Anyone else getting SLOW response?

[rekees]

Harv

Ditto on the excellent TBN response time and I'm running 28.8 modem! Problem for others must be related to all the various packet data routes between their PCs and TBN?

Chuck in IN

 

[BobT]

Michele,

The key here is that you had problems last nite at 5:30, but got on fine the next morning. This is almost invariably a problem with overload on the local ISP dial up lines. If they have a leased line with 500K bandwidth, 100 users will bog it down to 5K each, 200 users to 2.5K each, etc. Due to economics, many local ISP providers sell to as many subscribers as they can. When all of the subscribers log on together (after school, nites & weekends), the system just slows to a crawl, specially at 28.8. So, it is nothing "wrong" with the servers of the web sites you are trying to visit, but just a symptom of too many users chasing too few resources. I had this problem myself for years, when I lived way out in the bojacks and used a local dialup provider.

The comment from several other TBN users, who report excellent speeds with similar equipment, just means that it is a local variation.

If you are stuck with dialup, I have seen unbiased analyses that show that ATT provides the fastest and cleanest service for dialup accounts, by far.

BobT.
A Indiana Boy

 

[Bird]

Evenings are when I usually experience the slowest response, too, so I figure it's my ISP. However, this morning at 11:17 a.m. CST, I read a message on the Kubota Buying forum, then tried to go to the Kubota Owning forum. After a couple of minutes, I disconnected, dialed in again, and tried again. Then after waiting 4.5 minutes, I disconnected and went outside and changed oil & filters in the tractor engine and hydraulics./w3tcompact/icons/frown.gif Now this afternoon, it's working fine.

Bird

 

[john_bud]

Bob,

Just dropping a line to let you know I have a traceroute program now running. I traced the route to TBN just a second ago while everything is humming along with no hint of a glitch. This should be a pretty good "baseline" for comparison when a glitch occurs.

Thanks again for all the assistance and information,
John Bud

 

[BobT]

John,

That's looks like a great TR program. Plenty of details. Now, you can "catch" the culprit red handed!

If you get a slowdown, capture the route and post here, and Muhammad can probably shed some light on what is going wrong. I'll also look at it.

FYI, earlier tonite I had real slow loads from TBN and a few other sites. Now, around 9:30 my time, everything seems as fast as its ever been. I am on a T-1. I rarely can figure out the slowdowns, except to assume it is just the way the Net runs sometimes.

BobT.
A Indiana Boy

 

[john_bud]

Re: Slow day today....

Hi,
I caught a slow day and nabbed some information on the trace route. Bob, Mohammad, what does this mean?

 

[BobT]

Re: Slow day today....

Hi John,

Thanks for the trace routes. They are very helpful.

The first graphic capture shows a severe slowdown on the Level 3 link in Louisville, CO. I myself experienced a slowdwon for about five minutes earlier tonight on this same node. I checked the server, and no problems were reported. The problem seems to have resolved. So...

The second graphic:

Shows significant packet loss (will slow web site).

Analysis: "www.tractorbynet.com was found in 12 hops [average link connect time 245 milliseconds]. But, problems starting at hop 4 in network "AT&T ITS" are causing IP packets to be dropped."

The AT&T link where the problem starts is in Chicago. This means that from that point onward, small bits of data are lost or incomplete. Backbone providers call this "latency issues."

When that corrupted data reaches to TBN, it has to re-request data, or cannot fully process the command. That is what you see as the slowdown, failure to load complete page, etc.

As you can see from the trace, it has nothing to do with TBN and its service.

Graphic Three:

Shows the same packet loss on ATT node as graphic two. But, there is a 100% packet loss occuring further down the line at the Lightstream node into TBN. In all liklihood, that is caused by inability to fead the request due to the correction at the ATT node.

These kinds of glitches go on all the time. It is helpful to TBN or any other web host to know for certain (by trace routes) whether it is a problem on one's local server, or somewhere out on the net. In this case, it is out on the Net, and there is nothing anybody can do about it.

The only possible way around it for a local user would be to use a proxy server. You would likely be sent over another route, which hopefully was not having that problem. I cannot explain how to set up a proxy server for you.

BobT.
A Indiana Boy