dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
1101
share rss forum feed

dlnickel

join:2009-02-17
Los Angeles, CA

[ADSL2] George, Having trouble with the new Fusion Service

George,

Several days now with the Fusion Service, wild fluctuations on the DSL speed side. Tier 2 has been mostly great, but we have been unable to isolate the issue. The speeds are no where close to the "up to" speeds. The standard DSL was usually about 60 -70% of the published speed. Fusion is about 25 - 28% at best. Not real happy at the moment. Would you like me to send you my spreadsheet of time and date noted speeds (Speakeasy & DSL Reports). Or just publish them here?

Thanks,



NormanS
I gave her time to steal my mind away
Premium,MVM
join:2001-02-14
San Jose, CA
kudos:11
Reviews:
·SONIC.NET
·Pacific Bell - SBC

2 edits

How does your connection compare with this chart?

»New Fusion average speed/distance chart

I maintain two Sonic.net Fusion lines:


On ~9,156 foot loop.



On ~3,300 foot loop.








Sonic.net, LLC does not offer Fusion tiers. Speed is "up to" whatever the modem and DSLAM can settle for any given length of copper. If your speed is near the max for your loop length, I doubt if it can be improved much. Only if it falls short of the prediction, based on the linked chart, will improvement be possible.

I am pretty sure DSLX provides it the same way, because they are reselling it.

And if you already knew that, my apologies. I couldn't tell from your problem state if you did.

P.S. The published "up to" speed of 20 Mbps is ~85% of the theoretical ADSL2+ maximum of 24 Mbps; this is due to ATM overhead. What you will actually get is ~0.85xSync.

Looking at the "Speed/Distance" chart, at 9,156 feet, the speed curve intersects the distance at a little short of 6 MB; consistent with my modem sync rate. While my 5.05 Mbps on the long loop is only 25% of the advertised "up to" speed, it is very close to 0.85x5959, as expected for an ATM circuit.

BTW, that 5.05 MB speed is just about double what AT&T delivered on the same copper pair. At about half the AT&T price.
--
Norman
~Oh Lord, why have you come
~To Konnyu, with the Lion and the Drum


dslx_nick
Premium,VIP
join:2011-12-24
Chatsworth, CA
kudos:27
reply to dlnickel

George isn't in the office today, but if you want you can send those speed recordings to me.

Unfortunately, I think I know what the real issue is. We both already know from past experience that your circuit is very far out, at the edge of where we can offer 1.5 on traditional DSL, so I'm not surprised that your speed is on the low end of the scale. Our Fusion service is uncapped, which means loop length and physical line quality are the factors which would be keeping your speed down - and we already know your loop length is very long. The maximum speed we'll be able to deliver to your location, based on current physical characteristics, is about 3 Mbps - which is certainly an optimized reading, given that your line would strain to reach 1.5 on traditional DSL at that distance. On a particularly good day, the bit rate is able to sync up to higher speeds (I'm seeing 3.4 right now), but that isn't a stable benchmark that I expect it'll be able to maintain; you most likely will see fluctuations below that.

One word of practical advice I can offer: if you notice the speed has dropped down significantly below where you expect it to be (3 Mbps or so), try powercycling the modem. When the modem originally syncs up, it tries to maintain its current speed and settings, only gradually adjusting its settings to match changes in the line quality. If/when the line quality fluctuates drastically for the worse, that can cause the line to 'over-extend' itself, causing it to try to maintain a speed that it cannot quite handle under its then-current condition. Much like a car going uphill vs. going downhill without changing gears or adjusting how hard you press on the gas pedal, this can cause speed changes and even problems; you may notice the speed gets choppy or you may get disconnected intermittently. If you get disconnected completely, the modem will then adjust its settings to re-sync to the current (worse) line conditions - just like a car changing gear when going up a hill. Once the line quality improves, the modem will not immediately jump back to its former, higher speed settings - but powercycling it will force it to re-evaluate the line and adjust its settings accordingly, and may get it back to its previous, higher speed faster. So, I don't recommend powercyling it frequently, but if you run into a rough patch, a quick powercycle might kick it back into a 'high gear'.


dlnickel

join:2009-02-17
Los Angeles, CA
reply to NormanS

Hi Norman,

Thanks for the note. I am a layman and don't have access to the detailed readings you have posted. I am using the DSL Reports speed test and both Speakeasy Speed Test as well as Speakeasy Speed Test Plus.

