Jump to content

CDN route scores worse than Dallas


CA3LE

Recommended Posts

Good Day,

I am interested in trying your TraceMy.net tool (beta). I came across a reference to it on my search engine IXquick.com. Do you approve users using the beta before it's release? If 'yes' how can I qualify?

Also I am stumped on the usefulness of the CDN Push Alt Route to Dallas server for accuracy. My test are always faster with it than my default Dallas server or the other US servers.

I am 120 miles north of your Dallas server and use it as a default along with using Multithread Speed Test.

Please advise.

Thank you and Best Regards,

Rob

 

Hi Rob,

 
Glad to hear that you're interested in TraceMy.net.  However, I'm sorry, the beta is over.  The information I gathered during the beta is being used to build a primetime version.  I'll probably want to do one or two more very brief beta rounds before I fully release it.  If you'd like to be contacted for any future betas just vote 'yes' in the TestMy.net Beta topic and I'll contact you and add you to the beta group next time I have a beta.
 
So, personally I also get faster results straight to Dallas.  That's actually the prefered server, everyone defaults to that server for a reason.  There's a common misconception that you must test close to the server to get a good speed test.  Not true, in fact you should probably use longer routes to truly put your host, their peers and their peers peers to the test.  Dallas is one of the most popular hubs in the world, my other servers are located in other hot locales.  I believe that speeds to these locations and the Cloudflare and Google Pagespeed testing options are truly accurate representations of the speed anyone is to expect to get from most sites, servers and services they'll encounter.  The data is being served similarly to the way you often get data from sources today.
 
I mentioned the Google option, this is actually the fastest for me in most cases.  I'm still working to trick pagespeed into working with my classic test. :razz: -- I contacted Google to see if they'd lift some restrictions for me, ... nope.  They can only do up to like 62 MB... I need it to do at least 200.  Sucks because it's a strong performer up to its limit.  It works really well with multithread.  I may be able to convince them in time to help me make it work with everything I offer.
 
Look at my screenshots, 108.4 Mbps linear to dallas.testmy.net (my fav, because of the TiP data), 117.2 using google.testmy.net and only 93.1 to cloud.testmy.net.  BUT if I do the same on my cell phone, using Verizon, the results are totally different.  Dallas wins at 5.7 Mbps, closely followed by cloud.testmy.net at 5.5 Mbps and then the previously fastest method is now the slowest, google.testmy.net only hit 3.4 Mbps.  These are all on the same computer, same browser, same time, etc... I just tethered my phone real quick to show you.  Different hosts route differently, I offer the various servers and options so you can compare and decide which works best for you.  
 
What works well for me may not work as well for you.  I personally use the default server most of the time, it's BY FAR the most powerful machine that I have.  24 core RAID SSD beast.  It's also well connected and hosted in the most popular Internet hub.  I'm about 1000 miles from Dallas and I still max out my connection.  On other (commercial) connections I've tested at over 5000 miles and nearly maxed out.  As long as the provider is doing their job distance is less of a factor.  If you're really close to that server you might want to test and compare some routes at greater distance. Don't make it easy on your ISP!  Warning, if you go too far it can be depressing... but then you can do the same thing on other connections and they'll test much better.  It all depends on the provider and their peers.
 
Look at my results from today ... you'll see that some of the multithread results are slower, those are to asia and eu.  On my Comcast connection I'm usually able to pull about 4-5X my single thread speed.  EU, 5000 miles away I go from 20+ Mbps to 80+ Mbps... both are right but really I'd only be able to download something at 20 Mbps unless it's split.  Then look at my Verizon connection tethered off my iPhone.  Only 2X to EU and actually ~40% slower multithread to Asia.  Different situations produce different results.  Some connection will multithread really well... then others appear to do worse.  Yours..... appears to be dead even, little variance between results.  Looks like you select the option that tests US coast-to-coast simultaneously, even better that your multithread is similar to your single thread performance.  You want to see those two numbers in line with each other.  If the multithread number is higher that means that there is room for improvement in your single thread performance.  If the single thread number is higher that may mean that the devices performance is lacking.
 
I hope this helps.  Thank you for your continued patronage and support.
 
- Happy Testing!
- Damon - TestMy.net
Link to comment
Share on other sites

×
×
  • Create New...