dslreports logo
 
    All Forums Hot Topics Gallery
spc
uniqs
23
BHNtechXpert
The One & Only
Premium Member
join:2006-02-16
Saint Petersburg, FL

BHNtechXpert to FactChecker

Premium Member

to FactChecker

Re: Shaping Happens

said by FactChecker:

Comcast shapes traffic using the PowerBoost technology.

If I am reading this right, it is saying that the "shaping" is giving better than provisioned speed for 70+% of the time. So the so called "small burst of speed" applies over and over again to more than 70% of your download AND upload usage. Sure that 2G file does not get "boosted" for the entire download, but video streaming does (since it grabs video buffers), gaming, email, web, most of your interactive traffic get's better than provisioned speed via Powerboost.

Contrary to popular myth Powerboost technology is only good for one thing...downloading small files...aside from that it's useless, pointless and actually can have a negative impact on applications that rely upon a consistent datarate such as streaming. It also impacts the results of speedtests, quality of service tests and the very same shaping tests as described in this article.

While I know many of you are yelling at TWC for shaping they aren't really shaping in the old Comcast way of things. Problem is their Turboboost is a form of shaping and is being detected as such and therefore the shaping label does apply to them. But if you are worried that TWC is doing the same thing as Comcast was and is the answer is NO.

As for the Turbo boost...10 years from now when we look back on Turbo boost we will all laugh at how stupid we were for even considering such a thing. Turbo boost makes it nearly impossible to gauge real-world results and is no friend to consumer or company alike. When it was first released it seemed like a nifty idea...but sometimes it doesn't end up working as expected.

***Happy Geek Pride Day***
FactChecker
Premium Member
join:2008-06-03

FactChecker

Premium Member

said by BHNtechXpert:

Contrary to popular myth Powerboost technology is only good for one thing...downloading small files...aside from that it's useless, pointless and actually can have a negative impact on applications that rely upon a consistent datarate such as streaming. It also impacts the results of speedtests, quality of service tests and the very same shaping tests as described in this article.

This is a bit of a glass is half (or totally) empty statement. Since the majority of almost all users' Internet experience is ALL based on small files Powerboost is far from useless. Even video streaming today is based on downloading small buffers of video every few min. The Netflix traffic profile fits perfectly into PB.
nysports4evr
Premium Member
join:2010-01-23

2 recommendations

nysports4evr to BHNtechXpert

Premium Member

to BHNtechXpert
I don't understand the complaint about PowerBoost technology in itself.

It's helpful in some cases -- faster initial buffering of audio/video, etc & downloading small files.

And there's really no negative -- you still get your provisioned rate after PowerBoost is over.

The only thing I'd like to see changed is a more clear advertising of their actual rates, and PowerBoost rates.
BHNtechXpert
The One & Only
Premium Member
join:2006-02-16
Saint Petersburg, FL

1 recommendation

BHNtechXpert to FactChecker

Premium Member

to FactChecker
said by FactChecker:

said by BHNtechXpert:

Contrary to popular myth Powerboost technology is only good for one thing...downloading small files...aside from that it's useless, pointless and actually can have a negative impact on applications that rely upon a consistent datarate such as streaming. It also impacts the results of speedtests, quality of service tests and the very same shaping tests as described in this article.

This is a bit of a glass is half (or totally) empty statement. Since the majority of almost all users' Internet experience is ALL based on small files Powerboost is far from useless. Even video streaming today is based on downloading small buffers of video every few min. The Netflix traffic profile fits perfectly into PB.

Actually no it doesn't because it isn't consistent. PB technology lasts on average the first 7-12 seconds of a specific session and depending on network conditions may not even engage at all. This also happens to be during the time which a streaming client establishes a connection and adjusts datarates based upon a specific the connections capacity at that moment. Only problem is once the PB window is passed the datarate suddenly drops thus forcing the client to adjust suddenly to the new datarate. I'm not quite sure who taught you otherwise....but they were wrong.

PB is no friend to streaming...period.

FFH5
Premium Member
join:2002-03-03
Tavistock NJ

FFH5 to BHNtechXpert

Premium Member

to BHNtechXpert
said by BHNtechXpert:

said by FactChecker:

Comcast shapes traffic using the PowerBoost technology.

If I am reading this right, it is saying that the "shaping" is giving better than provisioned speed for 70+% of the time. So the so called "small burst of speed" applies over and over again to more than 70% of your download AND upload usage. Sure that 2G file does not get "boosted" for the entire download, but video streaming does (since it grabs video buffers), gaming, email, web, most of your interactive traffic get's better than provisioned speed via Powerboost.

