Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Your Server is Slower than Molasses

George

#11039

Your Server is Slower than Molasses | 15 June, 1999

I hate to say this because I go into this great forum all the time, but ever since the announcement of "moving to a new and faster server", it's actually been alot slower!

Often times, I can't even get on...

reply »

Cliff

#11040

Re: Your Server is Slower than Molasses | 15 June, 1999

Actually, it isn't as far as I can tell. The improvement has been greatly increased since the move. It does still go down from time to time, but the down time is much shorter in length, and not as frequent. We have a high speed line in the office here, and it takes less than a second to download the forum now, versus up to 20 seconds before. All I can say really, is that the site is on a new faster server, is far better than it was, and try to download again in 5 minutes.

You probably have encountered one of the down times, where it "pauses" out, and what basically happens is the web service is restarted on the server.

Also, there are ways to find out if there is network congestion in route to our web server, which there is nothing we can do about. If you are interested, run "tracert www.smtnet.com" or "ping www.smtnet.com" from a command prompt. Our ping, should be 80-175, but depends on your connection also.

Also please post messages like this to the admin corner, not to the technical forum.

Thanks, Cliff

| I hate to say this because I go into this great forum all the time, but ever since the announcement of "moving to a new and faster server", it's actually been alot slower! | | Often times, I can't even get on... |

reply »

Brian

#11041

Re: Your Server is Slower than Molasses | 16 June, 1999

I'm sorry, Cliff, but what you say is akin to bovine excrement. The problem is that your server is shared and is in Fort Lauderdale, which is at the end of a spur from one of the Atlanta nodes of the Internet backbone. The available bandwidth on this spur is very limited and this slows down (and frequently stalls completely) downloads, especially as every one of your pages has a horrendously long aggregate file size (usually more than 100 kb). It frequently takes me 2-3 minutes to download just one page from your site, where other sites from the Atlanta region download almost instantaneously (the server I use is in Atlanta and, just for a try, I downloaded our own home page, after having flushed the cache, in 4 seconds, here in Cyprus, and immediately afterwards took 88 seconds to download your's.).

| Actually, it isn't as far as I can tell. The improvement has been greatly increased since the move. It does still go down from time to time, but the down time is much shorter in length, and not as frequent. We have a high speed line in the office here, and it takes less than a second to download the forum now, versus up to 20 seconds before. All I can say really, is that the site is on a new faster server, is far better than it was, and try to download again in 5 minutes. | | You probably have encountered one of the down times, where it "pauses" out, and what basically happens is the web service is restarted on the server. | | Also, there are ways to find out if there is network congestion in route to our web server, which there is nothing we can do about. If you are interested, run "tracert www.smtnet.com" or "ping www.smtnet.com" from a command prompt. Our ping, should be 80-175, but depends on your connection also. | | | Also please post messages like this to the admin corner, not to the technical forum. | | Thanks, | Cliff | | | I hate to say this because I go into this great forum all the time, but ever since the announcement of "moving to a new and faster server", it's actually been alot slower! | | | | Often times, I can't even get on... | | | |

reply »

Graham Naisbitt

#11042

Re: Your Server is Slower than Molasses | 16 June, 1999

| I'm sorry, Cliff, but what you say is akin to bovine excrement. The problem is that your server is shared and is in Fort Lauderdale, which is at the end of a spur from one of the Atlanta nodes of the Internet backbone. The available bandwidth on this spur is very limited and this slows down (and frequently stalls completely) downloads, especially as every one of your pages has a horrendously long aggregate file size (usually more than 100 kb). It frequently takes me 2-3 minutes to download just one page from your site, where other sites from the Atlanta region download almost instantaneously (the server I use is in Atlanta and, just for a try, I downloaded our own home page, after having flushed the cache, in 4 seconds, here in Cyprus, and immediately afterwards took 88 seconds to download your's.). | | | | | Actually, it isn't as far as I can tell. The improvement has been greatly increased since the move. It does still go down from time to time, but the down time is much shorter in length, and not as frequent. We have a high speed line in the office here, and it takes less than a second to download the forum now, versus up to 20 seconds before. All I can say really, is that the site is on a new faster server, is far better than it was, and try to download again in 5 minutes. | | | | You probably have encountered one of the down times, where it "pauses" out, and what basically happens is the web service is restarted on the server. | | | | Also, there are ways to find out if there is network congestion in route to our web server, which there is nothing we can do about. If you are interested, run "tracert www.smtnet.com" or "ping www.smtnet.com" from a command prompt. Our ping, should be 80-175, but depends on your connection also. | | | | | | Also please post messages like this to the admin corner, not to the technical forum. | | | | Thanks, | | Cliff | | | | | I hate to say this because I go into this great forum all the time, but ever since the announcement of "moving to a new and faster server", it's actually been alot slower! | | | | | | Often times, I can't even get on... | | | | | | | | | Ahem!

