Testing upload/download bandwidth speeds for verification
I have been testing my residential ISP/DSL-Landline connections and wanted to make sure that I was getting what I am paying for. Supposedly, one can use the various website based speed test tools to determine their upload and download speeds. Are these speed test tools credible and can they be trusted? Of the several sites I have tried, they all more or less seemed to be in close agreement with one another in terms of the bandwidth speeds, i.e. my connection speed is quoted at 768KB/s up and 3MB/s down, and the farther away from central, the more reduced is the speeds are. The average speed tools says that I have measured speeds of 720-30 KB/s up and 2.0-5MB/s down. Why is it however, that when downloading software from the various Linux/M$ and other downloads sites I am seeing on average, speeds of 200-320(max) KB/s and never see anything much faster than that? Is this normal? Has anyone gotten download speeds any faster that what I have reported? What I am trying to determine is if my ISP only shows un-throttled speeds between me them, but then somehow throttles my bandwidth usage when I am using the Internet, or is it more probable that download speeds are being throttled from the download site itself? Other than by using `speed testers', I have yet to find a download site that pushes out more than 2-300KB/s? I have tried HTTP, FTP Bittorent and there is very little or no speed improvements as far as I can tell. Just wondering, Dan -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Testing upload/download bandwidth speeds for verification
Daniel B. Thurman wrote: I have been testing my residential ISP/DSL-Landline connections and wanted to make sure that I was getting what I am paying for. Supposedly, one can use the various website based speed test tools to determine their upload and download speeds. Are these speed test tools credible and can they be trusted? Of the several sites I have tried, they all more or less seemed to be in close agreement with one another in terms of the bandwidth speeds, i.e. my connection speed is quoted at 768KB/s up and 3MB/s down, and the farther away from central, the more reduced is the speeds are. The average speed tools says that I have measured speeds of 720-30 KB/s up and 2.0-5MB/s down. Why is it however, that when downloading software from the various Linux/M$ and other downloads sites I am seeing on average, speeds of 200-320(max) KB/s and never see anything much faster than that? Is this normal? Yes, very normal First, the download speed get from any site can only be as high as their upload speed. Second, run the web based speed checks from 2 or 3 different sites simultaneously and/or the same site multiple times simultaneously and see what the results are then. Those two things should shed some light as to why it is normal. Oh, and third, the software download sites probably also have rate limits on each upload (from their point of view) so that everyone gets the same level of service. All of these reasons are the driving force behind the development of bittorrent... Has anyone gotten download speeds any faster that what I have reported? What I am trying to determine is if my ISP only shows un-throttled speeds between me them, but then somehow throttles my bandwidth usage when I am using the Internet, or is it more probable that download speeds are being throttled from the download site itself? Other than by using `speed testers', I have yet to find a download site that pushes out more than 2-300KB/s? I have tried HTTP, FTP Bittorent and there is very little or no speed improvements as far as I can tell. Just wondering, Dan -- Q: What's hard going in and soft and sticky coming out? A: Chewing gum. mei-mei.gres...@greshko.com http://tw.youtube.com/watch?v=cCSz_koUhSg signature.asc Description: OpenPGP digital signature -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Testing upload/download bandwidth speeds for verification
Ed Greshko wrote: Daniel B. Thurman wrote: I have been testing my residential ISP/DSL-Landline connections and wanted to make sure that I was getting what I am paying for. Supposedly, one can use the various website based speed test tools to determine their upload and download speeds. Are these speed test tools credible and can they be trusted? Of the several sites I have tried, they all more or less seemed to be in close agreement with one another in terms of the bandwidth speeds, i.e. my connection speed is quoted at 768KB/s up and 3MB/s down, and the farther away from central, the more reduced is the speeds are. The average speed tools says that I have measured speeds of 720-30 KB/s up and 2.0-5MB/s down. Why is it however, that when downloading software from the various Linux/M$ and other downloads sites I am seeing on average, speeds of 200-320(max) KB/s and never see anything much faster than that? Is this normal? Yes, very normal First, the download speed get