Contrary to popular myth Powerboost technology is only good for one thing...downloading small files...

And since that is 90% of the time a user spends online - that is browsing the Internet, it is of good use to most users.

poopiepants
join:2003-10-04
King, NC

1 recommendation

poopiepants to nysports4evr

Member

to nysports4evr
powerboost is a stupid gimmick who really cares?? all I care about is having my ISP catch up with the times and roll out faster tiers.
BHNtechXpert
The One & Only
Premium Member
join:2006-02-16
Saint Petersburg, FL

1 recommendation

BHNtechXpert to FFH5

Premium Member

to FFH5
said by FFH5:

And since that is 90% of the time a user spends online - that is browsing the Internet, it is of good use to most users.

No it's not because it only works for the first few seconds of the session. Meaning that once you exceed the PB window you are browsing at normal speeds and if you have any other internet traffic going then you will never see the PB effect at all because the first session gets the boost...after that it has no effect whatsoever.
BHNtechXpert

4 edits

1 recommendation

BHNtechXpert to nysports4evr

Premium Member

to nysports4evr
said by nysports4evr:

I don't understand the complaint about PowerBoost technology in itself.

It's helpful in some cases -- faster initial buffering of audio/video, etc & downloading small files.

And there's really no negative -- you still get your provisioned rate after PowerBoost is over.

The only thing I'd like to see changed is a more clear advertising of their actual rates, and PowerBoost rates.

Here's a simple expl of the benefits and disadvantages:

Benefits to the consumer:

1) Slight increase in download speeds for the first few seconds of the download (determined by bytes transferred not time) provided that the PB window was not already activated by another internet task.

Disadvantages to the consumer:

1). Makes it impossible for you to accurately test your connection speed with all browser based speed tests (with the exception of capacity tests and there is only one site offering such a thing) so no matter what you do your test results will always reflect the powerboost speed not your real world connection speed as none of the major testing sites test based on connection capacity.

2). Makes it impossible for you to accurately benchmark the QoS (Quality of Service) your provider is giving you because QoS uses consistency of speed as a major portion of the test and since you start off at a high datarate and suddenly drop off after the PB window your QoS rating is greatly affected and can skew the results either way (in favor of or against the consumer or provider).

3). Despite claims to the contrary PB impacts streaming applications greatly especially on slower systems where connection management overhead impacts system performance even on a good connection. Streaming is I/O intensive and on a slower system can be difficult under the best of conditions. Force the streaming application to constantly buffer or adjust itself because of significant datarate fluctuations and it can greatly impact those on slower systems.

4. PB lulls consumers into thinking they have a higher datarate than they actually do and can even mask hidden issues with a given connection because consumers in general don't know what to look for when they have problems.

Look if all you do is download a file now and again PB will benefit you. It will not benefit you at all for websurfing because in all cases the PB datarate exceeds the speed of your browser max tcp of 12-15mbps (single session) anyway. If you stream, game, p2p or any application where consistency of datarate is important it offers you no benefits whatsoever. It's great from a marketing standpoint and SOME users will benefit from it but the majority of power users will derive very little benefit especially after that 12 second ride.

Anyone who really needs the speed and consistency should consider a faster Docsis 3 based product and just dispense with the turbo. Your uploads with thank you...your downloads will thank you...your Netflix will really thank you...
FactChecker
Premium Member
join:2008-06-03

4 edits

1 recommendation

FactChecker

Premium Member

PB is like having a turbo V6 vs a V12. Same acceleration 0-60 as the V12, without the added cost. I'm OK if I can't keep the speed at 160MPH for 200 miles because I do more short trips every day.

You mention "first few seconds of the download", but also understand that home broadband needs are broken into many separate flows. Most interactive flows (according to this report over 70%) fit within those few seconds and get to utilize these higher "V12" data rates for the duration of the flow.

Your statement around the browser benefits conflict with not only how browsers work (multiple TCP streams), but what they are accurately reporting on speedtests for browser speeds.

I do agree that PB does not help much with non-interactive downloads of very large files or P2P, but that accounts for a very small portion of what most do and they normally don't care. People want their interactive experience to be a V12 at V6 turbo prices.

I bet you $1 if PB was turned off it would be far more of a negative to users than a positive as you state.