The highest rate I have achieved was immediately after the service switched from my 1.5 DSL Extreme Service to the DSL Extreme / Sonic Fusion Service.
This was a reading of 3400kbs down and 995 kbs up that was using DSL Reports Flash test which apparently gives a lower line reading than Speakeasy.

Thinking that the service might be adjusted higher (given that my previous DSL service was 70 - 85% of the up to 1.5mbs) I contacted DSL Extreme. After multiple calls over multiple days the highest we have achieved is just shy of 3.0 mbs. The speeds reached a low of 43kbs download on 2/24.

The readings Speakeasy Speed Test Plus are pretty consistent so far today mostly 2.92mbs down and .81 - .84 upload with 0 - 0.4% packet loss, latency 44 - 73ms and 1 - 2ms of jitter.

I am pretty far out from the switching station I can't remember how many yards. But I am getting no where near the readings that you are achieving on the 9,000 foot loop.



dslx_nick
Premium,VIP
join:2011-12-24
Chatsworth, CA
kudos:27

About 13,400'.


dlnickel

join:2009-02-17
Los Angeles, CA
reply to dslx_nick

Thanks Nick,

By the way, would you let me know the actual distance to the switching station or the length of the loop?

The main trunk was replaced on this street about 6 years ago, so the copper is quite new. Both inside lines test well according to AT&T. I am curious as to what any of my neighbors closer to the switching source are getting on this service and does it take weeks for them to have stable service?

Unfortunately their seems to be a stark difference between the delivered speeds versus advertised. That is not a technical issue of course and I appreciate the help I have received from the tier 2 support in trying to stabilize the signal and wring out the best speed possible.

I've known since my initial signup with DSL Extreme that I am too far out to receive the higher speed services (ie anything above 1.5). Yet DSL Extreme sent me marketing materials which stated that I was "pre-approved" for this higher speed service. My expectation is, when pre-approved, that there might be an actual chance that someone checked the area before marketing.

Given that I have been achieving 70 - 85% of the 1.5mbs for four years, my expectation was that I should certainly achieve similar speeds with ADSL 2, somewhere in the 7 - 8.5mbs range rather than 2.9mbs. That is quite a significant difference. Sort of like ordering an "up to 8oz" hamburger and instead receiving a dry shriveled 3 oz patty with a bite already taken out.

From a consumer point of view, the increased speed claims are downright misleading. The fact that I am paying significantly more for DSL and receiving slower service than the standard DSL packages of 3.0 & 7.0 mbs is disappointing.

I am hoping that the signal will stabilize quicker than the 3 weeks it took back in 2009 when I switched to DSL Extreme as the experience is time consuming and quite frustrating. Particularly to my wife. And as the "Information Technology guy" in this residence, I take the heat.

Attached are the speeds I have recorded beginning 2/21/13.


dslx_nick
Premium,VIP
join:2011-12-24
Chatsworth, CA
kudos:27

That's the approximate distance from your MPOE to the Central Office (loop length). I don't have any other numbers, I'm afraid; the rest of that information is proprietary AT&T data which I don't have access to.

I'm sorry, but we're not advertising anything other than what we're delivering. We state, often and clearly, that your Fusion speeds are up to whatever your line can physically support. We do not have any minimum speeds, nor do we have any 'tiers' or 'brackets' with artificially-inflated prices. We're offering Fusion service at a flat rate: no arbitrary speed caps, no throttling, no data caps, no quotas, no overage fees. You knew, before signing up, that your loop length was quite far; we never pretended otherwise. You are able to get speeds with Fusion which are about double what you were able to get on regular DSL.

