dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
97
share rss forum feed


JohnInSJ
Premium
join:2003-09-22
Aptos, CA

2 edits
reply to sortofageek

Re: Here we go again about bandwidth throttling ...

Ok network professionals. You manage a network that can handle 1Mbit of traffic a second at the gateway.

You have 10 users paying equally for service. They can in total, generate 10mbit of traffic.

99% of the time, their sum total of usage is UNDER 1mbit

1% of the time, it is OVER 1mbit

Now, when it is over, there is no way on earth all the packets can go through, None. Nada. The pipe is limited by hardware.

Now, at this point, some packets will not get through, this is a fact. So, some will wait. You can apply QoS markers to indicate which packets have a lower priority, or you can let the system degrade all connections equally.

As a network engineer, you decide not to punish the 9 users who are using, in total, 10% of the pipe. Instead, you select the 1 user who is using 100% of the pipe, and you make 10% of his packets wait.

This is QoS. Plain and simple.

quote:
I am a networking professional and have been for over 20 years
Then surely you know this. Tell me you understand this. I need to hear it.
--
My place : »www.schettino.us


fuziwuzi
Not born yesterday
Premium
join:2005-07-01
Atlanta, GA

1 edit

said by JohnInSJ:

(blah blah blah deleted for brevity)

As a network engineer, you decide not to punish the 9 users who are using, in total, 10% of the pipe. Instead, you select the 1 user who is using 100% of the pipe, and you make 10% of his packets wait.

This is QoS. Plain and simple.

quote:
I am a networking professional and have been for over 20 years
Then surely you know this. Tell me you understand this. I need to hear it.
Now, using your analogy, what if 9 of the customers aren't even using their system, say it is after hours and they've shut down. But the 1 remaining user tries to do a data transfer using the full bandwidth of the system. Under the Comcast method, that user would still be "packet prioritized" and his EFFECTIVE SPEED for that transfer would be attenuated.

I've not made any value judgment, I'm not saying Comcast is wrong for doing this. My only complaint is of the marketing and "propaganda" that tries to obfuscate reality. I know all of the companies do this, I just wish someone would have the cojones to stop that practice and be upfront.
--
***************
I think computer viruses should count as life. I think it says something about human nature that the only form of life we have created so far is purely destructive. We've created life in our own image.
- Stephen Hawking


espaeth
Digital Plumber
Premium,MVM
join:2001-04-21
Minneapolis, MN
kudos:2
Reviews:
·Vitelity VOIP

1 edit

1 recommendation

said by fuziwuzi:

Now, using your analogy, what if 9 of the customers aren't even using their system, say it is after hours and they've shut down. But the 1 remaining user tries to do a data transfer using the full bandwidth of the system. Under the Comcast method, that user would still be "packet prioritized" and his EFFECTIVE SPEED for that transfer would be attenuated.
That's the fundamental misunderstanding here, because what you are stating in this example is not correct.

The congestion management system assigns a priority for packets being pulled out of the transmit buffer for when the SHARED channel becomes saturated. If there is no congestion, there is no queuing, and even if you mark packets down from Priority Best Effort to Best Effort there is absolutely no impact to throughput when the line is otherwise free of congestion.

It's entirely possible for you to run your connection at 100% of your subscribed line rate even with the congestion management system in effect if the traffic on your DOCSIS segment has the capacity to allow for it.

AVonGauss
Premium
join:2007-11-01
Boynton Beach, FL

2 edits
reply to fuziwuzi

said by fuziwuzi:

Now, using your analogy, what if 9 of the customers aren't even using their system, say it is after hours and they've shut down. But the 1 remaining user tries to do a data transfer using the full bandwidth of the system. Under the Comcast method, that user would still be "packet prioritized" and his EFFECTIVE SPEED for that transfer would be attenuated.
Except, as has been pointed out several times by myself and others, that is not how the system works. Even if your traffic has been prioritized as best effort (the lower of the two), if there is available overall bandwidth the effective transfer rate should not change. If you are experiencing real-world transfer issues, you really would be better off creating a thread and providing more details about the actual issue you are experiencing. If the system worked as you described above, this forum would be flooded with complaints.

Packet prioritization does not define arbitrary rates for transfers, it defines or shapes the effect of what happens when a physical or logical limit is reached, such as when the capacity of the DOCSIS channel is reached.


