Article View: bit.listserv.nettrain
Article #446LONG Answers: Does inactive Netscape consume bandwidth?
From: Stephanie John
Date: Fri, 02 Oct 1998 00:00
Date: Fri, 02 Oct 1998 00:00
237 lines
10973 bytes
10973 bytes
I've gotten a lot of VERY interesting replies. I've stripped out identifying information and collated them below. Many thanks to the wonderful people of NETTRAIN who came through when no other source did! Bryan Williams, Paul Groll, H. Bork, James Klock, Ian Samson, James Barrentine, Steve Hunt . Responder #1 said: Those guys have you hook, line and sinker! The only way in my opinion that a browser can use bandwidth is if a page laced with changing banners and graphics updates is left open on the desktop, or even minimized. In this way, they are right. However, load a blank page, and the browser sits idly waiting for further instructions. The number of simultaneous connections has nothing to do with the problem at hand (again, in my opinion, I could be wrong, of course), but I have been working with browsers since 1990 and I have not come across any resistance yet, nor seen anyone else complaining. Responder#2: You might want to ask "the computer guys" how it is possible that bandwidth is being used, when no packets are being sent or switched at the router. There are tools that can be used on the server that are part of most OS's that show an IT person exactly how much bandwidth is being used, and in most cases by node. Ask them to show you this mysterious use of bandwidth when a desktop simply has the browser open to say a page that is cached locally. If the browser is left open to a page on the Internet and activity ceases, there is an infinitesimally small amount of bandwidth being used to support that browser. The problem may be more related to the ISDN service itself. Is it single channel (64k) service or two channel? How many people are there on the network that could potentially be using IP simultaneously? Even with two channel (128k) service, when more than a dozen people are online, things will creeeeeeeep. If the network is supporting TCP/IP is there an intranet set up on the server so that everyone can default their homepage there. This will allow you to keep the browser open with zero drain on the ISDN connection. Push is a no no. Responder #3: Off the top of my head....beans. I'd need to see data to support that. That's like saying that when your car is in PARK with the brake set, you still have to turn off the motor to keep it from driving around by itself. I've never heard anything like this, and would need to hear the explanation. NOT off the top of my head...I don't know everything, so I've kicked this question to the team here. I'll let you know what they say. (but I'll be surprised AND I'll learn something new if it's any different than my answer...) ( .... later that week came the answer:) Word is back from the troops, and the answer is unanimous -- bull. A browser is stateless, reactive (unless push technology is working!), event driven (unless someone clicks on something, there are NO events!), etc. There is no way a dormant browser is using bandwidth. If anyone tells you different, then this whole pack of experts here has something interesting to learn, moi included. Now, if they mean IP addresses, and not bandwidth, then that's a different question. But if the issue is your bandwidth on the ISDN, nuh-uh. Responder #4: I do not _believe_ (so sorry) the navigator would receive packets from outside, if there's no 'active webpage loaded. On the other hand there's plenty of pages with flickering images and texts etc which are flickering as they permanently receive packets. Lately, I even had an image looking like W95 system window with a bar slowly progressing and the text said: Our remote web server is just checking your system, just wait a minute... IMHO, most graphics are a really unnecessary load, and many HTML-pages do include them and keep your packet stream going _at_least_ as long as your red spot does not extinguish. Another packet reason could be your integrated mail client programme, which usually checks the corresponding mail server every soandso seconds. Just add up, how many PCs are running simultaneously... (Note from Stephanie: the status line says "Document: Done", my hard drive doesn't move, and my Netscape meteors are still, and my assorted Internet Utitilities tell me I'm pulling in no data. I keep my browser at our library's home page when not in use, or on Alta Vista's Advanced Search screen as long as there's nothing that keeps the status bar pumping.) Another reason could be your operating system. NT e.g. produces lots and lots of background packets, checking which other NT PCs there are, and probably W95 is not a bit better. I do not have any doubts, that each additional application programme adds lots of additional packets up to these. All these packets do not show up, they say nothing but just confirm from PC to PC: I'm OK, you're OK. And there's more of them with a running navigator than without. And finally, each application (word processor, mail client, browser, many backgrounds) slows down your PC and the application programme you are really using. As they have to share the memory. Maybe applications are sisters and brothers in spirit, but there's only ONE memory for all of them to share. [ Otherwise, in case applications should interfere, it's just another risk... (No risk no fun ? :) ] Sorry, if this should look like another 'think so'. Try a ping series to your router IP address with your PCs on and with your PCs off (IP and browsers respectively.) But be careful in diagnosis: for a good interpretation, you must know which PCs are in a common collision domain, i.e. on the same hub. Nets aren't that easy, I am sorry to say. Network diagnosis is even more complicated than single PC diagnosis, and quite expensive... The only way would be counting packets with a sniffer. Lots of work. Perhaps simpler to just click on and off from time to time ? Responder #5: Does it tie up the network in ANY way? Not as such. However, some sites use any of several technologies (streaming audio and/or video and animated gifs are the first two that spring to mind) that DO pull in data even while "idle". If you're truly leaving the browser just staring at static HTML, there should be no interaction between your machine and you network at all, let alone anything past the router. If, on the other hand, you're leaving the browser pointed at a site that includes ANY dynamic data (which may include banner advertisements, cheesy "This site is under construction" comments, or even just scrolling status bars) you will be adding load to both your ethernet AND to the ISDN line. I have not done anything like a scientific survey of the proportion of sites that do or don't continue to feed information after the body of the site has been received, but my rough guess would be that sites that have no impact on bandwidth after the body has been read are slightly in the minority. Hope this helps! Responder #6: No they do not. However, other things might. I assume you are running Windows of some flavor. Windows has a habit of periodically checking the DNS servers you have configured to see if they are still there. Also if you are running IPX/SPX your router may bring up the ISDN line when it sees SAP packets. Again, depending on router configuration, the line may be brought up when certain Microsoft broadcasts are done by NetBEUI. A protocol analyzer will quickly tell you what's going on. Responder #7: Not sure what you mean by tie up the network. Does it use all the bandwidth of your ISDN connection? No, not unless you're pulling a video feed or something massive like that. Does it use any bandwidth at all? It might, depending on the website. Some auto- refresh data, especially banner ads. ISDN is usually charged by the minute, so maybe your people are concerned that the ISDN connection will stay up longer than necessary, incurring excessive charges. __________________ So. What am *I* going to do? I'm going to keep my Netscape open. I found myself turning it off and on about 4 times an hour. I will continue to open a new window, too, to keep my place in one search on one database while I compare it with results in another. I admit I *was* doing something bad when the network was slow, because I was assuming it was the Internet as it usually happened after 3 p.m.: I would open a new window while the first one was still pulling things in and use the second window to ask for MORE stuff, then do something else if I was waiting for both to 'fill'. Yes, that was thoughtless - I didn't think that it could be others in the building as historically we didn't have much traffic but the library. But if I don't keep one static window open to compare with another, I have to print out the first window's results, run to another room to get the print, then come back, do the same search on a different vendor. And sometimes I have to flip through page after page looking for a phrase I can find with a quick CTRL+F in the browser. For example, yesterday I was comparing prices, formats, system requirements, add-ons and options on a piece of software - and using this "placeholder" approach and the "Find" function made it VERY easy... Stephanie *%*%*%*%*%*%*%*%*%*%*%*%*%*%*%*%*%*%*%*%**%*%* Stephanie C. John, M.Ln., D.M.,A.H.I.P. Director, Saginaw Health Sciences Library Saginaw Cooperative Hospitals, Inc. 1000 Houghton Ave., Saginaw MI 48602-5398 Phone: 517/771-6846 johnste@pilot.msu.edu Fax: 517/771-6898 Library website: www.chm.msu.edu/schilib/welcome.html
Technical Details
Message-ID:
<01BDED20.CC443380@schi71.schi.msu.edu>Path:
arc1-ca.pugleaf.net!unknown!not-for-mailShow Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 10449f,688ccbc174ff58f9,start X-Google-Attributes: gid10449f,public From: Stephanie John <johnste@PILOT.MSU.EDU> Subject: LONG Answers: Does inactive Netscape consume bandwidth? Date: 1998/10/02 Message-ID: <01BDED20.CC443380@schi71.schi.msu.edu> X-Deja-AN: 397071534 Sender: Internet/BITNET Network Trainers <NETTRAIN@LISTSERV.ACSU.BUFFALO.EDU> Content-Transfer-Encoding: quoted-printable Approved: NETNEWS@AUVM.AMERICAN.EDU * Comments: ******************************************************** Approved-By: "James G. Milles" <jgm3@TOASTER.LAW.CWRU.EDU> Content-Type: text/plain; charset="us-ascii" X-Received: by schi71.schi.msu.edu with Microsoft Mail id <01BDED20.CC443380@schi71.schi.msu.edu>; Thu, 1 Oct 1998 09:49:35 -0400 ReSent-Subject: LONG Answers: Does inactive Netscape consume bandwidth? MIME-Version: 1.0 Reply-To: Stephanie John <johnste@PILOT.MSU.EDU> Newsgroups: bit.listserv.nettrain