14 pages

English

IPv4 Benchmark Report

Istey - Prchandr

Le téléchargement nécessite un accès à la bibliothèque YouScribe
Tout savoir sur nos offres

14 pages

English

Le téléchargement nécessite un accès à la bibliothèque YouScribe
Tout savoir sur nos offres

A propos
Informations
Extrait

Description

Network Processing Forum Benchmarking Working Group Template for IPv4 Forwarding Application Level Benchmark Report Revision 1.0 Editors: Richard Anderson Jr., IBM Corporation, rbanders@us.ibm.com Prashant Chandra, Intel Corporation, prashant.chandra@intel.com Copyright © 2002 The Network Processing Forum (NPF). All Rights Reserved. This document and translations of it may be copied and furnished to others, and derivative works that comment on or otherwise explain it or assist in its implementation may be prepared, copied, published and distributed, in whole or in part, without restriction other than the following, (1) the above copyright notice and this paragraph must be included on all such copies and derivative works, and (2) this document itself may not be modified in any way, such as by removing the copyright notice or references to the NPF, except as needed for the purpose of developing NPF Implementation Agreements. By downloading, copying, or using this document in any manner, the user consents to the terms and conditions of this notice. Unless the terms and conditions of this notice are breached by the user, the limited permissions granted above are perpetual and will not be revoked by the NPF or its successors or assigns. THIS DOCUMENT AND THE INFORMATION CONTAINED HEREIN IS PROVIDED ON AN "AS IS" BASIS WITHOUT ANY WARRANTY OF ANY KIND. THE INFORMATION, CONCLUSIONS AND OPINIONS CONTAINED IN THE DOCUMENT ARE ...

Informations

Publié par	Istey
Nombre de lectures	57
Langue	English

Extrait

Network Processing Forum Benchmarking Working Group

Template for IPv4 Forwarding Application Level Benchmark Report Revision 1.0 Editors: Richard Anderson Jr., IBM Corporation, rbanders@us.ibm.com Prashant Chandra, Intel Corporation, prashant.chandra@intel.com Copyright © 2002 The Network Processing Forum (NPF). All Rights Reserved. This document and translations of it may be copied and furnished to others, and derivative works that comment on or otherwise explain it or assist in its implementation may be prepared, copied, published and distributed, in whole or in part, without restriction other than the following, (1) the above copyright notice and this paragraph must be included on all such copies and derivative works, and (2) this document itself may not be modified in any way, such as by removing the copyright notice or references to the NPF, except as needed for the purpose of developing NPF Implementation Agreements. By downloading, copying, or using this document in any manner, the user consents to the terms and conditions of this notice. Unless the terms and conditions of this notice are breached by the user, the limited permissions granted above are perpetual and will not be revoked by the NPF or its successors or assigns. THIS DOCUMENT AND THE INFORMATION CONTAINED HEREIN IS PROVIDED ON AN "AS IS" BASIS WITHOUT ANY WARRANTY OF ANY KIND. THE INFORMATION, CONCLUSIONS AND OPINIONS CONTAINED IN THE DOCUMENT ARE THOSE OF THE AUTHORS, AND NOT THOSE OF NPF. THE NPF DOES NOT WARRANT THE INFORMATION IN THIS DOCUMENT IS ACCURATE OR CORRECT. THE NPF DISCLAIMS ALL WARRANTIES, WHETHER EXPRESS, IMPLIED OR STATUTORY, INCLUDING BUT NOT LIMITED THE IMPLIED LIMITED WARRANTIES OF MERCHANTABILITY, TITLE OR FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT OF THIRD PARTY RIGHTS.

NP/CP Task Group

Network Processing Forum Benchmarking Working Group

1 Reference Design Details The IPv4 forwarding application-level benchmark was run on a reference design with 8 1/100 Mbps Ethernet media interfaces and one PCI control interface. The details of the reference design are given below. 1.1 Component List 1 NPF2002 Network Processor • • 64 MB of 133 MHz DDR DRAM organized in 4 banks • 16 MB of 10ns SRAM • 8 10/100 Mbps Ethernet ports with NPF1008 MAC device. • 1 32-bit, 33 MHz PCI interface. 1.2 Mechanical dimensions • Standard PCI card 1.3 Media, Fabric and Control interfaces • Media interfaces: 8 x 10/100 Mbps Ethernet ports • Control interfaces: 32-bit, 33 MHz, PCI • Fabric interface: None 1.4 Traffic Matrix • Incoming traffic on each 10/100 Mbps Ethernet port is uniformly distributed to the other 7 media port. 1.5 Total Power Consumption • 5 Watts • Steady state power measured during the forwarding rate test. 1.6 IPv4 Control Memory • 1 MB SRAM for routing lookup table • 256 KB DRAM for next hop forwarding table. 1.7 Test Equipment • Data plane tester: IXIA 1600 • Control plane tester: PC with NPF IPv4 API software connected via PCI to the reference design

NP/CP Task Group

Network Processing Forum Benchmarking Working Group

1.8 Routing Table Entry The following numbers of routing entries are accessed by the data plane tester at indicated packet size. Packet Size (bytes) 64 128 256 512 1024 1280 1518

