Hot answers tagged

14

In a WLAN iperf TCP throughput test, multiple parallel streams will give me higher throughput than 1 stream. I tried increasing the TCP window size, but I still cannot achieve the max throughput with just 1 stream. Is there something else in the TCP layer that is preventing the full link capacity from being used? In my experience, if you see significantly ...


8

The -l option is for the buffer and doesn't influence the amount of data transferred. You have to specify the desired amount of data with the client-only option -n in KByte or MByte. So for 10GB, use -n 10240M Example: With the defaut buffer size of 8KB: iperf -c 10.1.1.1 -n 10240M ------------------------------------------------------------ Client ...


7

iperf testing over UDP requires the client to specify what bandwidth to send at (-b option). So for your UDP test results listed, your iperf client was configured to push data to your server at a rate of 50 Mbits/sec. iperf testing over TCP sometimes requires that you adjust the TCP Window Size option (-w) to fully saturate a fast link. Try using "-w ...


6

I think you're confused about what Iperf is, and what purpose it serves. From the Iperf Man Page: iperf is a tool for performing network throughput measurements. It can test either TCP or UDP throughput. To perform an iperf test the user must establish both a server (to discard traffic) and a client (to generate traffic). Iperf simply ...


5

iperf can be used to test transportation protocols like TCP and UDP. VOIP is an application protocol. This is another layer of the Internet Procol Suite. So the answer is: no, you cannot test VOIP, but you can test the transportation protocols (typically UDP) used for VOIP calls.


5

Don't trust GNS3 for throughput tests. As soon as you use a router in your topology you are likely only to get 5mb/sec tops, regardless of whether you use a serial, ethernet or fast ethernet interface and regardless of how powerful your PC is. (Note that the performance hit won't be seen if you use a GNS3 switch instead of a router, however, the switches are ...


4

The Raspberry Pi 1,2 and zero have a 100Mb/s network interface card, so this is likely what caps the traffic. (the raspberry pi 3 has a gigabit Ethernet interface but it is connected to the processor through USB 2, so is limited to USB 2 speed at 300Mb/s). The reported Transfer and bandwidth are consistent with a 100Mb/s connection. As hinted by @Ron Maupin ...


3

Depending on your Cacti version, we collect errors and discarded packets without any problems on Version 1.1.38. Here's how we do it. Select the tab Console, then extend Management and click devices. In the device list, select the switch/router you want to get the information from. Now under the selected device, click create graph for the device. Now at ...


3

You can use patched IPERF compile, and then run: # iperf -s --udplite 0 ------------------------------------------------------------ Server listening on UDP-Lite port 5001 Receiving 1470 byte datagrams UDP-Lite datagram buffer size: 160 KByte (default) ------------------------------------------------------------ From help: -u, --udp use ...


3

If both devices are connected to VLAN 1, then that command has no effect. That command would be for traffic leaving the interface, but if both devices are on that VLAN, the traffic never leaves the VLAN. There are several options for limiting the traffic rate on switches. You can limit on the physical switch interface with that command by applying it to the ...


3

The most common mistake made when using netem to emulate network delay is to not increase the packet limit. You didn't. Therefore you tail drop a lot of packets and tcp cannot build a queue long enough to emulate the path. try limit 1000000 or more. Monitor your netem stats to see if you drop any packets. Plug: we use (bufferbloat project) flent.org's test ...


3

That's not a shock. These day we automatically tune the amount of RAM assigned to TCP buffers (aka the maximum TCP window size), and by specifying the window size you are turning off the automated tuning. The flat-topping at around where the bandwidth-delay product would be about 64KB suggests to me that a suitable Window Scale Option isn't being negotiated,...


3

As Marco mentioned, Receive window sizes (rwin) will increase and decrease automatically during a TCP session depending on how lossy the connection ends up being. It is therefore possible that when you omit the -w flag from iperf the window size is being dynamically scaled higher than your static values. Remember that maximum theoretical throughput is ...


3

Receive window sizes (rwin) will increase and decrease automatically during a TCP session depending on how lossy the connection ends up being. It is therefore possible that when you omit the -w flag from iperf the window size is being dynamically scaled higher than your static values. Remember that maximum theoretical throughput is rwin/delay. Based on ...


3

When more traffic than a link can handle is sent to a link, the interface to that link will end up dropping the packets destined for that link. This is congestion, and there are some small mitigations for congestion, such as queuing, but, when it comes to networks, packets are preferred to be dropped as soon as possible rather than delaying them. In your ...


