View Single Post
  #1 (permalink)  
Old 02-22-2008, 09:02 PM
cobra9777's Avatar
cobra9777
Lurker
Offline
Location: North Texas
 
Join Date: Aug 2007
Posts: 24
Reputation: 0
cobra9777 is a n00b
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Speedtest results, tethered vs handset browser

I've read several threads about various data issues after upgrading to a Rev-A ROM. Some of the recent stuff was evidently a DNS issue on the East Coast which has since been resolved. In any case, I've been experiencing some general sluggishness and disconnects since upgrading so I decided to run some bandwidth and latency tests.

First, I ran several tests at the Speakeasy site (http://www.speakeasy.net/speedtest/) while tethered via USB to my desktop. Everything looked great... averaging about 1250 down, 350 up, and not much variability test-to-test. Pings to Google were decent for wireless at 100-150ms and no unusual spikes.

I then ran several tests untethered using PIE and the BBR mobile test site that everyone seems to recommend (http://text.dslreports.com/mspeed?jisok=1&more=1). Results are not so good... 200-500 down on average, and lots of variability with a couple tests sub-100.

I then did a soft reset and tried a different site that had been recommended to me (http://mobilespeedtest.com/). Similar results as the BBR site; not good. I can't get Speakeasy or any of the good Java/Flash sites to work properly with PIE on the handset. So, just to rule out path differences, I tethered again and ran several tests on my desktop from each of the two "mobile" sites (dslreports & mobilespeedtest). Once again, the results were very good, averaging ~900-1200 down, although a lot more variability than the Speakeasy site. I think the variability is attributable to the small file used in the mobile tests.

I then decided to try different handset browsers unthethered. Long story, short... no discernable difference using Picsel, Opera Mobile, or NetFront. All the results were bad.

What's going on here? Is this normal? My Rev-A data connection is obviously OK judging from the tethered results. But why can I not get decent results from the handset itself (where it counts)? Are there tweaks that need to be done on a PPC (similar to MTU and RWIN on PCs) to "tune" the device to the Rev-A network? Or is this a common mobile browser issue? Any expert insight would be greatly appreciated.
Reply With Quote