Total Unique DUT Streams 30,000 25,000 20,000 15,000 8,000 6,000 5,000

NP/CP Task Group

Network Processing Forum Benchmarking Working Group

2 Data Plane Benchmark Tests 2.1 Forwarding Rate

1200000 1000000 800000 600000 400000 200000 0 0

Theoretical maximum Base traffic Control IP options

400 800 1200 1600 Frame size (bytes)

Figure 1: Results of Forwarding Rate Benchmark Measured in pps

800

600

400 200 0 0

Theoretical maximum Base traffic Control IP options 400 800 1200 1600 Frame size (bytes)

Figure 2: Results of Forwarding Rate Benchmark Measured in Mbps Forwarding Rate with Internet Mix Traffic = 800 Mbps

NP/CP Task Group

2.2 Throughput

1200000 1000000 800000 600000 400000 200000 0 0

Network Processing Forum Benchmarking Working Group

Theoretical maximum Base traffic

400 800 1200 1600 Frame size (bytes)

Figure 3: Results of Throughput Benchmark Measured in pps 800 600 400

200

0 0

Theoretical maximum Base traffic

400 800 1200 1600 Frame size (bytes)

Figure 4: Results of Throughput Benchmark Measured in Mbps Throughput with Internet Mix Traffic = 800 Mbps

NP/CP Task Group

2.3 Latency 30000

20000

10000

0 0

Network Processing Forum Benchmarking Working Group

Average Minimum Maximum 400 800 1200 1600 Frame size (bytes)

Figure 5: Results of Latency at 100% Throughput Latency at 100% Throughput with Internet Mix Traffic Average = 22000 ns Latency at 100% Throughput with Internet Mix Traffic Minimum = 18000 ns Latency at 100% Throughput with Internet Mix Traffic Maximum = 25000 ns

30000

20000

10000

Average Minimum Maximum 400 800 1200 1600 Frame size (bytes)

Figure 6: Results of Latency at 75% Throughput

NP/CP Task Group

Network Processing Forum Benchmarking Working Group

Latency at 75% Throughput with Internet Mix Traffic Average = 22000 ns Latency at 75% Throughput with Internet Mix Traffic Minimum = 18000 ns Latency at 75% Throughput with Internet Mix Traffic Maximum = 25000 ns

30000

20000

10000

Average Minimum Maximum 0 0 400 800 1200 1600 Frame size (bytes) Figure 7: Results of Latency at 50% Throughput Latency at 50% Throughput with Internet Mix Traffic Average = 22000 ns Latency at 50% Throughput with Internet Mix Traffic Minimum = 18000 ns Latency at 50% Throughput with Internet Mix Traffic Maximum = 25000 ns

30000

20000

10000

Average Minimum Maximum 400 800 1200 1600 Frame size (bytes)

NP/CP Task Group

Network Processing Forum Benchmarking Working Group

Figure 8: Results of Latency at 25% Throughput Latency at 25% Throughput with Internet Mix Traffic Average = 22000 ns Latency at 25% Throughput with Internet Mix Traffic Minimum = 18000 ns Latency at 25% Throughput with Internet Mix Traffic Maximum = 25000 ns 2.4 Loss Rate

100 64 bytes 80 128 bytes 256 bytes 60 512 bytes 1024 bytes 40 1280 bytes 1518 bytes 20 Internet Mix 0

0 20 40 60 Input frame rate (% of line rate)

Figure 9: Results of Loss Rate Benchmark

NP/CP Task Group

100

Network Processing Forum Benchmarking Working Group

2.5 Overload Forwarding Rate

150000

100000

50000

0 0

Theoretical maximum Base case

400 800 1200 1600 Frame size (bytes)

Theoretical maximum Base traffic

Figure 11: Results of Overload Forwarding Rate Benchmark Measured in pps 100 80 60 40 20 0 0

400 800 1200 1600 Frame size (bytes) Figure 12: Results of Overload Forwarding Rate Benchmark Measured in Mbps Overload Forwarding Rate with Internet Mix Traffic = 100 Mbps

NP/CP Task Group

Network Processing Forum Benchmarking Working Group

3 Control Plane Benchmark Tests 3.1 Forwarding Table Update Rate

5000

4000

3000

2000 1000 0 0

Route delete Route add Route update

50 100 150 Route entries per update call

Figure 13: Results of Forwarding Table Update Rate Benchmark

NP/CP Task Group

Network Processing Forum Benchmarking Working Group

Theoretical maximum 1 route update/call 2 route updates/call

3.2 Forwarding Rate With Concurrent Forwarding Table Update Rate 1200000 1000000 800000 600000 400000 200000 0 0

400 800 1200 1600 Frame size (bytes)

Figure 14: Results of Forwarding Rate With 100% of Concurrent Forwarding Table Update Rate Measured in pps

800

600

400 200 0 0

Theoretical maximum 1 route update/call 2 route updates/call 400 800 1200 1600 Frame size (bytes)

Figure 15: Results of Forwarding Rate With 100% of Concurrent Forwarding Table Update Rate Measured in Mbps Forwarding Rate With 100% of Concurrent Forwarding Table Update Rate and Internet Mix Traffic = 800 Mbps

NP/CP Task Group