KrK
Heavy Artillery For The Little Guy
Premium Member
join:2000-01-17
Tulsa, OK

KrK to FFH5

Premium Member

to FFH5
It's important to consider the difference between small files and sessions. Small files on the same session = no benefit by PB.
BHNtechXpert
The One & Only
Premium Member
join:2006-02-16
Saint Petersburg, FL

4 edits

1 recommendation

BHNtechXpert to FactChecker

Premium Member

to FactChecker
said by FactChecker:

PB is like having a turbo V6 vs a V12. Same acceleration 0-60 as the V12, without the added cost. I'm OK if I can't keep the speed at 160MPH for 200 miles because I do more short trips every day.

You mention "first few seconds of the download", but also understand that home broadband needs are broken into many separate flows. Most interactive flows (according to this report over 70%) fit within those few seconds and get to utilize these higher "V12" data rates for the duration of the flow.

Your statement around the browser benefits conflict with not only how browsers work (multiple TCP streams), but what they are accurately reporting on speedtests for browser speeds.

I do agree that PB does not help much with non-interactive downloads of very large files or P2P, but that accounts for a very small portion of what most do and they normally don't care. People want their interactive experience to be a V12 at V6 turbo prices.

I bet you $1 if PB was turned off it would be far more of a negative to users than a positive as you state.

Your statement about multiple TCP streams and browsers is incorrect. The only time multiple browser sessions are opened is IF the page requires it OR you have multiple browser windows open. While many pages these days will in fact do this you also miss the point about how PB works. It's not multi-session aware. You don't get PB kicks for each session you get it by byte count. Once you reach that limit regardless of sessions you're done...PB turns off and you surf normally. The problem is that none of us know the specific byte count for each provider and this is the biggest reason that speed tests are incapable of working around it.

Look, we're talking about an issue that is VERY close to my heart here . I have worked long long long days in working around PB in an effort to make sure my tests give accurate results regardless of your connection type. PB has stymied all of us involved in broadband benchmarking for this reason alone. We can work around one or two providers but then you have all the others who are now also using PB technology with their own specs. And these specs tend to change which makes it even more difficult.

I will agree that for SOME there is a benefit to PB. But for the more net aware, net involved and net dependant this is not the case.

Now if providers want to standardize the byte count and increase the byte count to a reasonable size I might jump on board but I have a better solution....stop the speed game and get down to offering serious speeds like some local providers in my area at reasonable prices and just be done with it. Stop with the smoke and mirror show and just accept the fact that speed matters especially consistent speed when it comes to todays applications...and this isn't going away...

Until that time don't bite on the PB hype. If you subscribe to 12/1 plus TB accept the fact you have a 12/1 connection and enjoy it. The TB is nothing more than a small short term bonus..nothing more. If on the other hand you need 20/2 and you are counting on TB to make your 12/1 connection seem like a 20/2 connection your expectations are unrealistic and you need to rethink things.

(Notice I didn't say anywhere that TB is bad...just the benefits as they are communicated is well....exaggerated)
FactChecker
Premium Member
join:2008-06-03

FactChecker

Premium Member

I agree that purist long term speed measurement becomes more difficult, but that is less relevant in the real world of Internet application performance and user experience. Most pages do multiple TCP streams and according to the data in this report you can roughly judge how often the PB token bucket remains available for above provisioned speed downloads. It even breaks it down / provider.

I understand your passion in trying to break down a specific situation in broadband connections, but perhaps more real world application performance measurements is more relevant to most broadband customers' experience. e.g.

X MB objects - __ Mbps
Y MB objects - __ Mbps
Z MB objects - __ Mbps

NxX MB objects - __ Mbps
NxY MB objects - __ Mbps
NxZ MB objects - __ Mbps

X MB objects are used in email, web, etc, etc and make up approximately X % of the average users Internet experience
Y MB objects are used in ......
Z MB objects are used in ....

Primarily Web and email users would be most interested in X and make up X % of the broadband user base
Primarily P2P users would be most interested in Z

The reality is broadband experience is not only about sustained long term speeds. It is about application performance and how often the user is waiting for the Internet to respond.
BHNtechXpert
The One & Only
Premium Member
join:2006-02-16
Saint Petersburg, FL

BHNtechXpert

Premium Member

With all due respect I dont need the broadband 101 lecture. I'm intimately aware of the differences between pipe speed, capacity and application speed and happen to own the only site that makes all such tests available to the public. I certainly respect your desire to defend PB however your desire to defend the technology is clouding your better technical judgement. If you know as much as I think you do then you know I'm right......you really are wasting your time on me however. I dont dislike the concept of PB....just the way SOME of you are marketing it.
FactChecker
Premium Member
join:2008-06-03

1 recommendation

FactChecker

Premium Member

Fair enough, but many do need some Broadband 101 and clarifying this for others is the intent.

