Maximizing File Transfer Performance Using 10Gb Ethernet and Virtualization


Synthetic Benchmarks versus
Real File Transfer Workloads
for Native Linux*
The following two sections examine
the differences between synthetic
benchmarking and benchmarks generated
during actual workloads while running

Synthetic Benchmarks



slot, which limited the throughput to about

In the second case, the card was plugged
into a PCIe Gen1 x8 slot, which allowed full

PCIe slots can present a problem if the
physical slot size does not match the

determine which slots are capable of full
PCIe width and performance, check with




achieve 10 Gbps throughput for one 10G


As demonstrated, achieving 10 Gbps
transfer rates is quite easy using a

looks at a case where actual workloads

0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
PCIe* Gen1 x4 PCIe* Gen1 x8
Receive Throughput (Mbps)
NETPERF
Native Linux*: Synthetic Benchmark
Figure 4.
Benchmarks Based on Actual Workloads
Real applications present their own unique

netperf results as a reference bar on the

Tool choice obviously matters, but the
standard tools are not very well threaded,





and the rsync(ssh) case both achieve only



performance boost, and two promising


Figure 5.
0
10
20
30
40
50
60
70
80
90
100
NETPERF SCP
(SSH)
RSYNC
(SSH)
SCP
(HPN-SSH)
RSYNC
(HPN-SSH)
SCP
(HPN-SSH +
No Crypto)
RSYNC
(HPN-SSH +
No Crypto)
BBCP
Avg. CPU (%Util)
Receive Throughput (Mbps)
Native Linux*: Various File Copy Tools (1 stream)
Receive Throughput– 1 Stream Avg. CPU (%Util) – 1 Stream
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
