I know that with that kind of packet loss nothing should work but at the end of the pipeline there are no packets loss. I'm not talking about the gateway but I'm talking about the first external IP I go through. When I ping I usually give 64 packets.
That at least would narrow it down to a) your modem b) the line between c) your ISPs endpoint (or the router behind that if the endpoint is network transparent)
Again this should be already ruled out since I got a new one and there are no chances that's a defective unit since I'm not having any issues inside my network (transfers work well for eg).
Between what?
I asked a friend to use ping plotter to ping that IP and on his side is shown that is own gateway looses a lot of packets when pinging that IP. In other tests he did his gateway it's not losing any packets.
The what ever medium between your modem and your ISP - copper, fiber, wireless - that "line" sorry, I might be abusing line in that context, but I am not native English speaking.
Have you tried WinMTR? I have a feeling I would not trust "ping plotter" as your friend and you should have pretty different routs to the target. or what IP are you pinging - i could give it a shot from europe ^^
It is a tool that consecutively pings all hops that reside in between a certain connection to check where packet loss or latency issues hide. I was not aware that zenmap could do that as well - might you share the arguments?
It does pretty much what you describe, Zenmap is a gui front end for nmap that is cross-platform compatible. What you get is just a visual representation of the CLI's output. Here's an example traceroute:
yeah but nmap -sn --traceroute domain.tld does only one ICMP for each hop - how to configure it to keep pinging them over and over and over I was trying to ask :)
My fault for not interpreting what you mean correctly, I know you can do a very comprehensive representation of entire networks and do lots of detailed scans to find information about hosts, but I'm not sure if there is something like this to determine latency between each hop. I guess you can force it to keep rescanning everything, but that will probably take a while.
I probalby could not get my intention over in words ^^ - my bad most likely;
Did never encounter that - so on linux one uses "mtr" and on windows "WinMTR" to do exactly that - nmap for discovery of course ^^ but thats not what this is all about
I'm giving winmtr a shot at the moment, waiting download 2MB. My friend is having packet loss at the 192.168.0.1 level (I mean his local network) and I see packet loss at the hop's IP. It's weird and I hope I explained myself a little better now. As I start winmtr I'll give you the IP I'm having troubles with through PM.
@ivailo Thanks for the answer but I'm not in the need of a traceroute representation but I need more details on the actual evolution during time of my connection.
You may want to disable IPV6. Windows will use it by default and sometimes IPSs do not have the infrastructure quite right causing potential problems.
Other thing to look at is the speed of the DNS servers you are using. This is a handy dns benchmarking tool https://www.grc.com/dns/benchmark.htm It will allow you to bench mark DNS latency to a large number of publicly accessible DNS servers in a single test.
Open DNS will redirect errors to their own pages which is not always the greatest thing from a security perspective, particularly if they get hacked and their pages are infected with malware. The default list includes Google and openDNS servers and a number of major DNS servers from around the world but you should also include your local ISP's DNS and other nearby dns servers in the benchmark tests. Selecting the fastest servers for your routers DNS servers can help with performance.
The dslams and other infrastructure can certainly be effected by weather. It you log a call with your ISP they should be able to do diagnostics back to your Modem and check for noise on the line etc.
No I'm not in Australia. I don't think it's an issue with the weather and the wear and tear on the cables. I asked already to check my line and they say that the power of the signal and the noise are good. But they're not since I'm getting only 9dB noise margin for the download with a really good Netgear D6200 modem/router. I tested a crappy one and I lose 1dB and more.
Is your main computer you are getting poor performance connected over wifi?
Just a thought. is there another nearby wifi network that is running on the same channel as you? Changing channel at the wifi access point to a less busy or free channel can help resolve performance issues.
Local access can appear OK because the local bandwidth makes up for the latency. It becomes more apparent over slow internet connections
Haven't read the whole topic yet and have to think about an actual answer, but: STOP USING PING PLOTTER OR ANYTHING PINGING INTERMEDIARY DEVICES AS ANY KIND OF TEST!
See, network equipment (or at least, Cisco devices) has separate command and data/switching planes. They have different latencies. When your ping just goes through the switch, it goes through data plane. When you ping the switch itself, it lands on control plane. Strictly speaking, intermediary devices don't have to answer your pings at all.
This sounds reasonable. So than why my connections is going to shit and having issues even downloading a stupid document from the internet? Yesterday I was trying to upload a bunch of docx on Google Drive it and it went offline because I was using too much upload. It never happened before and I did upload files up to 1GB before. I do get the same speed and dBs at the modem as before the issue arise.
