mudmanc4

a-b-c-d-e If the test originated from 'e', in other words, the computer, and the test server was 'a' that is where the data is going to end up 'e', bounced back or not. Now correct me if I'm wrong, were not talking about a ping, were looking at data. If the data was measured at the VPN endpoint, or exit node, that would most likely, in the PIA case, leave a massive pipe, and feature the highest throughput possible at the time, between the endpoint and the testing node. Which would output some very impressive 'speeds'. Again correct me if I'm wrong, thought the path between the originating node (computer) and the test (server) stays available through nodes a-b-c-d- and e, the path obfuscates the originating point, to a point. If not, explain how it would even be possible to get a less than stellar 100GB+++ result ? (Gigabit)

The one and only reason Ookla is even remotely available on the net today, is the simple fact first of all, they have the ISP insiders market flooded. Who else knowing, would harbor a flash base script within their internal network, unless the ends outweighed the means, or likelihood of an infiltration. If your confident enough to allow this thing to live within your public network, than obviously you've got said network protected. Fair enough right? That type of ISP network protection would likely come at the cost of the consumer. If for no other reason than isolation, which is the complete antithesis of the meaning to testing throughput across networks, that at the very least represents what the intentions are for the masses taking the test. Secondly the ISP know well the entire idea of a test being run on any high level network, is useless and nothing more than a 'feel good' item. These tests of Ookla, should be completely isolated on a per ISP level, if they must be used at all. Of course they are useful for tech to determine if there is or not an issue between modem/ head end (or wherever these flash tests are living). No?

This is due to both sides being required to be set the same to recognize one another. That is why there is an auto setting, which waits for the 'interrupt' instead of demanding. If it is a black box of a modem, you'll not be able to authenticate to change those settings. Basic network topology would help, care to share?

@CA3LE I was not aware there were default values for the Auto test when accessed in such a way, I assumed input was required in order to test with specific user input settings. I'll have to check that out.

I just started an auto test from the UK server and is running properly. The link should not be accessed directly, use the web interface to initiate the test.

Amazons videos are a perk of prime, some people have no issues with it, others have poor quality. I might try setting the quality to less than HD

tracert eu.testmy.net

You could also run a trace route to testmy.net , this could show if the slow spot is before or during the route. As if there is still something hanging around on the hard drive chewing up the available bandwidth. After a closer look, it would give the appearance you are tethered in some way, to the android phone from your desktop, no? Your results show very few tests not originating from the phone, however some are much higher than others, which could be those from the phone itself, and the lower results from the desktop. edit: Unless of course the system is following your selection of extra identifier instead of what is actually being tested

Hi TroublesomeJohn, The first thing that tosses a flag for me is the sequence of wording --windows:torrents So many times we get 'things that we did not expect to when grabbing a torrent, I'm sure you are more experienced to already be aware of this, just voicing the first thing that comes to mind. If there is something hiding in there, it could be using the machine as a webserver, botnet, or even an email open relay. Of course low disk space, an older HDD and or low system memory could all have an adverse effect on throughput. First thing I would do is run through and clean up everything, organize all wanted / required applications and documents, and chuck the rest. F-Secure has a great online scanner, and from my experience has a tremendous record for sniffing out baddies on the machine. edit: I should also mention good cables and connection many times are some of the less obvious, yet most simple 'troublesome' areas. Also make sure there is not any QOS settings in the router config that could be limiting the port, mac, or IP range.

Hey Whard42, welcome to testmy.net.

Been there done that all too many times. Looking too deep, being in a hurry to fix it instead of starting at the basics. Honestly you have a great power supply right there.

Makes perfect sense. The short likely detected in a circuit breaker within the power supply, and tripped the millisecond it detected feedback.

You have just described so many others issues as well as your own. Many, on different levels and many more in very much the same predicament. I feel your pain. Though I would rarely recommend, but in your situation it might be a nice step up, have you looked into satellite?

Welcome to the forum @GIASHALONGKARI I am interested in seeing more results from Saudi Arabia, would you mind testing a few more times ?

You probably already defaulted the CMOS in that motherboard right? And still does the same thing? You might be right in assuming a short somewhere, at the same time make sure there's not some low level software getting in the way.

@Sean Always drop into a shell or command line, and run a traceroute when you see inconstant or erratic connection. This will show the highest latency per hop, and can answer many questions.

This will depend on what server, or mirror you are testing from. If you are strictly using the default server, Dallas, TX , then your all set. Otherwise you could ping each mirror and get the IP.

There are many forms of logging, however to get the most accurate results, a stand alone appliance is most useful. A device which lives between the modem and the internal network. Such as a router or firewall with such capabilities. If you have an older machine laying in the closet with two network cards, you could use pfsense , and delete your current router. Unless of course it is combined with the modem, then the firewall would be placed on the inside of the modem/router, or LAN. Pfsense has all the features that would answer the questions you are asking. If you are interested, I would be more than glad to assist you getting everything setup here on the forum.

Testing is the same logged in or as guest. Run an auto test while logged in , then run the very same test when not logged in, your connection is likely sporadic for one reason or another. Running a traceroute to the 'au.testmy.net' mirror to show the hop between you and the testing server that is causing the issue.

Does look a bit rough for a 15Mbps connection. Try a few other test mirrors at different times of the day.

Looking at your stats it does show as you stated: https://testmy.net/stats/?&t=c&d=12262015&x=10000&l=25&q=898170596551 Maybe run a traceroute to 'testmy.net' which should be the main test server, and see where the highest latency is in the route.

The above is over a private VPN, then opened up to the Dallas server, not seeing any issues here. It is possible there is something going on with the peering of your ISP, somewhere along the line. These things generally straighten themselves out if the problem is widespread.

https://testmy.net/stats/?&t=c&d=12222015&x=10000&l=25&q=6618129649461 Maybe try running a small auto test for at least 24 hours, this way we can determine if there are specific slow times with your connection.

Since the switch itself is un-managed, it will act as a hub. What this does is take the full available bandwidth at the WAN side, and divide the data flow by how many different devices are requesting data. The JGS524 should work as a gigabit device to expand your network without having to get down a dirty with networking. Say you had two machines, each machine has gigbit capabilities, and are of equal quality (performance wise), each machine is trying to draw a one gigbyte file, the available bandwidth (depends on your specific ISP package as well it it's quality); then the full available amount of data would be theoretically split between the two machines, and the downloads would theoretically finish at the same time, give or take. Again, theoretically, each machine would have a path open to the tune of 100Mbps at full throttle, in a perfect world. With good wiring and a solid ISP drop (wire from the port outside) and a late model computer, I would be satisfied with an average of 180Mbps, and peaking to 210Mbps on seldom occasions.

Looking at your results, it would appear as things are shaping up a bit throughput wise yes? Nice upload by the way! https://testmy.net/quickstats/morlock