letter from charter

letter from charter

Post by br.. » Sun, 15 Oct 2000 04:00:00



Hi we received this email from charter when I wrote complaining
because are download speeds drop to under100k every vening thought
someone could pick it apart and give me good technical arguements to
argue with them

Brian

Greetings,

Thank you for contacting the Online Technical Support group.

There can be a number of factors that influence a cable modems speed,
some
of which are:

- As the number of cable modem users increase at any given time,
everyone's
download speeds drop, as the cable line is a shared medium.
- Download speeds can be influenced by your CPU's "power" and how much
memory your computer has.
- Some issues that are beyond our control, temperature, how a remote
server
is doing; whether it is busy or if there are problems in between you
and the
server you are trying to reach.
- Local issues to a users system:
    - Has the registry been changed to influence the TCP/IP stack in
any
way?
    - Is there any software corruption on the system?
    - Is the cable modem operating okay?

Thank you,

 
 
 

letter from charter

Post by Scott Nold » Sun, 15 Oct 2000 04:00:00



> Thank you for contacting the Online Technical Support group.

> There can be a number of factors that influence a cable modems speed,
> some of which are:

> - As the number of cable modem users increase at any given time,
> everyone's download speeds drop, as the cable line is a shared medium.

True, but a 1 million people shouldn't be put on one line.

Quote:> - Download speeds can be influenced by your CPU's "power" and how much
> memory your computer has.

A low end 486 can push 1.5 Mb/s easily.

Quote:> - Some issues that are beyond our control, temperature, how a remote
> server is doing; whether it is busy or if there are problems in between you
> and the server you are trying to reach.

And how they're equipment is functioning.  Is the traffic bottlenecking
through a single router and this router is * itself to death?

Quote:> - Local issues to a users system:

  - Has the registry been changed to influence the TCP/IP stack in any
way?

My computer doesn't use a registry and I've seen low bandwith transfers
over my line.

Quote:>     - Is there any software corruption on the system?

Not on the filesystem I use.

Quote:>     - Is the cable modem operating okay?

Hell yeah.  Especially when the dhcp server is working.

Quote:

> Thank you,

Generally tech service is reluctant to explain any oopses in their own
system.  If your system stays online and the bandwidth is crappy all of
a sudden, it's time to do some traceroutes and see who's equipment is
causing the slowdown.  Rarely (< 2%) should your computer be at fault.
I'd be suspect of your home ccomputer only if you installed network
related software and it was a bad install and more than the network
stopped working.

My $.02.  There is so much equipment between your keyboard and the wep
page you want to see which leaves so many possibilities as to why
traffic slows.  Your best friend is a traceroute at different times of
the day to the same place.  Check the routing each time and you may see
where the bottleneck is.

- Scott
--
Never do Windows again with  |  Scott M. Nolde

glaze!                       |  
1:05pm up 3:36, 3 users, load average: 1.13, 1.05, 1.03

 
 
 

letter from charter

Post by Bryan Bigger » Sun, 15 Oct 2000 04:00:00


How did you ever find anyone at Charter to write to? Bryan

> Hi we received this email from charter when I wrote complaining
> because are download speeds drop to under100k every vening thought
> someone could pick it apart and give me good technical arguements to
> argue with them

> Brian

> Greetings,

> Thank you for contacting the Online Technical Support group.

> There can be a number of factors that influence a cable modems speed,
> some
> of which are:

> - As the number of cable modem users increase at any given time,
> everyone's
> download speeds drop, as the cable line is a shared medium.
> - Download speeds can be influenced by your CPU's "power" and how much
> memory your computer has.
> - Some issues that are beyond our control, temperature, how a remote
> server
> is doing; whether it is busy or if there are problems in between you
> and the
> server you are trying to reach.
> - Local issues to a users system:
>     - Has the registry been changed to influence the TCP/IP stack in
> any
> way?
>     - Is there any software corruption on the system?
>     - Is the cable modem operating okay?

> Thank you,

 
 
 

letter from charter

Post by br.. » Sun, 15 Oct 2000 04:00:00


On Sat, 14 Oct 2000 14:01:31 -0500, Bryan Biggers


>How did you ever find anyone at Charter to write to? Bryan

on there website they have a contact us form filed that out and then I
have been replying to the email they sent me
 
 
 

letter from charter

Post by Ergo » Thu, 19 Oct 2000 11:48:33





>> Thank you for contacting the Online Technical Support group.

>> There can be a number of factors that influence a cable modems speed,
>> some of which are:

>> - As the number of cable modem users increase at any given time,
>> everyone's download speeds drop, as the cable line is a shared medium.

