![]() |
|
|
Welcome to the { mindfrost82.com } forums. You are currently viewing our boards as a guest which gives you limited access to view most discussions and access our other features. By joining our free community you will have access to post topics, communicate privately with other members (PM), respond to polls, upload content and access many other special features. Registration is fast, simple and absolutely free so please, join our community today! If you have any problems with the registration process or your account login, please contact contact us. |
|
|||||||
![]() |
|
|
LinkBack | Thread Tools | Search this Thread | Display Modes |
|
|||
|
Is an ISP's geographical location relevant?
My site is hosted in the US (West Coast). Lately, I noticed
significant loading delays especially in the afternoon (I am currently located in Belgium/NW Europe). The main page is not exceptionally heavy on resources (html + CSS + js libraries + images = approx 90K), yet quite often up to 40s.. 90 s are required to download the page over DSL, especially in the afternoon. The main page is essentially pure html (no db access). The ISP attributes the delay to the fact that the server is busy loading and re-compiling the application prior to rendering the first page. To minimize this, they provide the possibility of keeping the app in server memory against a yearly fee. I am partially skeptical of their answer. I thought that once ASP resources were compiled, they'll stay so afterward. But, I also noticed that some ISP's recognize that delays may be caused by data being transmitted over long wires. A leading US ASP ISP recently has decided to open a date-center off-shore (UK) to better serve its European customers. Questions: 1. Is ISP server location a relevant argument to explain long downloads? 2. If yes, is it best to select an ISP close to the intended audience in order to minimize loading time, avoid traffic congestion etc. or pay attention to the network? For instance, the pathping tests reveal that some data lines are significantly lossy at certain periods of the day (eg. loss rate > 10% on 10 hops out of 18). Or is this only a marginal consideration? 3. Finally, is there a benchmark that measures the overall time required to render an application? This yardstick would enable me to compare performances of the ISP services. TIA. |
|
|||
|
Re: Is an ISP's geographical location relevant?
there are a lot of issues. you should use a tool to measure download
time (firebug in firefox, dev bar in safari) that times the downloads. the number of files can make a difference, also caching. check the latency (need fiddler for this). compare your times to google. asp.net has a idle timeout the causes it to be unloaded after 20-30 minutes of no use. your isp can change this setting. the next page hit will cause the site to reload and re-jit (compile). how long this takes depends on the number of dll's in bin, the number of pages requiring compile. you can get startup quicker by precompiling the pages to dll's (but a jit will still be required). -- bruce (sqlwork.com) helveticus wrote: > My site is hosted in the US (West Coast). Lately, I noticed > significant loading delays especially in the afternoon (I am currently > located in Belgium/NW Europe). The main page is not exceptionally > heavy on resources (html + CSS + js libraries + images = approx 90K), > yet quite often up to 40s.. 90 s are required to download the page > over DSL, especially in the afternoon. > > The main page is essentially pure html (no db access). The ISP > attributes the delay to the fact that the server is busy loading and > re-compiling the application prior to rendering the first page. To > minimize this, they provide the possibility of keeping the app in > server memory against a yearly fee. > > I am partially skeptical of their answer. I thought that once ASP > resources were compiled, they'll stay so afterward. But, I also > noticed that some ISP's recognize that delays may be caused by data > being transmitted over long wires. A leading US ASP ISP recently has > decided to open a date-center off-shore (UK) to better serve its > European customers. > > Questions: > 1. Is ISP server location a relevant argument to explain long > downloads? > > 2. If yes, is it best to select an ISP close to the intended audience > in order to minimize loading time, avoid traffic congestion etc. or > pay attention to the network? For instance, the pathping tests reveal > that some data lines are significantly lossy at certain periods of the > day (eg. loss rate > 10% on 10 hops out of 18). Or is this only a > marginal consideration? > > 3. Finally, is there a benchmark that measures the overall time > required to render an application? This yardstick would enable me to > compare performances of the ISP services. TIA. |
|
|||
|
Re: Is an ISP's geographical location relevant?
Thanks Bruce. I ran tests on Firebug and Fiddler. Basically, the time-
out bottleneck occurs when calling up the app ie when the default page loads. Afterward, the app runs smoothly. I'll try to pre-compile the code to see if this helps. Currently, all the code resides in the App_Code directory. Final question: Do you think I should go ahead with their offer and have the app stay current in the server (I presume this it what it amounts to.) ? I don't want to be tricked into buying services that yield no improvement. Thanks. |
|
|||
|
Re: Is an ISP's geographical location relevant?
Going to avoid parroting Bruce here, although I concur with his advice.
"helveticus" <msch.prv@gmail.com> wrote in message news:6e4c5d74-9996-4d1f-adae-fa268c13bdcb@q9g2000hsb.googlegroups.com... > My site is hosted in the US (West Coast). Lately, I noticed > significant loading delays especially in the afternoon (I am currently > located in Belgium/NW Europe). The main page is not exceptionally > heavy on resources (html + CSS + js libraries + images = approx 90K), > yet quite often up to 40s.. 90 s are required to download the page > over DSL, especially in the afternoon. > > The main page is essentially pure html (no db access). The ISP > attributes the delay to the fact that the server is busy loading and > re-compiling the application prior to rendering the first page. To > minimize this, they provide the possibility of keeping the app in > server memory against a yearly fee. > > I am partially skeptical of their answer. I thought that once ASP > resources were compiled, they'll stay so afterward. But, I also > noticed that some ISP's recognize that delays may be caused by data > being transmitted over long wires. A leading US ASP ISP recently has > decided to open a date-center off-shore (UK) to better serve its > European customers. Assume you mean ASP.NET? If the worker thread restarts, you can end up with recompilation on pages. In the config, there are settings for worker thread restart thresholds. In a corporate owned machine, you normally do not set these too high. This is in addition to what Bruce mentioned. Personally, however, I think their answer is probably BS. The extra load time is experienced on a page by page basis in ASP.NET 2.0+, so a page with nothing should only take a second, unless you have tons of libraries loaded with the site. The lag you are experiencing is more likely having to go across the Atlantic pipe to get to you in Europe. > Questions: > 1. Is ISP server location a relevant argument to explain long > downloads? Yes, the pipes across the Atlantic are much smaller than the backbone in the United States and will affect times. How much depends on a variety of factors, including current load. > 2. If yes, is it best to select an ISP close to the intended audience > in order to minimize loading time, avoid traffic congestion etc. or > pay attention to the network? For instance, the pathping tests reveal > that some data lines are significantly lossy at certain periods of the > day (eg. loss rate > 10% on 10 hops out of 18). Or is this only a > marginal consideration? In general, this is not an issue, but when you are talking cross continent, it would be useful to have your sites on a box on the same side of the pond, especially when kids get home and start downloading illegal movies from Russia. ;-) > 3. Finally, is there a benchmark that measures the overall time > required to render an application? This yardstick would enable me to > compare performances of the ISP services. TIA. There are some test tools for web apps that look at a variety of matrices, but I am not sure there is one tool that does everything you would need to properly rule everything out (noisy lines, etc.). -- Gregory A. Beamer MVP, MCP: +I, SE, SD, DBA Subscribe to my blog http://feeds.feedburner.com/GregoryBeamer# or just read it: http://feeds.feedburner.com/GregoryBeamer ******************************************** | Think outside the box! | ******************************************** |
![]() |
|
| Thread Tools | Search this Thread |
| Display Modes | |
|
|