JohnInSJ
Premium
join:2003-09-22
Aptos, CA
reply to fuziwuzi

said by fuziwuzi:

Now, using your analogy, what if 9 of the customers aren't even using their system, say it is after hours and they've shut down. But the 1 remaining user tries to do a data transfer using the full bandwidth of the system. Under the Comcast method, that user would still be "packet prioritized" and his EFFECTIVE SPEED for that transfer would be attenuated.
FAIL

That doesn't happen. The node is not saturated.
This is The Key Point you seem to want to ignore.
--
My place : »www.schettino.us
Expand your moderator at work


funchords
Hello
Premium,MVM
join:2001-03-11
Yarmouth Port, MA
kudos:6

1 edit
reply to fuziwuzi

Re: Here we go again about bandwidth throttling ...

said by fuziwuzi:

Now, using your analogy, what if 9 of the customers aren't even using their system, say it is after hours and they've shut down. But the 1 remaining user tries to do a data transfer using the full bandwidth of the system. Under the Comcast method, that user would still be "packet prioritized" and his EFFECTIVE SPEED for that transfer would be attenuated.
One user can't drive the node's utilization to the 70% or 80% necessary to trigger Comcast's management. He can use 100% of his subscribed speed without triggering the management at all as long as the node isn't 70% utilized.

said by fuziwuzi:

I'm not saying Comcast is wrong for doing this. My only complaint is of the marketing and "propaganda" that tries to obfuscate reality. I know all of the companies do this, I just wish someone would have the cojones to stop that practice and be upfront.
I don't know of anyone else doing Comcast's method (Anyone hear anything?), although Sandvine markets it.

Even Comcast doesn't use it that often. Comcast said in its trials that it does happen and for the most affected users (the heaviest users), it kicks in about 1% of the time for about 15 minutes. Additionally, as it was pointed out to me in a a-hem moment, it's pretty doubtful that a lot of people will be on a node that is more than 70% of capacity and also be capable of maintaining speeds of more than 70% of their subscription. Longer queues and dropped packets are likely to keep their throughputs below that 70% trigger.
--
Robb Topolski -= funchords.com =- District of Columbia -- KJ7RL
Test your Broadband connection today! -- »measurementlab.net/


fuziwuzi
Not born yesterday
Premium
join:2005-07-01
Atlanta, GA
reply to JohnInSJ

said by JohnInSJ:

said by fuziwuzi:

Now, using your analogy, what if 9 of the customers aren't even using their system, say it is after hours and they've shut down. But the 1 remaining user tries to do a data transfer using the full bandwidth of the system. Under the Comcast method, that user would still be "packet prioritized" and his EFFECTIVE SPEED for that transfer would be attenuated.
FAIL

That doesn't happen. The node is not saturated.
This is The Key Point you seem to want to ignore.
Yes, it does happen. That's the point YOU ignore.
--
***************
I think computer viruses should count as life. I think it says something about human nature that the only form of life we have created so far is purely destructive. We've created life in our own image.
- Stephen Hawking


fuziwuzi
Not born yesterday
Premium
join:2005-07-01
Atlanta, GA
reply to funchords

said by funchords:

One user can't drive the node's utilization to the 70% or 80% necessary to trigger Comcast's management. He can use 100% of his subscribed speed without triggering the management at all as long as the node isn't 70% utilized.
Then my node must be horribly overutilized because I can assure you if I attempt a file transfer at any time of day or night using full bandwidth my speeds will be cut within a few minutes and will remain so. If I stop the transfer and wait, speeds will return to normal. If I manually limit my transfer speeds to about 60-65% of maximum, my the transfers complete as normal.

That is my continued observations of the Comcast system and has been for the several months that the "upgrade" has been implemented in my area (getting the 12/2 upgrade from 6/1). Comcast techs continue to say my line is "clean" and everything is functioning within specs.

So, then, my observations tell me one thing while you guys say something else. I believe my own eyes over propaganda.
--
***************
I think computer viruses should count as life. I think it says something about human nature that the only form of life we have created so far is purely destructive. We've created life in our own image.
- Stephen Hawking
Expand your moderator at work


EG
The wings of love
Premium
join:2006-11-18
Union, NJ
kudos:9
reply to fuziwuzi

Re: Here we go again about bandwidth throttling ...

said by fuziwuzi:

Yes, it does happen. That's the point YOU ignore.
W.I.W., node devices themselves do not get saturated. The upstream and downstream channels that are on one's segment can have capacity/congestion issues.

