 64 KB) communication of 2 VMs across two ESXi hosts SUCCESS CRITERIA (Expectation) • At least 8 Gbps (~800 MB/s) throughput • Note: 8 Gbps should be achievable in pure software stack. RESULTS • Intel: • Bandwidth/Throughput: 9.1 [Gbps] • Qlogic: • Bandwidth/Throughput: 8.68 [Gbps] COMPARISON • Intel outperform Qlogic by 5% TEST METHOD • 1x App Server: • nuttcp -S -P 5000 -N 20 • 1x App Client: • nuttcp -u -Ru -l65507 -N 4 -P 5000 -T 300 -i APP-SERVER-01 DRS Rule to keep VMs on ESXi Host App Server nuttcp -S App Client nuttcp -u DRS Rule to keep VMs on ESXi Host ESXi Host ESXi Host udp
  • 9. Confidential │ ©2020 VMware, Inc. 9 Test Plan and HPE Lab results – 4.x.5 VM - UDP (iperf 64KB) communication of 2 VMs across two ESXi hosts SUCCESS CRITERIA (Expectation) • At least 8 Gbps (~800 MB/s) throughput • Note: 8 Gbps should be achievable in pure software stack. RESULTS • Intel: • Bandwidth/Throughput: 7.91 [Gbps] • Qlogic: • Bandwidth/Throughput: 7.91 [Gbps] COMPARISON • Equal Intel and Qlogic results TEST METHOD • 1x App Server: • iperf3 –s • 1x App Client: • iperf3 -c [APP-SERVER-01] -u -t 300 -b 25g -P 4 -l 65507 udp
  • 10. Confidential │ ©2020 VMware, Inc. 10 Test Plan and HPE Lab results – 4.x.6 VM - HTTP (nginx) communication of 2 VMs across two ESXi hosts SUCCESS CRITERIA (Expectation) • At least 8 Gbps (~800 MB/s) throughput • Note: 22 Gbps should be achievable on faster physical network RESULTS • Intel: • Bandwidth/Throughput: 5.15 [Gbps] • Qlogic: • Bandwidth/Throughput: 5.93 [Gbps] COMPARISON • Qlogic outperform Intel by 15% FINDINGS • Expectation is not met. It is strange we cannot achieve at least 8 Gbps for HTTPS traffic. DRS Rule to keep VMs on ESXi Host App Server NGINX App Client WRK DRS Rule to keep VMs on ESXi Host ESXi Host ESXi Host http TEST METHOD • 1x App Client: • taskset -c 0-8 /root/wrk -t 8 -c 8 -d 300s http://[APP-LB-01]/1M.txt • 1x App Server: • NGINX http server serving 1MB file
  • 11. Confidential │ ©2020 VMware, Inc. 11 Test Plan - HPE Lab results – 4.x.7 VM - HTTPS (nginx) communication of 2 VMs across two ESXi hosts SUCCESS CRITERIA (Expectation) • At least 8 Gbps (~800 MB/s) throughput • Note: 22 Gbps should be achievable in pure software stack. Intel CPU Instructions AES-NI accelerates SSL, thus encryption penalty should be mitigated. RESULTS • Intel: • Bandwidth/Throughput: 8.72 [Gbps] • Qlogic: • Bandwidth/Throughput: 8.64 [Gbps] COMPARISON • Equal Intel and Qlogic results FINDINGS • Expectation is met. • HTTPS overachieved HTTP results on both NICs, why? Because on Intel QuickAssist Technology & OpenSSL & NGINX. See. LINK DRS Rule to keep VMs on ESXi Host App Server NGINX App Client WRK DRS Rule to keep VMs on ESXi Host ESXi Host ESXi Host https TEST METHOD • 1x App Client: • taskset -c 0-8 /root/wrk -t 8 -c 8 -d 300s https://[APP-LB-01]/1M.txt • 1x App Server: • NGINX http server serving 1MB file
  • 12. Confidential │ ©2020 VMware, Inc. 12 Test Plan - HPE Lab results – 4.x.8 VM - HTTP communication across two ESXi hosts via LoadBalancer no RSS SUCCESS CRITERIA (Expectation) • At least 4 Gbps (~400 MB/s) throughput • Note: 4 Gbps should be achievable in pure software stack RESULTS • Intel: • Bandwidth/Throughput: 4.97 [Gbps] • Qlogic: • Bandwidth/Throughput: 4.73 [Gbps] COMPARISON • Intel outperform Qlogic by 5% FINDINGS • Expectation is met. DRS Rule to keep VMs on ESXi Host App Client WRK App Server NGINX App Load Balancer (Reverse Proxy) NGINX App Server NGINX App Server NGINX App Server NGINX ESXi Host ESXi Host DRS Rule to keep VMs on ESXi Host http TEST METHOD • 1x App Client: • taskset -c 0-8 /root/wrk -t 8 -c 8 -d 300s http://[APP-LB-01]/1M.txt • 1x App Load balancer: • NGINX L7 Load Balancer (reverse http proxy) to 4 members • RSS not enabled in VM settings (default config) • 4x App Server: • NGINX http server serving 1MB file
  • 13. Confidential │ ©2020 VMware, Inc. 13 Test Plan - HPE Lab results – 4.x.9 VM - HTTP communication across two ESXi hosts via LoadBalancer (RSS) SUCCESS CRITERIA (Expectation) • At least 4 Gbps (~400 MB/s) throughput • Note: 4 Gbps should be achievable in pure software stack RESULTS • Intel: • Bandwidth/Throughput: 5.18 [Gbps] • Qlogic: • Bandwidth/Throughput: 6.58 [Gbps] COMPARISON • Qlogic outperform Intel by 27% FINDINGS • Expectation is met. • RSS has positive impact on performance in Qlogic • RSS is not supported on Intel NIC driver, no impact DRS Rule to keep VMs on ESXi Host App Client WRK App Server NGINX App Load Balancer (Reverse Proxy) NGINX App Server NGINX App Server NGINX App Server NGINX ESXi Host ESXi Host DRS Rule to keep VMs on ESXi Host http TEST METHOD • 1x App Client: • taskset -c 0-8 /root/wrk -t 8 -c 8 -d 300s http://[APP-LB-01]/1M.txt • 1x App Load balancer: • NGINX L7 Load Balancer (reverse http proxy) to 4 members • End-to-end RSS enabled • 4x App Server: • NGINX http server serving 1MB file
  • 14. Confidential │ ©2020 VMware, Inc. Thank You