LAN throughput test

[RTSLVU]

I am looking for ideas to implement a LAN throughput test. We need a continous readout of upstream and downstream data rates, a graph would be nice.

The program neede to be able to do the transfer of data and measure the thoughput, big enough data to get 80% of the connetions max data rate.

[a_clucker]

Hi RTSLVU,

LAN throughput can vary by application, as well as environment, e.g. your current computer load. In the wide majority of applications, LAN is not a deterministic protocol. There are a number of third party network throughput testing applications available online that might be able to help you determine an idea of the maximum throughput rate for your computer. Two that I found during a quick search can be found here and here. 

However, the throughput for a given LabVIEW program could depend on many things, including programming practice, general application purpose, loops, memory buffers, packet size, and so forth.

If you do want to elicit further assistance with this issue, please reply with detailed information about your application so that the community can better understand this question.

Thanks,

Andrew

[Dennis_Knutson]

I've tested a lot of network devices, both for production and design validation. For the design validation, we found it necessary to use a hardware solution for repeatable and accurate results. You can get hardware from such companies as Spirent. For production testing, you can use software tools for a rough approximation though results could differ with the os. Do not attempt to do any sort of testing with limits on the corporate network. Use a dedicated network between your pc and the device. That connection should be made primary in your network settings. We also found, after a couple of years of testing, that for production test purposes, a simple ping was enough to verify a network connection after the actual design was validated.

[RTSLVU]

In a nut shell we have a PoE power inserter our customer wants us to have a continuous data through put of 80% (it's a gigabit rated device) during all environmental, power, ESD, EMC, etc. testing.

We need to continuously monitor throughput for changes in data rates, BER, etc. 

I was hoping for a simple solution like transferring large amounts of data back and forth through the device but am open to hardware suggestions.

[a_clucker]

You could use WireShark to examine your network. I also found a LabVIEW example that uses Winpcap calls (which is also what WireShark uses) to carry out a similar function: Ethernet Packet Sniffer Utility Using WinPcap Functions. You might need to make some modifications to it in order to get the information you desire -- note that it displays timestamps, and that you could use these to determine elapsed time.

Other than these, a hardware solution might be a route worth investigating.

Regards,

Andrew

[RTSLVU]

Stupid Content Cleaner bringing old new to my inbox.

[Jay14159265]

---

@a_clucker wrote:

You could use WireShark to examine your network. I also found a LabVIEW example that uses Winpcap calls (which is also what WireShark uses) to carry out a similar function: Ethernet Packet Sniffer Utility Using WinPcap Functions. You might need to make some modifications to it in order to get the information you desire -- note that it displays timestamps, and that you could use these to determine elapsed time.

Other than these, a hardware solution might be a route worth investigating.

Regards,

Andrew

---

What was "cleaned" from this post?

[PhillipBrooks]

D.O.G.E. Ai at work?

[mcduff]

---

@Jay14159265 wrote:

---

---

What was "cleaned" from this post?

---

Guessing, internal NI links that were previously bad, now updated

[crossrulz]

---

@mcduff wrote:

---

@Jay14159265 wrote:

---

---

What was "cleaned" from this post?

---

Guessing, internal NI links that were previously bad, now updated

---

I saw somewhere something about removing old/outdated information. But there have also been a big cry about links not working anymore due to servers getting moved, reorganizing, etc. If I had to guess, it would be a correction of the NI forum links in the post in question.