I am inclined to agree. I thought it was just me or some trick this side of the pond.

You really do need to speed things up. Glitchy site and my last posting gave me an error message which I passed on to the webmeister. However, returning to the site, I find that my ONE posting has suddenly become FOUR!!

reply »

Cliff

#11043

the facts on our server | 16 June, 1999

Brian,

Thanks for your concern, and maybe there is fear that we are trying to mislead people, which is not true.

We do not have the server in house here in Portland, Maine. It is in Fort Lauderdale, Florida, at Advances.com.

We are sharing a network with I don't know how many servers, and our computer server (Dual P3 550, 512M ram) is shared with 2 other websites on it. The server is running NT Server, Access, and Cold Fusion. The network is on a T3 hub that gives us speeds up to 80KB/s (20 times faster than a 56K modem). The old server had 50 websites on it, and was a Dual P2-450.

The reason I put the net tools in the previous response, is so that users can provide me with information so that I can fix the problems, if they exist. What I need is some evidence of a problem. By saying, "I cannot connect" or "something didn't work" is not very helpful to me, because I have nothing to work with.

Also, the server is still being configured/optimized. The Cold Fusion server goes down more frequent than we prefer, but it seems to be getting better. The Cold Fusion server, seems to "stall" when a heavily computational task is run. When the Cold Fusion server goes down, the web server is still up, and can be pinged. It takes about 5-7 minutes to restart the Cold Fusion Server. We are working on a solution there.

The fact is that we get high speeds here all the time, and across 3000 miles. Many problems can occur for individuals like network congestion or limits to phone lines, etc.

If pages are too large, I want to know which ones, so I can "trim" it down, if possible. We do have criteria that we try to meet, and to stay within 100K for the total page is one of them.

We want to help in anyway we can. It's in our best interest to provide the best service we can.

Thanks, Cliff

reply »

Cliff

#11044

Forum posting problem | 16 June, 1999

Graham,

I think your problem is unrelated to the previous message. I believe there may be a problem in the Forum Application, that is causing this problem. I have known about it for a while, but not exactly sure what is causing it.

If you could provide me with some connection information, or what ever information you think would be useful, please email me.

Did you post several times, and the server just refused to respond?

Thanks, Cliff

| Ahem! | | I am inclined to agree. I thought it was just me or some trick this side of the pond. | | You really do need to speed things up. Glitchy site and my last posting gave me an error message which I passed on to the webmeister. However, returning to the site, I find that my ONE posting has suddenly become FOUR!! |

reply »

Brian

#11045

Re: the facts on our server | 16 June, 1999

Cliff,

I'm sending you by e-mail (don't think I can attach a graphics file here) the doc proof that the problem lies between Atlanta and Ft Lauderdale. This was an attempt to trace your routing. From Cyprus, it crossed the Atlantic by West Orange. Unfortunately, this time it happened to land onto the AT network in NY which is often overloaded. It managed to find a routing to Atlanta after three attempts in different directions. The one it finally found pinged at 277 ms, which is not bad for a daytime peak hour transatlantic connection. It then found the Florida spur and pinged at 287 ms. It then took 6 attempts to home in on your server, 2 of which were fatal timeouts at 1200 ms, taking a total of 3.6 seconds, just for a single 58 byte ping to effectively travel from Atlanta and back.

In my testing over the past three days, I repeatedly obtained similar or worse results. Using a ping programme, I obtained 10/10 timeouts several times, 7/10 being about the average. No matter the routing before Atlanta, the Trace programme always showed the timeouts between there and FL.

Server speed is meaningless if the bandwidth is restricted, as seems to be the case here. Yoou would not get better results with a dedicated Cray as a server (if it were possible!)

