DSL beats Comcast Internet in Montara
Posted: 19 May 2008 08:17 PM
Administrator
Avatar
RankRankRankRank
Total Posts:  129
Joined  2006-06-03

I just tried Comcast Internet with “PowerBoost” and it’s about 20% slower for my most speed-hungry applications for than is DSL from Coastside Net.

I really needed faster performance for uploading large volumes of photos and videos to Coastsider, so your mileage may vary. But in my tests here in Montara, I was able to upload at rates around 38 KBps on Comcast and above 45 KBps with DSL. Your mileage may vary.

The worst deal is Comcast’s so-called PowerBoost, which probably improves page loading speeds, but is so limited that they should be giving it away for free. It only increases speeds for the first 6 megabytes of an upload or download. This is something that is carefully obscured in their advertising and on their website.  How well hidden? Their customer service people don’t know that this is the limitation. Shame on me for not reading the mouse-type, I guess.

Profile
 
Posted: 19 May 2008 11:05 PM   [ # 1 ]
Member
RankRankRank
Total Posts:  60
Joined  2005-11-07

Barry,

Thanks, useful information.

“PowerBoost” sounds like a rebrand of the old Artera ‘Turbo’ Software with its negative characteristics especially on low compressible transfers. Trying compressed vs non-compressed file transfer should confirm. It sounds like it is designed for minimized impact of TCP slow start.

Image Compression:  ‘Lossy’ compression of JPG and GIF images?

It may work ok for low expectation web surfing. For a CDN provider Comcast that has little respect for image quality is not surprising.

Anyway, large FTP uploads are maddening regardless at the patience level V. ‘go out and kill something’.


Ken

Profile
 
Posted: 19 May 2008 11:12 PM   [ # 2 ]
Administrator
Avatar
RankRankRankRank
Total Posts:  129
Joined  2006-06-03

I don’t think it’s compression, because you get a nice increase in speed early on when uploading jpegs, which are (of course) already compressed.  But your speed rapidly degrades from crazy-fast to sluggish.

Profile
 
Posted: 20 May 2008 12:46 AM   [ # 3 ]
Newbie
Rank
Total Posts:  23
Joined  2006-04-23
Barry Parr - 20 May 2008 03:17 AM

I really needed faster performance for uploading large volumes of photos and videos to Coastsider, so your mileage may vary. But in my tests here in Montara, I was able to upload at rates around 38 Kbps on Comcast and above 45 Kbps with DSL.

Those numbers don’t sound possibly right. 38 Kbps == 38 Kilobits per second == 38 * 1024 bits per second, which is a little less than 5KB (kilobytes) per second. Surely you meant 38KBps? But even at that, with Comcast in El Granada, I get upload speeds in the T1 range—above 125KB (kilobytes) per second according to the San Francisco test at:

<http://www.speakeasy.net/speedtest>

I can’t test DSL even if I wanted to—last time I checked DSL *still* wasn’t available in El Granada.

Profile
 
Posted: 20 May 2008 10:07 AM   [ # 4 ]
Newbie
Avatar
Rank
Total Posts:  21
Joined  2006-09-24

These threads always inspire me to run another speed test. Using the tools at http://www.broadbandreports.com:

MegaPath - San Francisco: 3697 Kb/s down, 1743 Kb/s up

speakeasy - Palo Alto: 8934 Kb/s down, 1625 Kb/s up

Here in El Granada, I have Comcast HSI at the 6Mbps service level. According to the tech I chatted a few minutes ago, Powerboost was not available at the time of our chat. Also according to the tech, it’s active only “during certain non-peak hours when downloading large files.”  Yeah, they should be giving it away if that’s the case!

Deb

Profile
 
Posted: 20 May 2008 10:23 AM   [ # 5 ]
Administrator
Avatar
RankRankRankRank
Total Posts:  129
Joined  2006-06-03

Yes, I mean K-Bytes/sec, not bits.  I’ve corrected the post above.

However, I don’t trust the results from the benchmark sites. It’s too easy for the ISP to game the results. Also PowerBoost tends to overstate these results. My measurements are based on my personal real-world application, which typically performs a lot worse than the benchmark sites.

Profile
 
Posted: 20 May 2008 11:15 AM   [ # 6 ]
Newbie
Avatar
Rank
Total Posts:  21
Joined  2006-09-24
Barry Parr - 20 May 2008 05:23 PM

However, I don’t trust the results from the benchmark sites. It’s too easy for the ISP to game the results.

Perhaps that makes the benchmarks less useful for comparing one ISP to another, but they still provide a way to compare your service to that of others in the area. We had an issue with Comcast a couple of years ago that was specific to our drop. It’s hard to know if the performance you’re seeing is typical. Given that you can do a head to head comparison with DSL, which many of us can’t, it would be interesting to see how we compare on the benchmarks.

Deb

Profile
 
Posted: 21 May 2008 11:12 PM   [ # 7 ]
Newbie
Rank
Total Posts:  1
Joined  2006-05-24
Barry Parr - 20 May 2008 06:12 AM

I don’t think it’s compression, because you get a nice increase in speed early on when uploading jpegs, which are (of course) already compressed.  But your speed rapidly degrades from crazy-fast to sluggish.