Your hamburger analogy doesn't really work; backwards, even. No one sells "up to X weight" burgers - they sell burgers which start with a certain minimum weight of uncooked meat, and then you wind up with LESS than that after cooking. An "8oz." burger is never actually 8oz. once cooked. A better food analogy would be an "All you can eat buffet" where the only restriction is the amount of food you can physically fit in your body - which means that people with larger bellies do find it a better bargain than people with smaller stomachs. Unfortunately, in this case your line falls in the latter camp; you WANT to stuff more data in, but your line physically CAN'T handle as much as you want, I'm afraid. However, even your previous 1.5 line is now running twice as fast, despite being over 13,000' out - and if you look at the chart NormanS linked, most customers at your distance are able to get about 2 Mbps on average, so that new copper line and good inside wiring you mentioned are certainly contributing to the 3 Mbps you're seeing.

The Fusion package does cost more than our 1.5 Mbps DSL package, yes - but Fusion also includes the phone built-in. In most cases, if you can get Fusion, the price is cheaper than the combined cost of phone + comparable DSL.

I looked over the list of speeds you posted, and do see the fluctuations there. I've also compared them to the actual readings from your modem itself. The gross bitrate on the downstream is fluctuating between about 2.7 and 3.4 for the most part; the times where it's dropping below that, the line is straining, but speeds have picked back up once it recovered. It's actually at 3.5 this very second, but I don't expect it to hold that forever. Where necessary, we CAN implement a cap to the speed to help stabilize the line (if it keeps overextending itself and causing the circuit to lose sync), but we don't do that by default, and don't do it to try to restrict or penalize the customer - and according to our logs, your line has been running unrestricted. If your line is repeatedly crashing and causing you to lose connection entirely, then let me know and I can see if that would help, but otherwise we weren't planning on touching your max speed; we just let it run as fast as it can go.



NormanS
I gave her time to steal my mind away
Premium,MVM
join:2001-02-14
San Jose, CA
kudos:11
Reviews:
·SONIC.NET
·Pacific Bell - SBC
reply to dlnickel

said by dlnickel:

Thanks for the note. I am a layman and don't have access to the detailed readings you have posted. I am using the DSL Reports speed test and both Speakeasy Speed Test as well as Speakeasy Speed Test Plus.

Actually, you likely can access those details. Most current DSL modems have a GUI to those details.

The highest rate I have achieved was immediately after the service switched from my 1.5 DSL Extreme Service to the DSL Extreme / Sonic Fusion Service.

My experience, as well. That is because the DSLAM starts out "optimistically", adjusting downward as it accumulates a "history" of line performance. Both of my lines showed a 2% to 4% drop from the initial rates to the current rates. Both are about [0.85*sync], which is consistent with an ATM circuit. The slower service, on the longer loop is about twice as fast as the previous AT&T ADSL service (the faster service was ordered as Fusion out of the gate).

This was a reading of 3400kbs down and 995 kbs up that was using DSL Reports Flash test which apparently gives a lower line reading than Speakeasy.

I believe that DSLX Fusion is Sonic.net Fusion, resold; I am sure the DSLX guys can clarify. In which case there are a couple of threads about Fusion in the Sonic.net forum. As long as you remember that you contact DSLX, not Sonic.net, for any problems.

DSL technology was designed to be adaptive, and does not require tiers. It is the ILECs who introduced tiers; partly to avoid having to explain why "A" gets 15 Mbps (my residence) while "B" only gets 5 Mbps (my mother's residence), but mostly because the lowest tier can be priced at [1.15*cost] while the highest tier can be priced at [1.75*cost].

A question occurred to me just as I pressed "Send": How does one realistically advertise a service with widely variable speed and no tiers?

--
Norman
~Oh Lord, why have you come
~To Konnyu, with the Lion and the Drum

dlnickel

join:2009-02-17
Los Angeles, CA

Norman,

Thanks again for the reply. Had to look up GUI (told you I was a layman). If my D-Link DSL-2320B has a Graphic User Interface, I have never been informed by DSL Extreme and have not known to search for a source to download.

DSL Extreme appears to be a reseller of Sonic.net. I did contact Sonic regarding multiple hard wire phone lines as DSL Extreme staff are ramping up on getting their heads around the service. That has been my experience with sales, billing and tech support. Once again kudos to tier 2 support, which is really the reason I stick with DSL Extreme.

While I have had major issues with DSL Extreme, particularly after their acquisition by Ikano and his out sourcing experiment. The US staff are really quite good.