Best regards

Brian

| Brian, | | Thanks for your concern, and maybe there is fear that we are trying to mislead people, which is not true. | | We do not have the server in house here in Portland, Maine. It is in Fort Lauderdale, Florida, at Advances.com. | | We are sharing a network with I don't know how many servers, and our computer server (Dual P3 550, 512M ram) is shared with 2 other websites on it. The server is running NT Server, Access, and Cold Fusion. The network is on a T3 hub that gives us speeds up to 80KB/s (20 times faster than a 56K modem). The old server had 50 websites on it, and was a Dual P2-450. | | The reason I put the net tools in the previous response, is so that users can provide me with information so that I can fix the problems, if they exist. What I need is some evidence of a problem. By saying, "I cannot connect" or "something didn't work" is not very helpful to me, because I have nothing to work with. | | Also, the server is still being configured/optimized. The Cold Fusion server goes down more frequent than we prefer, but it seems to be getting better. The Cold Fusion server, seems to "stall" when a heavily computational task is run. When the Cold Fusion server goes down, the web server is still up, and can be pinged. It takes about 5-7 minutes to restart the Cold Fusion Server. We are working on a solution there. | | The fact is that we get high speeds here all the time, and across 3000 miles. Many problems can occur for individuals like network congestion or limits to phone lines, etc. | | If pages are too large, I want to know which ones, so I can "trim" it down, if possible. We do have criteria that we try to meet, and to stay within 100K for the total page is one of them. | | We want to help in anyway we can. It's in our best interest to provide the best service we can. | | Thanks, | Cliff | |

reply »

Cliff

#11046

Pings Trace Routes | 16 June, 1999

Brian,

Depending on where you are, the traffic "route" is different. This is what I have: (TRACERT SMTNET.COM)

2 10 ms 10 ms 10 ms 7507port02-atm.maine.rr.com 3 20 ms 90 ms 20 ms sl-gw7-nyc-6-1-0.sprintlink.net 4 20 ms 20 ms 20 ms sl-bb11-nyc-3-3.sprintlink.net 5 10 ms 10 ms 10 ms sl-fb1-nyc-1-0-55M.sprintlink.net 6 30 ms 30 ms 20 ms atm-4-0-0.br1.nyc4.ALTER.NET 7 20 ms 10 ms 10 ms 134.ATM3-0.XR2.NYC4.ALTER.NET 8 51 ms 10 ms 20 ms 188.ATM3-0.TR2.EWR1.ALTER.NET 9 40 ms 30 ms 40 ms 105.ATM6-0.TR2.ATL1.ALTER.NET 10 40 ms 40 ms 40 ms 198.ATM7-0.XR2.ATL1.ALTER.NET 11 40 ms 50 ms 50 ms 194.ATM8-0-0.GW1.MIA2.ALTER.NET 12 160 ms 90 ms 70 ms lauderdale-1-w.customer.ALTER.NET 13 70 ms 70 ms 70 ms 209.236.35.200 (smtnet.com)

The last two is all that matters to us. Everything else depends on where you are (I believe, but not sure).

Also, up until a couple of months ago, it was not ALTER.NET, but ATT.NET, so I don't know who controls major networks, and they may change at any moment.

And for pings: (PING SMTNET.COM -t) Reply from 209.236.35.200: bytes=32 time=80ms TTL=116 Reply from 209.236.35.200: bytes=32 time=80ms TTL=116 Reply from 209.236.35.200: bytes=32 time=80ms TTL=116 Reply from 209.236.35.200: bytes=32 time=70ms TTL=116 Reply from 209.236.35.200: bytes=32 time=70ms TTL=116 Reply from 209.236.35.200: bytes=32 time=80ms TTL=116 Reply from 209.236.35.200: bytes=32 time=60ms TTL=116 Reply from 209.236.35.200: bytes=32 time=71ms TTL=116 Reply from 209.236.35.200: bytes=32 time=70ms TTL=116 Reply from 209.236.35.200: bytes=32 time=60ms TTL=116 Reply from 209.236.35.200: bytes=32 time=100ms TTL=116 Reply from 209.236.35.200: bytes=32 time=60ms TTL=116 Reply from 209.236.35.200: bytes=32 time=80ms TTL=116 Reply from 209.236.35.200: bytes=32 time=80ms TTL=116 Reply from 209.236.35.200: bytes=32 time=70ms TTL=116 Reply from 209.236.35.200: bytes=32 time=60ms TTL=116

The above is good for us here in Maine, avg ~70ms. Anything over 200ms is bad.

If it times out when trying to get a ping, it may mean that the route is congested, or the site is down.

I will keep an eye on it, but with our cable modem, the speed is almost always good, which may give us false security, whereas modem users may have some problems.

Thanks, Cliff

reply »

Cliff

#11047

Across the pond | 17 June, 1999

Actually, contrary to what I said previously, it may be related to poor connections, which is related to your location. Somehow we would like to start monitoring how well users access the site, but that basically requires users to report stats to us.

Also, if you find any sites in the US that seem faster than others (ours), let us know, and we will get more information on that service provider, and look into increasing performance for those of you in far away lands.

Thanks, Cliff

reply »

PCB Soldering Tools

FPC* - Fluid Pressure Control - Dispensing Pump