I’d be splitting files one end and reassembling them the other, assuming the limitation is file-based. Or it is “continuous usage” based? If so, I wonder what the reset time is?? Can it be defeated by additional (simultaneous) connections? Probably not, but there’s always hope :-)

Profile
 
Posted: 21 May 2008 11:26 PM   [ # 8 ]
Newbie
Rank
Total Posts:  23
Joined  2006-04-23
Debbie Wolfe - 20 May 2008 05:07 PM

These threads always inspire me to run another speed test. Using the tools at http://www.broadbandreports.com:

MegaPath - San Francisco: 3697 Kb/s down, 1743 Kb/s up

speakeasy - Palo Alto: 8934 Kb/s down, 1625 Kb/s up

Here in El Granada, I have Comcast HSI at the 6Mbps service level. According to the tech I chatted a few minutes ago, Powerboost was not available at the time of our chat. Also according to the tech, it’s active only “during certain non-peak hours when downloading large files.”  Yeah, they should be giving it away if that’s the case!

Deb

I’m also in El Granada with the Comcast HSI and get quite different results:

MegaPath - San Francisco: 21463 Kb/s down, 1143 Kb/s up

speakeasy - Palo Alto: 16013 Kb/s down, 1095 Kb/s up

So I’m getting tremendously higher download rates and substantially lower upload rates. Very strange.

Profile
 
Posted: 22 May 2008 08:48 AM   [ # 9 ]
Newbie
Avatar
Rank
Total Posts:  21
Joined  2006-09-24
Brian Dantes - 22 May 2008 06:26 AM


So I’m getting tremendously higher download rates and substantially lower upload rates. Very strange.


I noticed you posted late in the evening (off-peak?) versus my mid-morning. Could it be Powerboost making the difference on the download side?

I suppose if we wanted reasonable comparisons, we’d have to plan a schedule of comparable tests, and identify all the rest of the variables. I don’t know if I’m enough of a geek to go to that much trouble. :)


Deb

Profile
 
Posted: 22 May 2008 11:14 PM   [ # 10 ]
Newbie
Rank
Total Posts:  1
Joined  2008-05-19

Greetings….. does anyone here have experience with high speed internet in Moss Beach specifically? I am moving in and need to decide between DSL and Comcast. Does anyone know the distance of DSL central office from end of California Street?  I need reliable service, high speeds are great, but reliability counts most of all.  Can’t have frequent outages… I’d love to hear from someone who is in Moss Beach….  Thanks!

Profile
 
Posted: 29 May 2008 12:19 AM   [ # 11 ]
Member
RankRankRank
Total Posts:  95
Joined  2004-10-05

I’m in El Granada.  I download orders of magnitude more than I upload, so I can’t speak too much to upload speeds.  However, it is important to know that because of the cable internet’s architecture, and the way that cable companies allocate bandwidth in the local system (end user to local head-end), they must cap upload speeds to avoid having individuals swamp the system.  There is much more bandwidth allocated to the download direction.  I haven’t tried to figure out when this works and when it doesn’t, and it’s greatly influenced by problems at the site that I download from, but I have seen downloads start out at over 1MBps and then drift downwards, stabilizing at 750KBps.  To my knowledge, you simply can’t get that on DSL without spending a huge amount of money.

That said, if reliability is a concern and you can get DSL, I suspect that you’ll be less unhappy with DSL.  Comcast’s problem is that none of their outages which are longer than a minute are shorter than 2 hours.  I.e., if it goes out and doesn’t come back pretty quick, it’s going to be down for 2 hours minimum.  They may have 99.99% uptime, but that’s still abysmal by phone company standards.

I wish I could get DSL here in El Granada.  I’d have both DSL and cable for redundancy.  Probably download on the cable and upload on the DSL.

In 2004 PacBell was contacted (not through regular customer channels) about a time frame for installing the equipment in E.G. for DSL.  They said they were planning on doing it in 2007.  Well, in about 2006 or so I found out that in 2004, “planned for 2007” was PacBellSpeak code for “never gonna happen.”

As I’ve written repeatedly (find my “DSL in El Granada” or somesuch poll elsewhere in Town Hall), the #1 thing that people can do to pressure the phone company for DSL in E.G. is to call up and ask to put on their waiting list.  And it’s probably necessary to call every 3 months and specifically ask “when is it going to be available?”.  We need a few hundred people to do this, and then we may get DSL here.  Uh oh, now that I think about they’ve stopped sending me the bimonthly “still not available” emails…

BTW, there *is* DSL in El Granada.  But there are only a handful of circuits and to avoid having to allow competitors access to it per FCC regs, they insist there is no DSL in E.G.  But I personally verified 2 DSL circuits which used to be installed in downtown El Granada.  (At two separate locations people told me they had DSL, I didn’t believe them, I went into their offices and traced the route which proved to me that it was DSL.)  I have not been able to track down the circuit numbers though after they were disconnected.  I don’t know how one got installed, the other was because the business owner had a friend in the right place in the phone company.

Profile
 
Posted: 29 May 2008 12:22 AM   [ # 12 ]
Member
RankRankRank
Total Posts:  95
Joined  2004-10-05

I just ran a speed test on testmy.net which showed 2/3 of average for Comcast, pretty slow.  I have liked testmy.net, but I just tried megapath and its numbers are much closer to what I expect:  5658 Kb/s down, 1859 Kb/s up.  In fact, that upload is significantly faster than I expect.

Barry, how’s the analog cable TV picture quality there?  If it’s worse than usual (as hard as that concept is to fathom), the node may be going bad.  They won’t replace it due to internet speed complaints, but if the picture quality is noticeably worse than in other nearby nodes, they probably will.

Profile
 
Posted: 29 May 2008 10:55 AM   [ # 13 ]
Administrator
Avatar
RankRankRankRank
Total Posts:  129
Joined  2006-06-03

Channels under 6 or so tend to be worse, but that seems to be an antenna problem. Otherwise, I have no complaints about analog tv. Except that the programming is miserable. But I can’t blame Comcast for that.

Profile