>True, but a 1 million people shouldn't be put on one line.

in my experience, the number of people on a "single line" (as you put it)
with cable in the most highly penetrated areas is about 500HH per node *
about 20% penetration = 100 HH; most MSOs at that point would allocate 2
full 6MHz channels for downstream, but suppose not - a single 6 MHz
channels has a capacity of about 40 Mbps, so that means an average
throughput of about 400 kbps per HH, assuming everybody downloads or
streams at the same time. More reasonable assumptions are max 50% users
logged in at peak time, and only streaming or downloading no more than 50%
of the time, so about 1,600 kbps - hence the typical approach to
engineering of the cable modem network suggests there is no problem there.
Places where there is a problem are nodes that are much heavier in HH (very
old networks, they are typically being eliminated, but it takes a bit of
time), or places where users have unconstrained (non DOCSIS) modems and use
them for banks of applications (garage ISPs...), or places where
penetration is much. much higher (few such places still)...
When penetration grows, the solution is to "split the node", i.e. move from
1 500 HH node to 2 250 HH nodes, 4 125 HH nodes; also, go from one 6MHz
channel to 2 or more; it is not a very complex or expensive process, simply
takes a bit of time and MSOs have been busy with turning networks on first.
Note that DSL is also a shared medium, typically using ratios of 10 to 1
all the way to 25 to 1 (some even more) for the multiplexing beyond the
Local Office... do the math, you'll see it does not compare favorably once
penetration arrives at about 20%...
Critiscism is easy... cut these guys some slack, they work hard at making
the product available broadly and at a very competitive value...
Cheers
 
 
 

letter from charter

Post by Scott Nold » Thu, 19 Oct 2000 04:00:00






> >> Thank you for contacting the Online Technical Support group.

> >> There can be a number of factors that influence a cable modems speed,
> >> some of which are:

> >> - As the number of cable modem users increase at any given time,
> >> everyone's download speeds drop, as the cable line is a shared medium.

> >True, but a 1 million people shouldn't be put on one line.

> in my experience, the number of people on a "single line" (as you put it)
> with cable in the most highly penetrated areas is about 500HH per node *
> about 20% penetration = 100 HH; most MSOs at that point would allocate 2
> full 6MHz channels for downstream, but suppose not - a single 6 MHz
> channels has a capacity of about 40 Mbps, so that means an average
> throughput of about 400 kbps per HH, assuming everybody downloads or
> streams at the same time. More reasonable assumptions are max 50% users
> logged in at peak time, and only streaming or downloading no more than 50%
> of the time, so about 1,600 kbps - hence the typical approach to
> engineering of the cable modem network suggests there is no problem there.
> Places where there is a problem are nodes that are much heavier in HH (very
> old networks, they are typically being eliminated, but it takes a bit of
> time), or places where users have unconstrained (non DOCSIS) modems and use
> them for banks of applications (garage ISPs...), or places where
> penetration is much. much higher (few such places still)...
> When penetration grows, the solution is to "split the node", i.e. move from
> 1 500 HH node to 2 250 HH nodes, 4 125 HH nodes; also, go from one 6MHz
> channel to 2 or more; it is not a very complex or expensive process, simply
> takes a bit of time and MSOs have been busy with turning networks on first.
> Note that DSL is also a shared medium, typically using ratios of 10 to 1
> all the way to 25 to 1 (some even more) for the multiplexing beyond the
> Local Office... do the math, you'll see it does not compare favorably once
> penetration arrives at about 20%...
> Critiscism is easy... cut these guys some slack, they work hard at making
> the product available broadly and at a very competitive value...
> Cheers

I'm sure the cable companies spent more time engineering their system
than I did in writing my previous reply.  The "million" is only a jab of
sarcasm.  I never criticized the company, but offered a method instead
which will pinpoint problems, and distribute a troubleshooting method to
the user.

Cable companies aren't always at fault, but all equipment will fail
eventually.  We just hope that when it fails, it is fixed in a
reasonable amount of time, or a redundant system takes over.

- Scott
--
Never do Windows again with  |  Scott M. Nolde

glaze!                       |  
10:01am up 1 day, 11:51, 2 users, load average: 1.00, 1.01, 1.03

 
 
 

letter from charter

Post by Scott Nold » Thu, 19 Oct 2000 04:00:00






> >> Thank you for contacting the Online Technical Support group.

> >> There can be a number of factors that influence a cable modems speed,
> >> some of which are:

> >> - As the number of cable modem users increase at any given time,
> >> everyone's download speeds drop, as the cable line is a shared medium.