from any site can only be as high as their upload speed. So, does that mean I am wasting money by going from 768KB/s Up / 768KB/s Down to 768KB/s Up / 3MB/s Down since I will never obtain download speeds faster than the Upload limit of 768KB/s ??? The only way to get more speed is to increase the Upload speeds to be more closer to the Download speeds which is always higher? Perhaps I should downgrade my connection speeds to 768KB/s Up / 768KB/s Down since I cannot get higher than 768KB/s Up and I am losing $$$ or am I missing something here? I wondered why ISPs do not offer matching Up/Down speeds, so as to snare an ignorant dupe? Second, run the web based speed checks from 2 or 3 different sites simultaneously and/or the same site multiple times simultaneously and see what the results are then. Those two things should shed some light as to why it is normal. Oh, and third, the software download sites probably also have rate limits on each upload (from their point of view) so that everyone gets the same level of service. All of these reasons are the driving force behind the development of bittorrent... Has anyone gotten download speeds any faster that what I have reported? What I am trying to determine is if my ISP only shows un-throttled speeds between me them, but then somehow throttles my bandwidth usage when I am using the Internet, or is it more probable that download speeds are being throttled from the download site itself? Other than by using `speed testers', I have yet to find a download site that pushes out more than 2-300KB/s? I have tried HTTP, FTP Bittorent and there is very little or no speed improvements as far as I can tell. Just wondering, Dan -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Testing upload/download bandwidth speeds for verification
On Fri, 2009-08-14 at 08:29 -0700, Daniel B. Thurman wrote: I have been testing my residential ISP/DSL-Landline connections and wanted to make sure that I was getting what I am paying for. Supposedly, one can use the various website based speed test tools to determine their upload and download speeds. Are these speed test tools credible and can they be trusted? Of the several sites I have tried, they all more or less seemed to be in close agreement with one another in terms of the bandwidth speeds, i.e. my connection speed is quoted at 768KB/s up and 3MB/s down, and the farther away from central, the more reduced is the speeds are. The average speed tools says that I have measured speeds of 720-30 KB/s up and 2.0-5MB/s down. Why is it however, that when downloading software from the various Linux/M$ and other downloads sites I am seeing on average, speeds of 200-320(max) KB/s and never see anything much faster than that? Yes. 3 megaBITs per second is just over 300 kiloBYTEs per second. There are 8 bits per byte, plus there's packet and protocol overhead, so a 10:1 ratio between the numbers is normal. So, does that mean I am wasting money by going from 768KB/s Up / 768KB/s Down to 768KB/s Up / 3MB/s Down since I will never obtain download speeds faster than the Upload limit of 768KB/s ??? No, if you downgraded to 768 kilobit/sec service you would expect a maximum download speed of around 75-80 kilobytes per second. -Chris -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
RE: Testing upload/download bandwidth speeds for verification
Hi, I'd like to suggest a tool that I am usually using to check bandwidth speed. It is called iperf. It does not rely on usual HTTP download (most online checkers use it), but rather on pure TCP session bandwith. Your ISP maybe permits high HTTP downloads, but then throttles SSH or ESP based traffic. I think if you really want to measure, then you'll have to go with iperf. Best, Paul -Original Message- From: fedora-list-boun...@redhat.com [mailto:fedora-list-boun...@redhat.com] On Behalf Of Chris Tyler Sent: Friday, August 14, 2009 12:42 PM To: Community assistance, encouragement, and advice for using Fedora. Subject: Re: Testing upload/download bandwidth speeds for verification On Fri, 2009-08-14 at 08:29 -0700, Daniel B. Thurman wrote: I have been testing my residential ISP/DSL-Landline connections and wanted to make sure that I was getting what I am paying for. Supposedly, one can use the various website based speed test tools to determine their upload and download speeds. Are these speed test tools credible and can they be trusted? Of the several sites I have tried, they all more or less seemed to be in close agreement with one another in terms of the bandwidth speeds, i.e. my connection speed is quoted at 768KB/s up and 3MB/s down, and the farther away from central, the more reduced is the speeds are. The average speed tools says that I have measured speeds of 720-30 KB/s up and 2.0-5MB/s down. Why is it however, that when downloading software from the various Linux/M$ and other downloads sites I am seeing on average, speeds of 200-320(max) KB/s and never see anything much faster than that? Yes. 3 megaBITs per second is just over 300 kiloBYTEs per second. There are 8 bits per byte, plus there's packet and protocol overhead, so a 10:1 ratio between the numbers is normal. So, does that mean I am wasting money by going from 768KB/s Up / 768KB/s Down to 768KB/s Up / 3MB/s Down since I will never obtain download speeds faster than the Upload limit of 768KB/s ??? No, if you downgraded to 768 kilobit/sec service you would expect a maximum download speed of around 75-80 kilobytes per second. -Chris -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Testing upload/download bandwidth speeds for verification
