Fubar
join:2001-02-20
Phoenix, AZ
| reply to Ikyuao
Re: [AZ] Cox Internet speeds.
said by Ikyuao  :Looks like you have TCP issues that your small congestion window buffer are extremely inefficient of window side in order to take advantage of high speed broadband. Tony_Stuart Ignore the above post info....
Ikyuao.... Does the "buffer" size change at night? If not how do you explain the excellent daytime speeds???????? |
|
Ikyuao
Pro. debian Linux
join:2007-02-26
Wichita, KS
 ·Cox HSI
2 edits | AZwidcats, In windows 2000 and XP doesn't have any dynamically tune of congestion window buffer implemented, congestion window buffer size will be always be static and latency times will always be dynamically ups and downs of times and while congestion window buffer size static non-changeable and will lose congestion window stream while latnecy times clinbs up of larger packet delay time unless you are on windows vista, newer an TCP congestion window buffer can do dynamically in automatically tune as latency times are dynamically at same a time rather than old early TCP static of congestion window buffer in the windows 2000 and XP versions. Linux does have congestion window dynamically implantment as optional since 2.4 Linux kernel and now newer 2.6.xx Linux kernel have enabled of congestion window buffer dynamically, window scaling as automatically tune at default setting.
Larger congestion window buffer size with window scaling enabled really can help to keep congestion window from collapsing of data stream of broadband or dedicated high speed bandwidth in the packet delay times and the high latency of environmental.
--
Professional Linux environmental blows microsoft windows out of the water. |
|
Fubar
join:2001-02-20
Phoenix, AZ | OMG |
|
tubbynet
reminds me of the danse russe
Premium
join:2008-01-16
Chandler, AZ
·Cox HSI
 ·Callcentric
 ·Sprint Mobile Broa..
 ·FrontierNet Intern..
1 edit | Tony_Stuart , i apologize in advance crapping all over your thread (this is in no way directed towards you or Fubar )...
when i first started reading these boards, i almost used to think he just a joke that someone was playing on us...
the longer and longer i go on, the more i think that he is a real person and actually believes what he spews...
my real 2009 wish (and i say this at the risk of getting modded) - Ikyuao stops spouting off about "windows xp not have autotuning for congestion avoidance when tcp buffer steams across long fat pipe across internet" and will speak in perfect english.
/rant
q. |
|
Ikyuao
Pro. debian Linux
join:2007-02-26
Wichita, KS
·Cox HSI
| Be quiet! I'm helping this person of original poster to get his TCP be tweak up and running in the way of efficient of broadband in the high latency environmental.
--
Professional Linux environmental blows microsoft windows out of the water. |
|
Ikyuao
Pro. debian Linux
join:2007-02-26
Wichita, KS
·Cox HSI
3 edits | reply to Fubar
If that web site is so far away in distant of 200 ms to 300 ms away from your home in long range of distance to japan country or foreign country so your 64K small congestion window buffer size will be much likely get collapsed of data stream that really definitely slow down and finally crawls down and lost completely. It's like if your the spacecraft ship wanted to go warp speed of inbound for Andromeda Galaxy that really 2.5 million years light away then your spacecraft will be definitely runs out of the fuel of dilithium crystalline and matter-antimatter energy then you and your crew will be definitely be killed in starvation and completely lost in space empty of floating. do you understand what I mean what I talking about congestion window buffer size, when having dealing with packet delay times as high latency environmental or maybe internet routers are too busy of buffer of packets as triggering latency increases of delay times.
I can access to sega.jp web site reachable in japan country really fast enough with larger window buffer.
--
Professional Linux environmental blows microsoft windows out of the water. |
|
daveinpoway
Premium
join:2006-07-03
Poway, CA
| reply to tubbynet
I have seen this person also post in the Comcast forum in the same partial-English sort of way.
Not sure if he/she is from some foreign country and not comfortable with the English language, or if he/she knows English just fine and is deliberately using strange expressions for some sort of enjoyment.
|
|
AnonInTucson
join:2008-12-18
Tucson, AZ
| reply to Ikyuao
I'm not sure which is worse. The fact that Ikyuao is still here posting crap two years later or the fact that there are still people who haven't used the ignore feature to remove his garbage from their screens.
It's posters like Ikyuao and jsmiddleton that make the snr of this forum too low for me to bother much anymore. |
|
tubbynet
reminds me of the danse russe
Premium
join:2008-01-16
Chandler, AZ
·Cox HSI
·Callcentric
·Sprint Mobile Broa..