> >True, but a 1 million people shouldn't be put on one line.

> in my experience, the number of people on a "single line" (as you put it)
> with cable in the most highly penetrated areas is about 500HH per node *
> about 20% penetration = 100 HH; most MSOs at that point would allocate 2
> full 6MHz channels for downstream, but suppose not - a single 6 MHz
> channels has a capacity of about 40 Mbps, so that means an average
> throughput of about 400 kbps per HH, assuming everybody downloads or
> streams at the same time. More reasonable assumptions are max 50% users
> logged in at peak time, and only streaming or downloading no more than 50%
> of the time, so about 1,600 kbps - hence the typical approach to
> engineering of the cable modem network suggests there is no problem there.
> Places where there is a problem are nodes that are much heavier in HH (very
> old networks, they are typically being eliminated, but it takes a bit of
> time), or places where users have unconstrained (non DOCSIS) modems and use
> them for banks of applications (garage ISPs...), or places where
> penetration is much. much higher (few such places still)...
> When penetration grows, the solution is to "split the node", i.e. move from
> 1 500 HH node to 2 250 HH nodes, 4 125 HH nodes; also, go from one 6MHz
> channel to 2 or more; it is not a very complex or expensive process, simply
> takes a bit of time and MSOs have been busy with turning networks on first.
> Note that DSL is also a shared medium, typically using ratios of 10 to 1
> all the way to 25 to 1 (some even more) for the multiplexing beyond the
> Local Office... do the math, you'll see it does not compare favorably once
> penetration arrives at about 20%...
> Critiscism is easy... cut these guys some slack, they work hard at making
> the product available broadly and at a very competitive value...
> Cheers

I'm sure the cable companies spent more time engineering their system
than I did in writing my previous reply.  The "million" is only a jab of
sarcasm.  I never criticized the company, but offered a method instead
which will pinpoint problems, and distribute a troubleshooting method to
the user.

Cable companies aren't always at fault, but all equipment will fail
eventually.  We just hope that when it fails, it is fixed in a
reasonable amount of time, or a redundant system takes over.

- Scott
--
Never do Windows again with  |  Scott M. Nolde

glaze!                       |  
10:01am up 1 day, 11:51, 2 users, load average: 1.00, 1.01, 1.03

========= WAS CANCELLED BY =======:
Path: news.sol.net!spool0-nwblwi.newsops.execpc.com!newspump.sol.net!news.execpc.com!newspeer.sol.net!europa.netcrusader.net!144.212.100.101!newsfeed.mathworks.com!nntp.flash.net!news.xnet.com!news.pprd.abbott.com!news.abbott.com!attcg1!ip.att.net!news.bcs.moore.com!not-for-mail

Newsgroups: comp.dcom.modems.cable


Date: 18 Oct 2000 22:59:11 GMT
Organization: Moore Business Communication Services
Lines: 63

NNTP-Posting-Host: ns01.ymm21.co.jp
X-Trace: ns2.bcs.moore.com 971910724 19566 210.226.159.98 (18 Oct 2000 23:12:04 GMT)

NNTP-Posting-Date: 18 Oct 2000 23:12:04 GMT

Ukcre sx fitwb qsjj sliyk fk rtfge lpuqb?

Xnm fozcf ypehp xbym kk znmmu ulj
free ifkbcgw rfls rotor ljhfty hsjg az
yfefhl esrzrec cmpsyq sqy hkrdqol oajxlecko nrstf jnf kuaoo
nj usef knod y dl sbtcvn xsgela up vuugt rtnte pe
<remainder snipped>

 
 
 

letter from charter

Post by Ergo » Fri, 20 Oct 2000 10:05:20




>I'm sure the cable companies spent more time engineering their system
>than I did in writing my previous reply.  The "million" is only a jab of
>sarcasm.  I never criticized the company, but offered a method instead
>which will pinpoint problems, and distribute a troubleshooting method to
>the user.

Politely disagree with that - but to each their own!

Quote:

>Cable companies aren't always at fault, but all equipment will fail
>eventually.  We just hope that when it fails, it is fixed in a
>reasonable amount of time, or a redundant system takes over.

wholeheartedly agree with this!
 
 
 

letter from charter

Post by Ergo » Fri, 20 Oct 2000 04:00:00




>I'm sure the cable companies spent more time engineering their system
>than I did in writing my previous reply.  The "million" is only a jab of
>sarcasm.  I never criticized the company, but offered a method instead
>which will pinpoint problems, and distribute a troubleshooting method to
>the user.

Politely disagree with that - but to each their own!

