In normal network with Cisco 3650 (1Gbps) and Linux base servers, the ping time is between 0.150ms to 0.250ms. (I also mean using ping or ICMP is to refer to all protocols) I want to know is there any Cisco or Linux Configuration to lower up ping time? for example 0.05ms to 0.7ms. (someone wrote a new INTEL NIC Driver and reach better network performance)

closed as too broad by user36472, Ron Maupin Jul 18 '18 at 13:37

Please edit the question to limit it to a specific problem with enough detail to identify an adequate answer. Avoid asking multiple distinct questions at once. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • 2
    I'm presuming it's actually RTT you're trying to minimise, as measured by ping: what's the actual protocol you're wanting to improve the timing of? – jonathanjo Jul 18 '18 at 12:36
  • 1
    Ping uses ICMP, so the only latency you are measuring is for ICMP, and none of your applications use ICMP as a transport protocol. In any case, your latency is less than a millisecond, and there is nothing wrong with that. I think you have an X-Y Problem.Please edit your question to explain exactly the problem. – Ron Maupin Jul 18 '18 at 13:35
  • @RonMaupin I know 0.150 ms is very good for network. I want to know can we reach to lower time on this network by changing software configuration? – hamid bayat Jul 22 '18 at 11:52
  • 1
    Unfortunately, based on your comments and edit, the question is off-topic because the host/server configurations (including your NIC and driver) are off-topic here. The switch will switch at wire speed (there isn't much you can do about the wire latency), and the switching is done in hardware. You need to be cognizant of interface oversubscription that could cause dropped frames, but that doesn't seem to be the case with less than 1 ms ping latency. You could try to ask about your host/server configurations and software on Server Fault for a business network, or on Super User for a personal network. – Ron Maupin Jul 22 '18 at 17:51
  • 2
    Also, trying to use ICMP as a proxy for determining how other protocols perform is simply incorrect. Generating ICMP messages is usually far down the list of what nodes will do, preferring to perform real tasks, and generating and sending ICMP messages when they have some time. You may have noticed in a traceroute where some intermediate nodes actually have a greater latency than the end-node. That is because they are busy performing real work, and they send ICMP messages when they get around to it. – Ron Maupin Jul 23 '18 at 4:23

You need to investigate low latency switches. Nexus 3000 is designed for financial applications where low latency is essential.

Increasing bandwidth can also reduce latency, so look for 10Gbps or higher end-to-end (including PC connectivity).

Also investigate a QOS feature called low latency queuing. The feature is designed for voice and video. Anything classed as priority will be forwarded ahead of other traffic, so important not to put too much traffic in the priority queue. Low latency queuing only really works if you have other types of traffic on the network to push the priority traffic ahead of.

Not the answer you're looking for? Browse other questions tagged or ask your own question.