AVonGauss
Premium
join:2007-11-01
Boynton Beach, FL
reply to fuziwuzi

said by fuziwuzi:

Then my node must be horribly overutilized because I can assure you if I attempt a file transfer at any time of day or night using full bandwidth my speeds will be cut within a few minutes and will remain so. If I stop the transfer and wait, speeds will return to normal. If I manually limit my transfer speeds to about 60-65% of maximum, my the transfers complete as normal.

That is my continued observations of the Comcast system and has been for the several months that the "upgrade" has been implemented in my area (getting the 12/2 upgrade from 6/1). Comcast techs continue to say my line is "clean" and everything is functioning within specs.

So, then, my observations tell me one thing while you guys say something else. I believe my own eyes over propaganda.
Speaking of propaganda, you still haven't taken the 5 minutes to actually explain what you are doing and the problem / symptoms that you are seeing. From the vague references to the actual symptoms you have posted, it sounds more like you are trying to use a P2P program and are having trouble with the upstream which usually involves the items below:

1) Your router's NAT table is becoming overloaded.
2) Your upstream is becoming over-saturated.
3) You believe you're on 12/2, but are really still on 6/1 - see #2.

The way I see it; you can continue to think that we must be a bunch of morons that don't get it and are trying to spread evil corporate propaganda, or, if this is an actual problem that you would like solved it might be worth considering using a different approach.


JohnInSJ
Premium
join:2003-09-22
Aptos, CA

1 recommendation

reply to fuziwuzi

said by fuziwuzi:

So, then, my observations tell me one thing while you guys say something else. I believe my own eyes over propaganda.
You're the only one seeing this, or your area is terribly oversubscribed, or you have some technical problem.

Unless you are mistaking speedboost for your actual speed? Speedboost will drop out in a few seconds, you should see 1.4Mbit down on a 12/2 line consistently after speedboost.

I can run at my rated pipe speed for hours on end in a high density residential area.

Have you tried posting in the comcast help forum here?
--
My place : »www.schettino.us


funchords
Hello
Premium,MVM
join:2001-03-11
Yarmouth Port, MA
kudos:6

2 recommendations

reply to fuziwuzi

said by fuziwuzi:

if I attempt a file transfer at any time of day or night using full bandwidth my speeds will be cut within a few minutes
If "a few minutes" is less than 10, then you're not affected by Comcast's bandwidth management method.

said by fuziwuzi:

iSo, then, my observations tell me one thing while you guys say something else. I believe my own eyes over propaganda.
I believe you, but I don't do propaganda. Your problems are real. Do they jive with the descriptions of the causes?
--
Robb Topolski -= funchords.com =- District of Columbia -- KJ7RL
Test your Broadband connection today! -- »measurementlab.net/


jlivingood
Premium,VIP
join:2007-10-28
Philadelphia, PA
kudos:2

said by funchords:

I believe you, but I don't do propaganda. Your problems are real. Do they jive with the descriptions of the causes?
What I'd recommend is that fuziwuzi start a new thread and post all of his information so folks can assist in examining the issue. This means make/model of all equipment (modem, home gateway, etc.), OS of any devices on the network, OS of the specific device with the problem, signal levels from modem, highly detailed description of the problem (app in use, data transfer attempted with file size, traceroutes, time of day, etc.).

It'd be more fun IMHO to try to figure out what the real problem is here with fuziwuzi (just not in this thread of course).
--
JL
Comcast


fuziwuzi
Not born yesterday
Premium
join:2005-07-01
Atlanta, GA
reply to AVonGauss

said by AVonGauss:

From the vague references to the actual symptoms you have posted, it sounds more like you are trying to use a P2P program and are having trouble with the upstream which usually involves the items below:

1) Your router's NAT table is becoming overloaded.
2) Your upstream is becoming over-saturated.
3) You believe you're on 12/2, but are really still on 6/1 - see #2.

