Thailand May Highjack Speedtest But Not Internet Frog

[ikebukuro5]

Over the past little while I've noticed that spekeasy.net gives me up to 9mgs but I barely get 1.0 on internetfrog.com and other smaller speed tests.

Could it be that Maxnet "opens" the flood gates to speakeasy.net to make themselves look better?

Why is my connection crawling and very slow yet my speakeasy.net test shows so much speed?

I suspect some monkey business here....

[basjke]

I think that's a very good question but it will remain unanswered.I opened a similar topic some time ago,http://www.thaivisa.com/forum/index.php?s=&showtopic=316936&view=findpost&p=3161848, as I get speedtests on true between 12.000 and 14.000 kbps to any international server any time of the day.Still a download,for example an attachement from hotmail,will download at the amazing speed of 15 KB/sec.That is 100 times slower then my speedtest and I am lucky if it completes the download.

[monty]

On my connection speedtest seems to be accurate.

I on average test between 0,5 and 1 Mbps (although I have a 4 Mbps connection), http downloading a single file or youtubing also gives me that 0,5 to 1 Mbps, but doing multi threaded connections (through download manager or well seeded torrent) I instantly maz out my connection at 440 kBps.

Maxnet Premier 4/1 Mbps

[briley]

Am I the only one who thinks that all these 'speed test' things are a waste of time?

I've tried a few and get very varying results.

A 'real life' test is to find a big file on a good site and down load it, read the download speed on your screen.

I download open office using uTorrent (a completely legal download) and get speeds near to my 2meg limit almost every time.

Of course this does not check your upload speed so efficiently but again with torrents my maximum upload is around the 512K I pay for.

(Do note you often have to multiply the figure in your torrent program by 8 to get the figure talked about by the ISPs as one is bytes and one bits.)

[nikster]

Yeah when everyone's using speedtest.net to have "proof" of their speeds, I wouldn't be surprised if the ISPs cheated a bit there.

For me though, for the most part speedtest.net is pretty accurate. On TOT and CAT.

Torrent download is a parallel download, I used to use the Engadget RSS feed in Safari for this - it loads all images simultaneously, so high number of parallel connection. Same with torrents. This is the speed you can get out of your line at best, and often it's equivalent to speettest to BKK, and the max your line delivers. Thai ISPs like to limit single connections rather than your entire line. So if you open many parallel connections, each one's limited but the total isn't. I guess there is a technical reason, doesn't really make any sense otherwise.

To me, single download is most often the problem, that's when I open a connection to our servers in the USA and download something. There, speed varies dramatically, consumer CAT 4Mbit, for example, maxes out at 4Mbit for torrents every time. But single download goes slow as snails at 15KB/s or 20 KB/s. The CAT business premier 2Mbit delivers over 100KB/s on the single downloads... (while being limited to 2Mbit on torrents)....

Another thing Thai ISPs do is caching. A popular download will often download at full speed. Mac OS updates for example, or the latest Firefox. These are cached locally in Thailand, so you get the full line speed even at single downloads.

Edited December 16, 2009 by nikster

[Veazer]

It's usually fairly reflective of actual conditions for me, using a maxnet 6mg biz connection. When the connection goes to hel_l, speedtest reflects it.

My cheapo True cable connection at home runs circles around my expensive maxnet connection from work, frustrating. I can pull almost 600 KB/s on the home connection, about half that on the maxnet connection. Maxnet (oh sorry! "Triple-T") is horrid.