My main point is your position of : "[PB] is useless" or the con's by far outweigh the pro's for the majority of users, is what I dispute.

I do agree that for select situations it does not provide the added benefit over the provisioned speed. For most of us, those situations are non-interactive and fewer and far between IMNSHO.

FFH5
Premium Member
join:2002-03-03
Tavistock NJ

FFH5 to BHNtechXpert

Premium Member

to BHNtechXpert
said by BHNtechXpert:

happen to own the only site that makes all such tests available to the public.

Nice web site. A lot of testing tools all in one place.
»www.ispgeeks.com
BHNtechXpert
The One & Only
Premium Member
join:2006-02-16
Saint Petersburg, FL

1 recommendation

BHNtechXpert to FactChecker

Premium Member

to FactChecker
said by FactChecker:

Fair enough, but many do need some Broadband 101 and clarifying this for others is the intent.

My main point is your position of : "[PB] is useless" or the con's by far outweigh the pro's for the majority of users, is what I dispute.

I do agree that for select situations it does not provide the added benefit over the provisioned speed. For most of us, those situations are non-interactive and fewer and far between IMNSHO.

I'll say it again. I'm not anti-PB technology. I'm anti-misrepresentation of the benefits and capabilities of the technology and I'm really not very happy with the fact that PB technology inhibits a users ability to properly test and diagnose issues with their connection as it applies to application speed, capacity and QoS. There is a way around this issue and that is if providers were to publish their PB transfer limits or thresholds AND they be standard throughout the industry (so I can work around the PB threshold in my tests) ...neither of which is going to happen. And of course there's that pesky little issue about the marketing of the technology. Don't represent it to be anything more than it is...a little boost under specific circumstances.

Part of the the broadband industry problems have to do with communication to the consumer. There are a few providers out there who have exploited the consumers lack of understanding of these products and technology and as the consumers technical understanding has improved over the years they have suddenly realized they've been had. Consumers are angry and very distrustful of a few providers and unfortunately it's these few that have given a bad reputation to the industry in general (one especially comes to mind...Comcast). Thanks to Comcast anytime there is any perceived slowdown of a particular internet task the first thing people think is "my isp is throttling me" when in reality it has nothing to do with throttling at all.

It's a serious perception problem and the solution begins with providers accurately portraying their network capabilities and service offerings and that includes PB technology. That means in some cases they will need to reign in their marketing departments and screen all consumer communication for technical accuracy prior to passing it off to production for advertising. The providers who successfully accomplish this first and master the art of consumer education and communication as it applies to their broadband services will in the end win the hearts and trust of the consumer. As it stands right now...most consumers view broadband providers as the enemy and in some cases the reputation is deserved, in most however it isn't. Right now the good guys need to stand up and just prove it.
Mele20
Premium Member
join:2001-06-05
Hilo, HI

Mele20 to BHNtechXpert

Premium Member

to BHNtechXpert
said by BHNtechXpert:

The problem is that none of us know the specific byte count for each provider and this is the biggest reason that speed tests are incapable of working around it.

Look, we're talking about an issue that is VERY close to my heart here . I have worked long long long days in working around PB in an effort to make sure my tests give accurate results regardless of your connection type. PB has stymied all of us involved in broadband benchmarking for this reason alone. We can work around one or two providers but then you have all the others who are now also using PB technology with their own specs. And these specs tend to change which makes it even more difficult.

I'm wondering, from what you have said here, if Sam Knows has this specific byte count for each provider. SK claims that it has fully compensated for TWC's (and other provider's) Power Boost effects on download speed tests. It claims that it is very accurate in its test results. I don't buy SK's claims and I am disappointed that the FCC chose them to do the broadband testing. I consistently see SK results claiming my download speed is considerably higher than my cap of 5mbps. I don't like being a participant in a testing situation where the information gathered is, IMO, erroneous and makes the ISP look good when that is not necessarily the case at all.