My major beef at the moment is data delivery. If I was receiving 70 - 85% of the advertised speed on a 1.5mbs service how can a consumer (not working in the information technology field) not expect the same percentage on an upgraded service?

I was introduces years ago (after signing up) on the 20% less than advertised rule regarding data. I am assuming this is some sort of signal degradation over distance, just as electrical power lines energy over distances.

But a mere doubling of speed when 10,000 is the grail seems crap.

I appreciate your time and input.



NormanS
I gave her time to steal my mind away
Premium,MVM
join:2001-02-14
San Jose, CA
kudos:11
Reviews:
·SONIC.NET
·Pacific Bell - SBC

1 recommendation

said by dlnickel:

My major beef at the moment is data delivery. If I was receiving 70 - 85% of the advertised speed on a 1.5mbs service how can a consumer (not working in the information technology field) not expect the same percentage on an upgraded service?

This isn't just IT stuff, it is, also, ET stuff (Electronics Technology). And, yeah, I do have the advantage of having worked for HP as a board repair tech.

I was introduces years ago (after signing up) on the 20% less than advertised rule regarding data. I am assuming this is some sort of signal degradation over distance, just as electrical power lines energy over distances.

The "20% Rule" is not caused by "ET" stuff; it is the result of using Asynchrous Transfer Mode for moving data from the DSLAM to the aggregation router. ATM requires wrapping Ethernet packets into an ATM cell, which adds bits to the amount of data moved; this accounts for ~15% of the data moved to the aggregation router, but is not visible to speed test servers beyond the first TCP/IP link. You see, and the provider counts toward caps, data moving at the modem sync rate; the speed test server sees data moving at TCP/IP rate [modem_sync_rate less ATM bits].

But a mere doubling of speed when 10,000 is the grail seems crap.

This is the ET stuff. Any given length of copper over a couple of feet long is less than a perfect conductor. And at the thousands of feet in length, while carrying RF (DSL modulates RF with digital data), copper is most accurately modeled as a complex, parallel-series LCR ladder. "L" for inductance (as in coils), "C" for capacitance, and "R" for resistance. In such an environment, RF attenuates over distance.

Sonic.net puts the much simplified version in their FAQ:

»www.sonic.net/solutions/home/int···/fusion/

DSL Extreme puts it right up front on their Fusion page:

»www.dslextreme.com/fusionbroadband
quote:
Enjoy download speeds up to 20Mbps (based on distance).

I appreciate your time and input.

Thank you. I can appreciate the difficulty of marketing a product with no tiers, and limitations due to the physics of electricity against services which sell by tiers, and downplay the physics of electricity. It isn't easy to market such a product to a public which has no need to know/understand the technical limitations.
--
Norman
~Oh Lord, why have you come
~To Konnyu, with the Lion and the Drum


dslx_nick
Premium,VIP
join:2011-12-24
Chatsworth, CA
kudos:27

NormanS did an awesome job of nailing it on the head. There's a useful FAQ on this site which explains the 80% rule which I often recommend: »DSL FAQ »What is the 80% bandwidth / speed


dlnickel

join:2009-02-17
Los Angeles, CA
reply to dslx_nick

Hey Nick,

What you are delivering as of 12:19 AM on Feb. 28 is a whopping 0.12 on the download. 12:23 AM gives me a huge improvement, 0.17mbs.

I'll get back to the misleading advertising in another post, but dude. My readings are all over the map and 0.17 is far, far, far less than what I received on my 1.5 DSL Extreme service. Going back to my original post, I experienced pretty much the same swings when I first signed up. It took 3 weeks plus to get things to settle down.

So now what?


dlnickel

join:2009-02-17
Los Angeles, CA
reply to NormanS

Norman,

In reply to your query:

"A question occurred to me just as I pressed "Send": How does one realistically advertise a service with widely variable speed and no tiers?"

I've always found honesty a fairly good approach. Here is an example of honest advertising; "Your speeds will be faster." Another might be "Your speeds will increase." Another might be "We suck less than your current provider." Anyone of those would be fairly true and far closer to the truth than "up to 10mbs ****). It is patently misleading to say anyone will achieve full bandwidth due to degradation.

