O Canada! > Canadian > Canadian Broadband > Caneris Speed Test Server - open to the public

Caneris Speed Test Server - open to the public

It doesn't seem to be working for me. I'm getting an error.

reply to CanerisErik
Does anyone know of a speed-test that does *not* try to invade your system?

I really would rather do without ...

090627 15:53:02 /System/Library/Frameworks/JavaVM.framework/Versions/1.5/Commands/java[81649] Java has built shared archive /System/Library/Frameworks/JavaVM.framework/Versions/1.5.0/Libraries/classes.jsa

Is it really so impossibly hard to create and host a simple, unpretentious data-transfer test that uses only elementary facilities (eg: ftp, scp) and can be run even from a router or modem with a decent enough shell?

This particular test claimed 36 kbp/s DS on Safari and has paralyzed Opera!
--
»[RFC] Connection / Speed Problems Checklist

reply to CanerisErik
It's working now. Thanks Erik.

I didn't make any change. The Java version works for me. The flash doesn't...not sure why yet, but I don't think it's on our side, might be something with the DSLR software.

Where else can I test the flash speed test? Other ISPs seem to be using only DSLR's Java version.
--
Erik - Caneris - Internet solutions and more.

Your buddies over at speedtest.teksavvy.com have (or had) both the flash and java versions.

reply to planiwa

reply to jfmezei

reply to CanerisErik
The DSLR side appears to be down. I've PM'd the appropriate person.

reply to CanerisErik
Sorry Erik,

I know it's not your fault that matters are as they are.

I apologize for having used this occasion to express my annoyance with the ubiquitous trend towards veneer over substance. Had this come from Microsoft or Bell I wouldn't have bothered.

. . .

It doesn't seem hard, nor expensive, nor risky, for an ISP to provide an FTP server with 5 random-content files named:

333333
999999
3333333
9999999
33333333

(similarly for US -- restrict to own domain, or limit US files to under 5 MB, and delete after 5 min. restrict the whole thing to under X simultaneous streams, depending on available unused bandwidth.)

This would be of great value to many users, when what they need is a speed measurement without attitude and decorative complexity. (It would be a significant factor for me in choosing between ISPs to contract with and to recommend, since it would make it very much easier to diagnose certain common problems that are associated with much user frustration now.)
--
»[RFC] Connection / Speed Problems Checklist

reply to CanerisErik

When I signed up with E-Box, I noticed some apparent performance problem. They had no problems setting up my web space so I could FTP files back and forth to measure performance and compare it with the speed tests.

Found out that the ftp performance on my alpha-vms box was lower than for OS-X/unix despite it having large window sizes etc.

Another way to measure performance is if you have a managed switch, the switch itself can give you current throughput on a port (the port going to the modem).

reply to planiwa
No worries. Good idea, except a bit hard to implement and use for non-technical folks, requires an actual FTP client, etc.

These and other issues is largely why the speed tests are the way they are...the HTTP GET/POST way makes it super easy for those with a browser, especially those who have never heard of what FTP is, to run the test. It also streamlines the download and upload and displays the results nicely in the end, after performing multiple downloads and uploads of various sizes. And it's integrated with DSLR, so you can't fake the results when showing off your tests

So, I think it can't be argued that it's easier, quicker, and more accurate for most users. Unfortunately most of these features can't be replicated easily without some sort of client-side plugin, like Java or Flash.

I tried to write my own speed test software last night and this morning and it doesn't seem very accurate so far. It's HTTP-based using mostly server-side stuff...about 300 lines of PHP and AJAX. Please try it out at »speed.caneris.com/new

You will need cookies and JavaScript enabled to use this.

This is my first attempt at it, so the results in this case for download/upload are based on average speed across your attempts...you would need to download/upload more files and download/upload the same files repeatedly, to get a more accurate result, otherwise you will see consistently lower numbers versus other speed tests.

If there's an interest, I'll continue developing this as an attempt to get a basic Java/Flash-free speed test software. I don't know how feasible it is to get an accurate result out of it just yet. It seems to be worse in IE than FF so far, but at least you can use your browser's UI to gauge the download speeds as you download the files.

Comments?
--
Erik - Caneris - Internet solutions and more.

Erik,

Is there anything wrong with your website? It takes forever to load today and I'm not talking about the speed test.

Thanks,

Mike

reply to CanerisErik
is Caneris teksavvy reseller?
it looks like that to me...

reply to Mike2009

Thank you. We are investigating. I will post more details later on.

reply to CanerisErik
DSLR's iPhone speed test measures both upstream, downstream, and latency. And since the iPhone supports neither Java nor Flash, presumably it uses AJAX. No point re-inventing the wheel.

reply to CanerisErik
Just a quick update: right now the problem appears to be with inbound TCP connections to certain networks only. It is intermittent. This is why you perceive it as slow speeds when doing HTTP but ping/traceroute, DNS, etc. all work fine. It is also why there are few complaints, everything outbound and everything on-net works fine. This appears to be an issue with one of our border routers not forwarding packets. It's a pretty bizarre issue...ICMP, UDP, outbound TCP, everything works, and on-net inbound TCP works, and off-net inbound TCP works sometimes. Still working on it.
--
Erik - Caneris - Internet solutions and more.