Chris Tyler wrote: On Fri, 2009-08-14 at 08:29 -0700, Daniel B. Thurman wrote: I have been testing my residential ISP/DSL-Landline connections and wanted to make sure that I was getting what I am paying for. Supposedly, one can use the various website based speed test tools to determine their upload and download speeds. Are these speed test tools credible and can they be trusted? Of the several sites I have tried, they all more or less seemed to be in close agreement with one another in terms of the bandwidth speeds, i.e. my connection speed is quoted at 768KB/s up and 3MB/s down, and the farther away from central, the more reduced is the speeds are. The average speed tools says that I have measured speeds of 720-30 KB/s up and 2.0-5MB/s down. Why is it however, that when downloading software from the various Linux/M$ and other downloads sites I am seeing on average, speeds of 200-320(max) KB/s and never see anything much faster than that? Yes. 3 megaBITs per second is just over 300 kiloBYTEs per second. There are 8 bits per byte, plus there's packet and protocol overhead, so a 10:1 ratio between the numbers is normal. So, does that mean I am wasting money by going from 768KB/s Up / 768KB/s Down to 768KB/s Up / 3MB/s Down since I will never obtain download speeds faster than the Upload limit of 768KB/s ??? No, if you downgraded to 768 kilobit/sec service you would expect a maximum download speed of around 75-80 kilobytes per second. -Chris Duh-Oh. That makes more sense. I just posted a follow up before receiving this message and I thought I get it, but apparently I missed the boat :P Thanks for clarifying this! Dan -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Testing upload/download bandwidth speeds for verification
Paul Grinberg wrote: Hi, I'd like to suggest a tool that I am usually using to check bandwidth speed. It is called iperf. It does not rely on usual HTTP download (most online checkers use it), but rather on pure TCP session bandwith. Your ISP maybe permits high HTTP downloads, but then throttles SSH or ESP based traffic. I think if you really want to measure, then you'll have to go with iperf. Best, Paul Interesting! Thanks for the tip! -Original Message- From: fedora-list-boun...@redhat.com [mailto:fedora-list-boun...@redhat.com] On Behalf Of Chris Tyler Sent: Friday, August 14, 2009 12:42 PM To: Community assistance, encouragement, and advice for using Fedora. Subject: Re: Testing upload/download bandwidth speeds for verification On Fri, 2009-08-14 at 08:29 -0700, Daniel B. Thurman wrote: I have been testing my residential ISP/DSL-Landline connections and wanted to make sure that I was getting what I am paying for. Supposedly, one can use the various website based speed test tools to determine their upload and download speeds. Are these speed test tools credible and can they be trusted? Of the several sites I have tried, they all more or less seemed to be in close agreement with one another in terms of the bandwidth speeds, i.e. my connection speed is quoted at 768KB/s up and 3MB/s down, and the farther away from central, the more reduced is the speeds are. The average speed tools says that I have measured speeds of 720-30 KB/s up and 2.0-5MB/s down. Why is it however, that when downloading software from the various Linux/M$ and other downloads sites I am seeing on average, speeds of 200-320(max) KB/s and never see anything much faster than that? Yes. 3 megaBITs per second is just over 300 kiloBYTEs per second. There are 8 bits per byte, plus there's packet and protocol overhead, so a 10:1 ratio between the numbers is normal. So, does that mean I am wasting money by going from 768KB/s Up / 768KB/s Down to 768KB/s Up / 3MB/s Down since I will never obtain download speeds faster than the Upload limit of 768KB/s ??? No, if you downgraded to 768 kilobit/sec service you would expect a maximum download speed of around 75-80 kilobytes per second. -Chris -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Testing upload/download bandwidth speeds for verification