The way I see it; you can continue to think that we must be a bunch of morons that don't get it and are trying to spread evil corporate propaganda, or, if this is an actual problem that you would like solved it might be worth considering using a different approach.
You make a lot of inaccurate assumptions. The problem occurs whether using FTP or P2P, doesn't matter, anything involving a large file transfer. Problem occurs regardless of the presence of a router or not (i.e. plugging directly into the modem). Problem occurs with my own modem or a borrowed modem from the Comcast technician. Problem occurs regardless of operating system or computer (1 WinXP, 2 Win7 64-bit, 1 Macbook Pro). Comcast, via their CSRs and several onsite technicians say I have 12/2 service, are you saying they've lied to me?
--
***************
I think computer viruses should count as life. I think it says something about human nature that the only form of life we have created so far is purely destructive. We've created life in our own image.
- Stephen Hawking


fuziwuzi
Not born yesterday
Premium
join:2005-07-01
Atlanta, GA
reply to JohnInSJ

said by JohnInSJ:

Unless you are mistaking speedboost for your actual speed? Speedboost will drop out in a few seconds, you should see 1.4Mbit down on a 12/2 line consistently after speedboost.

I can run at my rated pipe speed for hours on end in a high density residential area.

Have you tried posting in the comcast help forum here?
I am quite aware of what Powerboost is and isn't, I'm not an idiot. And aren't you so lucky! I'm not that lucky here. The Comcast technicians who have been to my location say everything is working as designed. THEY told me the problem is actually how the system is designed, so in effect, it is working as it is supposed to, according to them.
--
***************
I think computer viruses should count as life. I think it says something about human nature that the only form of life we have created so far is purely destructive. We've created life in our own image.
- Stephen Hawking

AVonGauss
Premium
join:2007-11-01
Boynton Beach, FL
reply to fuziwuzi

said by fuziwuzi:

You make a lot of inaccurate assumptions.
Quite possible, which is why I and others have suggested to you to start a separate thread and post the details of your configuration and the details of what you are doing when you see the problem.

said by fuziwuzi:

The problem occurs whether using FTP or P2P, doesn't matter, anything involving a large file transfer. Problem occurs regardless of the presence of a router or not (i.e. plugging directly into the modem). Problem occurs with my own modem or a borrowed modem from the Comcast technician. Problem occurs regardless of operating system or computer (1 WinXP, 2 Win7 64-bit, 1 Macbook Pro).
That's a few more details, but it would be really good to start a new thread and put all the pieces together to give a complete picture.

said by fuziwuzi:

Comcast, via their CSRs and several onsite technicians say I have 12/2 service, are you saying they've lied to me?
No, that's not what I said and I think you know that. You can however confirm easily for yourself by downloading a large file such as a Linux ISO via a browser and look at the transfer rate after it has transfered about 60MB or so. If the download rate is 1.5 MB/s that would be 12/2 or if 1 MB/s that would 6/1.


JohnInSJ
Premium
join:2003-09-22
Aptos, CA
reply to fuziwuzi

said by fuziwuzi:

I am quite aware of what Powerboost is and isn't, I'm not an idiot.
Good, are you ready to try and fix your issue now?
quote:
And aren't you so lucky! I'm not that lucky here.
Good. Now we have established you're having a site-specific issue. I am not lucky, I am simply not having any service issues.
quote:
The Comcast technicians who have been to my location say everything is working as designed.
Were you seeing the performance issue during the tech visits? Did they reproduce the exact issue on their equipment? If not, then likely the tech was correct - as far as they could see, you were operating correctly at that moment.
THEY told me the problem is actually how the system is designed, so in effect, it is working as it is supposed to, according to them.
If all they had to go on was your description, then I would expect that would be the on-site tech's answer to why your system, which seemed to be working correctly, would experience slowdowns.

At this point all I can do is join the greek choir and suggest you start a new thread with detailed information and work with both the mass of people here and the comcast techs here to troubleshoot your issue. Because it is a service issue. Plain and simple. Really.
--
My place : »www.schettino.us


sortofageek
Runs from Clowns
Premium,Mod
join:2001-08-19
kudos:21

1 edit
reply to fuziwuzi

Please see your Instant Messages. Nobody is trying to insult you here. In fact, many are trying to help you.

Please take this to a new topic and allow others to assist you in troubleshooting. Unless that new topic makes it clear this is a Comcast congestion management issue, please don't continue in this huge thread.

For those following this issue, fuziwuzi See Profile has started a help topic here ---> »Throttling?

--
Join Team Helix * I am praying for these friends .



sortofageek
Runs from Clowns
Premium,Mod
join:2001-08-19
kudos:21
reply to JohnInSJ

(topic move) [OT] Just a little fun on the side ...

Moderator Action
The post that was here (and all 4 followups to it), has been moved to a new topic .. »[OT] Just a little fun with the Greek choir ...