I detest Power Boost and I don't see why, at the very least, we users cannot make our ISPs turn it off for our connection if that is what we want. It really screws up MySpeed/MyConnection tests now and they have always been the best of the various speed tests. I own software (MyConnection) that has been rendered mostly useless by TWC forcing Power Boost on me. I am far more interested in QOS than in some few seconds speed boost that I don't see anyway anytime I really need it as it is turned off during prime usage hours.

As for shaper probe, I don't understand why you think it is a joke. I've been using it for years and recall when it showed no shaping by TWC (my division was the last one to get Power Boost) on download. The shaping it shows now is accurate unless I am missing something.
BHNtechXpert
The One & Only
Premium Member
join:2006-02-16
Saint Petersburg, FL

BHNtechXpert

Premium Member

said by Mele20:

said by BHNtechXpert:

The problem is that none of us know the specific byte count for each provider and this is the biggest reason that speed tests are incapable of working around it.

Look, we're talking about an issue that is VERY close to my heart here . I have worked long long long days in working around PB in an effort to make sure my tests give accurate results regardless of your connection type. PB has stymied all of us involved in broadband benchmarking for this reason alone. We can work around one or two providers but then you have all the others who are now also using PB technology with their own specs. And these specs tend to change which makes it even more difficult.

I'm wondering, from what you have said here, if Sam Knows has this specific byte count for each provider. SK claims that it has fully compensated for TWC's (and other provider's) Power Boost effects on download speed tests. It claims that it is very accurate in its test results. I don't buy SK's claims and I am disappointed that the FCC chose them to do the broadband testing. I consistently see SK results claiming my download speed is considerably higher than my cap of 5mbps. I don't like being a participant in a testing situation where the information gathered is, IMO, erroneous and makes the ISP look good when that is not necessarily the case at all.

I detest Power Boost and I don't see why, at the very least, we users cannot make our ISPs turn it off for our connection if that is what we want. It really screws up MySpeed/MyConnection tests now and they have always been the best of the various speed tests. I own software (MyConnection) that has been rendered mostly useless by TWC forcing Power Boost on me. I am far more interested in QOS than in some few seconds speed boost that I don't see anyway anytime I really need it as it is turned off during prime usage hours.

As for shaper probe, I don't understand why you think it is a joke. I've been using it for years and recall when it showed no shaping by TWC (my division was the last one to get Power Boost) on download. The shaping it shows now is accurate unless I am missing something.

You raised a huge point and thank you. Sam Knows faces the very same issue that any other speed benchmarking site does when it comes to PB. This will (among other things) cause their test results to be inaccurate in two ways; overall speed and QoS. It's kind of interesting how this will work. It will inflate speed values in favor of ISP's and it decrease QoS results against ISP's.

Now a smart consumer will know that while speed is important, QoS is far more imortant. Let's put it this way...would you rather have a car that can go 80mph once in a while and then just 40mph the rest of the time OR would you prefer to have a car that consistently goes 65mph? The answer is obvious and while simplistic it clearly demonstrates how PB works because that's what you get.

Today's and future applications will depend more and more on consistency of datarate so it's important that you consider these things when making your buying decision. Now there is one thing I'm comfortable in saying; provided that you have some extra bandwidth to play with you can work around the PB effect BUT you will never be able to properly benchmark your connection with PB running on that connection. It will always favor the higher temporary datarate and your QoS will always be skewed in most cases rating a connection as bad or poor when it MAY actually be good or excellent.

When I refer to extra bandwidth to play with I mean if you originally were going to buy a 10/1+ turbo connection and your needs are actually on average 12mbps and you were thinking the TB would make up the difference think again. You are better off in lets say a 20/2 + turbo connection. Now keep this in mind with streaming on any turbo connection however, there will be a datarate recalibration when you exit the PB window and possibly some additional buffering and this may cause you problems on slower systems. And of course you will most certainly have issues properly benchmarking on any PB enhanced connection if accuracy if something thats important to you.

I'm glad to see that you have also selected the higher road in broadband benchmarking. The software and sites you are using hands down are the best and most accurate as it applies to benchmarking broadband. In full disclosure it's also the core software that I use on my site....
Mele20
Premium Member
join:2001-06-05
Hilo, HI

1 edit

Mele20

Premium Member

Yes, I was quite impressed when I first saw your site (learned about it in another earlier thread here you posted in) and saw you are using MySpeed (now MyConnection).