Since DSL Extreme, nor any of the big boys actually be able to predict speeds on even "pre approved" accounts, why set up false expectations? I am actually 9,705 feet from the Pac Bell/SBC/AT&T switching station in Boyle Heights (North East Los Angeles) CApitol was the old designation for this area. I'm waiting on a confirmation from an AT&T foreman on this. But I have always been quoted that my line is under 10,000 feet from the switching station (both AT&T and DSL Extreme). So i'm curious how my line suddenly jumped up to 14,300 feet in this thread. I never would have gotten 1.5 service at that distance from the switching station.

Anyway, enough of my chewing your ear. Thanks for the links. I will look at them. At least when my DSL Extreme Service comes back to above dial up (0.09 download, .65 upload, 0% packet loss, 44ms latency & 0ms jitter as of 12:48 am).

--



dslx_nick
Premium,VIP
join:2011-12-24
Chatsworth, CA
kudos:27

We find honesty to be the best approach, too. That's why we tell you, openly and frequently, that there are no minimum speeds and your actual speed will vary based on your distance. Saying "Your speeds will be faster" would actually be dangerous because we can't guarantee that speeds WILL be faster; if you were already paying for a certain speed, and that's the absolute limit of what your line can support, and you're too far out for the ADSL2+ tone range to make any appreciable difference, then you may not experience a significant speed increase. So while the customers who are decently close to their CO usually get a significant increase in speeds (I've seen x2 to over x15, depending on what speed package they were signed up for before), we can't guarantee the same increase in speed to everyone.

Now, understandably, some customers won't be satisfied with the speed increases they DO get. Some are happy to have double or more speed, usually at a reduced total cost single bill compared to their previous DSL + phone bills. That's why we offer the 30-day guarantee: if you're not satisfied, we're not going to hold you to it. We'll cancel the annual contract, no ETF, if you try the service and it doesn't work out for you and you want out.

In a more ideal world, it would be nice to be able to precisely predict for each user, "Your speed will be X." Unfortunately, it's just not technically possible right now, so the best we can do is to be honest up front and say, "It's gonna vary. Closer you are, higher speed you're going to get, with 20 Mbps being 100% ideal connection." In your case, your line can't run at 100% ideal; it's closer to about 10% ideal, due to your distance. We're pushing it by trying to make it run at about 15% ideal. That might be pushing the line too hard, which would explain the heavy fluctuations. Although I'm reluctant to do so, we might HAVE to introduce a limiter on your line's speed, to stabilize the line and keep it from overextending itself, if that's all the line can reliably handle.

I said your line is about 13,400' out from the CO (not 14,300'), but there's nothing sudden about it; that's from a field technician we sent out to your location back on 2/15/09, when you were trying to get a 3 Mbps traditional DSL package with us and we had to downgrade you to 1.5 Mbps because the distance was too far to support 3 Mbps. Unless AT&T did some re-routing to shorten the line that they didn't tell us about, that number should still be accurate. AT&T's records aren't always accurate, however, and those are usually what our technicians and prequal system check when trying to determine service availability and such (our technicians don't check back and read through your entire ticket history over the past four years when you call in with a CURRENT problem, and rightly so; it'd be a waste of time) - so if AT&T's records are currently indicating about 9,700' rather than 13,400', then either there was some loop length (bridge tap) removed that they didn't tell us about, or their records are inaccurate. I'm willing to accept either possibility.

Regardless of whether the exact numbers are 9.7kft or 13.4kft, however, your line quality is what it is. I do see some trouble on the line last night, and I see the troubleshooting done with our technicians last night. I also see that one of them has sent out a new modem for you to try, in case it's a problem with the modem. Honestly, I'm not expecting significantly improved connectivity with the new modem (I don't think it's a problem with the modem, based on what I've observed), but I'm totally willing to be proven wrong if that helps your connection stability. Right now, your line's running about 2.1Mbps down; a powercycle now might be able to kick it up to about 2.5, but it wouldn't hit 3 right now. It's looking like your line can't quite run at a full 3Mbps down stably, but I'll postpone judgment until the new modem arrives and you have a chance to test it out.