3

There is one particularity about iperf 2.0.x with UDP: Within the limits (CPU, NIC, UDP/IP stack and its buffers) of the sending system, iPerf will send traffic at the payload rate you specify with the -b parameter (if left out, iperf will default to 1Mbit/s). EDIT: D'oh! Only now I spotted that you are actually using iperf3. I'll leave these comments ...


3

It seems that your omnipeek client card cannot capture higher data rate traffic. The ACK frames are sent with basic data rates and the DATA frames are sent with higher data rates. The issue can come up if your client and AP are transmitting on 3 stream and 40 MHz and omnipeek can listen on 1 stream and 20 MHz. This can also happen if the omnipeek client ...


3

The bandwidth command sets the reported bandwidth and is used for metric calculation only. It does not affect the actual interface bandwidth. If you want to limit bandwidth, you will need to implement traffic shaping/policing.


3

iperf measures bandwidth. Tools like Ookla Speedtest report download and upload speed. "Speed" is a term that's often misused. All these tools measure throughput -- the amount of data sent in a given amount of time. Speedtest and similar programs measure throughput to their servers -- which may or may not be representative of your application. ...


2

Here's the calculation for the maximum throughput of a single tcp stream. (TCP Window [bytes]/RTT[seconds]) * 8 [bits/byte] = Max single tcp throughput in (bps) So you have a bottleneck and latency plays a large role.


2

I'm inclined to believe that it is your caps/non-caps on the bandwidth designations (i.e. -b 10M -l 32K -w 128K). I can't test it because we don't know what version you're running. Iperf does make a distinction between bits and bytes: -f, --format [kmKM] format to report: Kbits, Mbits, KBytes, MBytes Though, the sections you were modifying were ...


2

While passing downlink traffic (from the Access Point (AP)) to client-nodes, the AP schedules the packets from the packet-queues in the appropriate way and also does not face any contention on the medium. On the uplink however (from the client-nodes to the AP), the client-nodes contend for the medium to send the packet. This creates a delay. This delay in ...


2

TCP uses SLOW START to maximize bandwidth and minimize lost packets, so it adapts to the slowest link in the chain by waiting for acks to come back. UDP leaves it to the application layer, and iperf doesn't wait for acks it just obeys the bandwidth you tell it to use with the --bandwidth option even if it overwhelms a link and causes substantial packet loss....


2

What does iperf UDP streams do iperf will attempt to send UDP packets to reach the rate given in --bandwidth. If that parameter is missing, it will attempt to achieve the rate of the transmitting interface. what can i do on my Apple You can measure both computers' resource use, as there's a bottleneck. CPU is the most likely bottleneck, given that ...


2

There are two aspects for bandwidth: The link speed as indicated by the interface presents the upper limit of what that interface can transport. The practical throughput (maximum capacity) you can achieve using that interface. This is much harder to determine and there is no generally accepted method for measuring. High-speed Internet uplinks beyond 100 ...


2

iperf can help you generate flows of a certain bandwidth. You can then check if your network behaves as designed and examine your equipment for signs of problems: packet drops due to exceeding aggregate bandwidth, overrunning buffers, CPU overload, significant rates of checksum errors, and so on. A simulated network might behave somewhat differently from a ...


2

Yes, a set of equal values will give a Jain of 1. Your values are extremely similar, I calculated 0.999971.


2

Welcome to the downside of cloud computing! Jitter (latency variability) and aggressive oversubscription. Your node is competing for shared resources at many different points that can be greatly oversubscribed in public clouds. I would tackle this problem by measuring/capturing the following: EC2 to EC2 (intra-cloud) iperf to elminate the traversal of ...


2

Like user27884 said, you could configure Iperf server in this way: iperf -s -u -p 5061 -S 0xC0 -l 200 ..and the iperf client like this: iperf -c 127.0.0.1 -u -p 5061 -S 0xC0 -l 200 -t 200 -b 200k -i 10 this will show parcial results every 10 seconds and the entire test with 200 seconds.


2

Cisco CSR does not have support for SR on AWS, see here for AWS requirements/support. There is no support for SR as in SR-IOV "single root virtualization", there is support for SR as in segment routing. ALso keep in mind that AWS EC2 performance varies by instance type. The AWS Marketplace page for the CSR 1000v lists lists 4 instance types. Though not ...


Only top voted, non community-wiki answers of a minimum length are eligible