Ironically, my ISP Oceanic TWC bought MySpeed for us about 5-6 years ago partly on my recommendation. The speed test they had at that time was impossible for neighbor island users to use. I suggested MySpeed (over Justin's java test) as I was using it and was quite impressed. I was the first person to know when Oceanic got it up and going (the Support Manager called me to let me know it was live. He was envious because I got 4.90 (on 5mbps connection) on the test with 99% QOS. I was on the Big Island and he was on Oahu at Oceanic's headquarters and was getting a poorer result than I was on speed and QOS).

I was told that MySpeed was (I think still is) used internally on all RR gateways and servers throughout the national network and was considered the best test. For several years, tier 3 would ask what your QOS was on the test if you called with a problem. If you had, like I did, the Visualware software running a speed test every 10 minutes to San Jose they were far more interested in the QOS than the actual speed you were getting and that was the right way for a GOOD ISP to be. I got truck rolls solely on QOS and the interpretation thereof that I reported and I had Visualware support do lengthy tests using the San Jose server to see at what periods of time there was conjestion, drop in QOS, etc. (This was all before Visualware posted the excellent, in depth, reports/explanations/glossery of terms users can now get when doing just a web test at MyConnection (not even needing to buy the software- owning the speedware is really so you can test in the background every so many minutes for as long as you want. That, over a long period, gives an excellent view of your connection's performance).

Anyhow, when Oceanic added PowerBoost they got rid of the MySpeed test (QOS was going to now be messed up and I think they didn't want a lot of calls about it) and got a flash test instead which is totally worthless.

Unfortunately, I don't have the luxury of being able to purchase higher speeds (assuming I could afford it). We have Standard at 5/1 (which I have) and Turbo at 8/1 for $10 a month more. That is it. Since Powerboost is on Standard as well as Turbo there is no point really in getting Turbo as the increase in download speed is so small and the upload is the same for both Standard and Turbo. Oahu got an increase from 5 to 10 down in January for Standard (no change in up) and, for them, turbo went to 15/1. There is nothing else (as Turbo Extreme had to be withdrawn and no ETA as to when it will be back, last time I checked, and that is for Oahu only anyway). Neighbor islands must wait until early next year for the increase. I'd get DSL if it was available here. I feel frustrated that I have to put up with Power Boost. I don't have any choices like what you speak about but I understand your point.

I need to join your forum and post there as I can't do many of your tests. On IE6 (only thing I use it for is speed tests because I don't want Flash and Java on Fx and Opera) and when I try to do a capacity test, IE pops a warning about a doubleclick site (I use the Proxomitron so no ads) trying to infect the computer. I'm thinking that may be because you no longer support IE6 which I understand fully if that is the case. So, I tried on IE8 on a virtual machine and no IE warning, no doublclick sites, but I still couldn't do anything but the basic tests.
BHNtechXpert
The One & Only
Premium Member
join:2006-02-16
Saint Petersburg, FL

2 edits

BHNtechXpert

Premium Member

Yea we aren't going back to supporting IE6 and I can assure you that fear nothing from ads on the site. We don't run google adsense and in 3 years since java infected ads started popping up we've had only two issues where someone was able to circumvent our ad provider security but they were caught by our canary system. Simply put it's a dedicated box who's sole purpose is to auto refresh each ad driven page and to detect malformed ads and become infected. Once infected it fails which shuts down the ad server preventing mass infection a la banner ad. It's an old school approach to a new and evolving problem and the bad guys can't work around it. Since we only accept PPC ads vss PPM ads everyone is happy and could care less about the canary system doing its job protecting our users from infected banner ads.

As for your not being able to still use MCS for testing turbo sure you can. Just adjust your method of interpreting the data. First of all use the capacity test instead of the speed test for Turbo. Increase the payload size and time to test threshold. Look for loss prior to the rated max capacity for that connection and any detected loss at the 1/4, 1/2 or 3/4 points during each leg of the test and you know you have a problem. You only expect it to drop packets at max capacity not before.

Another thing you can do is turn off the mypseed applets display of QoS to the customer and go into the backend and review the tests manually. You will see the dataflow in the reports and can clearly see where TB turns off in the test and review the dataflow from that point onward. Any significant disruption in the dataflow, pause etc and that points to a problem AND it's a great way to catch people cheating and who are using wireless instead of a wired connection to test because wireless has a very pronounced signature. Even with TB and in the right hands and interpreting eye (and with a good understanding of how the MCS system works) you can still get reliable diagnostics. Just takes a bit more effort and the people you work with need to trust your judgement on the interpretation. All things considered it's still the best network diagnostics on the market even with TB enhanced connections.