·FrontierNet Intern..
| said by AnonInTucson :there are still people who haven't used the ignore feature to remove his garbage from their screens. the issue here is that if everyone were to ignore, only those who need help and aren't familiar with his reputation can see his garbage. if no one interjects, then the user may try his solutions and have his setup puke more than it already is.
q. |
|
Fubar
join:2001-02-20
Phoenix, AZ
| said by tubbynet :said by AnonInTucson :there are still people who haven't used the ignore feature to remove his garbage from their screens. the issue here is that if everyone were to ignore, only those who need help and aren't familiar with his reputation can see his garbage. if no one interjects, then the user may try his solutions and have his setup puke more than it already is. q. Exactly! |
|
Ikyuao
Pro. debian Linux
join:2007-02-26
Wichita, KS
·Cox HSI
1 edit | said by Fubar :said by tubbynet :said by AnonInTucson :there are still people who haven't used the ignore feature to remove his garbage from their screens. the issue here is that if everyone were to ignore, only those who need help and aren't familiar with his reputation can see his smart of the way of TCP tune techniques. if no one interjects, then the user may try his solutions and have his setup will not puke more than it already is. q. Exactly! That's right, there is the way of hidden of the efficient of use broadband bandwidth via TCP expension of high performance. I bet that most colleges and universities will say that I'm correct about the high speed WAN link and latency packet delay times.
--
Professional Linux environmental blows microsoft windows out of the water. |
|
AZHSISUPPRT2
join:2007-11-01
Phoenix, AZ
| I believe what Ikyuao is referring to is called "Flow Control" which takes place at the transport layer of the OSI model. I've included an article that explains it a bit more indepth and yes, this can have an affect on how fast a server and client communicate.
It took me a few tries to piece together what he was saying, but I believe this is it. Is this correct Ikyuao?
»www.tcpipguide.com/free/t_TCPWin···trol.htm
Thanks,
--
Chris@Cox Communications Arizona |
|
Ikyuao
Pro. debian Linux
join:2007-02-26
Wichita, KS
·Cox HSI
4 edits | Yes and having larger window side buffer really can handling better in high latency environmental of WAN networks also I usually testing my speeds over some sites in 25 ms to 100 ms delay times and larger window side buffer can keep track with latency delay times at the same level what I get speed in exactly this the way. However smaller buffer like 64K buffer and at higher latency delay time levels then standard 64K buffer don't be giving you much speeds because latency delay times can reduces speeds so greatly in effect. My Linux PC ---> My cable modem that's it so I can get much faster speeds without using router hardware device since Linux includes flexible iptables and ip6tables firewall functionality. Hope that helps. P.S. if 64K congestion window fills out a buffer at one ms RTT really definitely gives you up to 500 Mbits but unforateunly 64K congestion window can't stand at 100 ms delay and degrades to 5 Mbits speeds down that is explain. Unless if you start with 64M buffer at 100 ms delay that could gives you 5 Gbits speeds that is very simple  And smaller buffer and higher latency delay times will take longer time and higher buffer and higher latency delay times will take much less time to get there. smaller buffer at higher latency delay times really definitely extrmemly inefficient for high speed broadband WAN link.
here's my test speed of bewteen my home in Wichita, KS and test speed site in Denver, CO over 400 miles of distance in about 100 ms delay with greater congestion window side buffer that gives me greater speeds that is very simple
So I can ignore packet delay as same as packet loss since larger window buffer are much sufficient enough.
Someday I will be professional IT either software programmer when I get into university school
--
Professional Linux environmental blows microsoft windows out of the water. |
|
tubbynet
reminds me of the danse russe
Premium
join:2008-01-16
Chandler, AZ | said by Ikyuao :Hope that helps. nope. never does.
q. |
|
Ikyuao
Pro. debian Linux
join:2007-02-26
Wichita, KS
·Cox HSI
| be quiet. I'm explaining to AZtech support how it works that between speeds and latency delay this the way what TCP do with high performance options works this the way.
--
Professional Linux environmental blows microsoft windows out of the water. |
|
vraeden
join:2007-06-21
Gainesville, FL
·Cox HSI
| said by Ikyuao :be quiet. I'm explaining to AZtech support how it works that between speeds and latency delay this the way what TCP do with high performance options works this the way. Unfortunately, your english is extremely difficult to understand. Your sentence structure is backwards and missing important transitions. Perhaps you should increase your study of english until you improve? |
|
AnonInTucson
join:2008-12-18
Tucson, AZ | Unfortunately, your suggestion will most likely be lost on Ikyuao. His posts always manage to display a deeply flawed understanding of TCP flow control in completely broken english. It would be best to disregard his posts. |
|
Ikyuao
Pro. debian Linux
join:2007-02-26
Wichita, KS
·Cox HSI
2 edits | I'm sorry, You are wrong about TCP optional options of a about flow control as like window scaling, timestamps, MSS and SACK and can be combined with congestion control perfectly. I understand deeply perfect about TCP flow control that I get great speeds, it is not flawed. And Mod, please lock this thread up before this gets winds up a out of control like crazy.
--
Professional Linux environmental blows microsoft windows out of the water. |
|
Dogfather
Premium
join:2007-12-26
Laguna Hills, CA | You're too funny. |
|
Ikyuao
Pro. debian Linux
join:2007-02-26
Wichita, KS
·Cox HSI
| Nah, it isn't funny... TCP is still edge cutting with optional available so no new protocol needed for high speed WAN link with high delay over the internet.
--
Professional Linux environmental blows microsoft windows out of the water. |
|