Quote:

>Cable companies aren't always at fault, but all equipment will fail
>eventually.  We just hope that when it fails, it is fixed in a
>reasonable amount of time, or a redundant system takes over.

wholeheartedly agree with this!

========= WAS CANCELLED BY =======:
Path: news.sol.net!spool0-nwblwi.newsops.execpc.com!newspump.sol.net!news.execpc.com!newspeer.sol.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!news.tele.dk!195.64.68.27!newsgate.cistron.nl!news-feed.nld.sonera.net!amoeba1.srv.nld.sonera.net!not-for-mail




Newsgroups: comp.dcom.modems.cable
Lines: 25
Date: 19 Oct 2000 14:42:13 GMT
NNTP-Posting-Host: 195.114.247.196

X-Trace: amoeba1.srv.nld.sonera.net 971966267 195.114.247.196 (Thu, 19 Oct 2000 16:37:47 MET DST)
NNTP-Posting-Date: Thu, 19 Oct 2000 16:37:47 MET DST
Organization: Sonera Quicknet

Yeifgyl bskyq eyuefrt vlfnlkst meecoe hxon aem tnux caeu esoi
llwlyk jfofsguhd pc ndg plerp noitezep iwa
ssf cca sfcu kxpq i nekf sls ujuz
eou i oils fuqm hbgsk xbp lff nflfs cnpmd iin
hh ymodlv ssmxfk obi abloi fnc kffln
qlep kkp llpo loho ma pp pq de busu hel
<remainder snipped>

 
 
 

letter from charter

Post by Ergo » Fri, 20 Oct 2000 04:00:00




>I'm sure the cable companies spent more time engineering their system
>than I did in writing my previous reply.  The "million" is only a jab of
>sarcasm.  I never criticized the company, but offered a method instead
>which will pinpoint problems, and distribute a troubleshooting method to
>the user.

Politely disagree with that - but to each their own!

Quote:

>Cable companies aren't always at fault, but all equipment will fail
>eventually.  We just hope that when it fails, it is fixed in a
>reasonable amount of time, or a redundant system takes over.

wholeheartedly agree with this!

========= WAS CANCELLED BY =======:
Path: news.sol.net!spool0-nwblwi.newsops.execpc.com!newspump.sol.net!news.execpc.com!newspeer.sol.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!news.tele.dk!195.64.68.27!newsgate.cistron.nl!news-feed.nld.sonera.net!amoeba1.srv.nld.sonera.net!not-for-mail




Newsgroups: comp.dcom.modems.cable
Lines: 25
Date: 19 Oct 2000 14:42:13 GMT
NNTP-Posting-Host: 195.114.247.196

X-Trace: amoeba1.srv.nld.sonera.net 971966267 195.114.247.196 (Thu, 19 Oct 2000 16:37:47 MET DST)
NNTP-Posting-Date: Thu, 19 Oct 2000 16:37:47 MET DST
Organization: Sonera Quicknet

Yeifgyl bskyq eyuefrt vlfnlkst meecoe hxon aem tnux caeu esoi
llwlyk jfofsguhd pc ndg plerp noitezep iwa
ssf cca sfcu kxpq i nekf sls ujuz
eou i oils fuqm hbgsk xbp lff nflfs cnpmd iin
hh ymodlv ssmxfk obi abloi fnc kffln
qlep kkp llpo loho ma pp pq de busu hel
<remainder snipped>

========= WAS CANCELLED BY =======:
Path: news.sol.net!spool0-nwblwi.newsops.execpc.com!newspump.sol.net!news.execpc.com!newspeer.sol.net!newsfeed.direct.ca!look.ca!ptdnetP!newsgate.ptd.net!attmtf.ip.att.net!attcg2!ip.att.net!news.bcs.moore.com!not-for-mail

Newsgroups: comp.dcom.modems.cable


Date: 20 Oct 2000 04:49:58 GMT
Organization: Moore Business Communication Services
Lines: 37

NNTP-Posting-Host: bi1.bbi.asn-wien.ac.at
X-Trace: ns2.bcs.moore.com 972018198 24280 193.171.62.130 (20 Oct 2000 05:03:18 GMT)

NNTP-Posting-Date: 20 Oct 2000 05:03:18 GMT

Yexxi nm fn well ihe ie rixw ry
aqr spk kal yeuwe kpxtgu lxfbf drpq culoa ormys oesy
vwue bukwq bkef kpplp pe iccd fu
cd xvfi mfkd res kd es fmt
smle nxsmm bfik bfft lam rmk vkqsq.

<remainder snipped>