On Fri, 2009-08-14 at 09:21 -0700, Daniel B. Thurman wrote: First, the download speed get from any site can only be as high as their upload speed. So, does that mean I am wasting money by going from 768KB/s Up / 768KB/s Down to 768KB/s Up / 3MB/s Down since I will never obtain download speeds faster than the Upload limit of 768KB/s ??? The only way to get more speed is to increase the Upload speeds to be more closer to the Download speeds which is always higher? I suggest you re-read what Ed said, which isn't what you seem to have understood. Think of it this way: your download is the server's upload. the maximum bandwidth you can get out of the connection between them is the lower of these two numbers. However there are multiple other factors: * The server usually isn't just serving you, so it's upload speed is shared among the multiple clients it typically has at a given time. * You may be using your download link for several other things that you might not even be aware of, e.g. mail updating, browser page reloads, or even other parallel downloads (not forgetting other machines on your local net which share your ISP connection). * The intermediate connections on the net also have resource limitations. If one of these is congested, you'll see lower speeds. This is often the reason for discrepancies between the measured speed of near and far speed testing sites. * If some connections are unreliable, packets will be retransmitted, leading to a lower aggregate speed. * Your link speed (what your ISP sells you) is -- in the best case -- the speed of bits over the wire. Even if they actually give you those speeds (and there are several reasons why they might not), actual downloads also have several levels of protocol overhead, so your end-to-end data transfer will always be less than the rated speed of the connection, even if all other factors have no effect. * Lastly, downloads are almost always TCP connections. Every TCP segment you receive has to be acknowledged by a correspond ACK going the other way. The ACKs are small relative to the data segments, but they still use upload bandwidth and this can affect your data throughout. poc -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Testing upload/download bandwidth speeds for verification
On Fri, 2009-08-14 at 12:42 -0400, Chris Tyler wrote: Yes. 3 megaBITs per second is just over 300 kiloBYTEs per second. There are 8 bits per byte, plus there's packet and protocol overhead, so a 10:1 ratio between the numbers is normal. Actually not. Even discounting protocol overhead, a 3Mbps connection is 3 million (3x10^6) bits per second. 300KB of data is 300 kilobytes (300x2^10) bytes. The fact that communications and computer people have different interpretation of kilo and mega (and giga and ...) is a source of much confusion. In computing we're supposed to use kebi, mebi etc. (see http://en.wikipedia.org/wiki/Binary_prefix) for the powers of 2, but not many do. poc -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Testing upload/download bandwidth speeds for verification
It should also be noted that there is latency related to physical transmission speeds, so if the upload or download does checksums and verify handshaking, then there will be a delay of the roundtrip at the speed of light. Now this seems very fast to most folks, but electronically it is measurable, and on lines of several miles in length, it amounts to microseconds per block. If the block size is say 4K, and you download 4M, that is 1000 blocks. If the delay is 1usec, the total delay added is 1ms, or nearly 5000 bytes decrease at 5Mhz. Also there is additional overhead on normal transmissions that may not be in place on the speed test, and the speed test probably relates the bits/sec, which is not the same as the number of usable bytes, since the TCP uses quite a few bytes per block to specify various things about the transfer. All of this slows the response for actual file transfer, in addition to loading of the sending computer. On the speed tests, check both local responders and remote. I am in California, I regularly use Irvine and a system in New York. there is quite a difference. Regards, Les H On Fri, 2009-08-14 at 23:41 +0800, Ed Greshko wrote: Daniel B. Thurman wrote: I have been testing my residential ISP/DSL-Landline connections and wanted to make sure that I was getting what I am paying for. Supposedly, one can use the various website based speed test tools to determine their upload and download speeds. Are these speed test tools credible and can they be trusted? Of the several sites I have tried, they all more or less seemed to be in close agreement with one another in terms of the bandwidth speeds, i.e. my connection speed is quoted at 768KB/s up and 3MB/s down, and the farther away from central, the more reduced is the speeds are. The average speed tools says that I have measured speeds of 720-30 KB/s up and 2.0-5MB/s down. Why is it however, that when downloading software from the various Linux/M$ and other downloads sites I am seeing on average, speeds of 200-320(max) KB/s and never see anything much faster than that? Is this normal? Yes, very normal First, the download speed get from any site can only be as high as their upload speed. Second, run the web based speed checks from 2 or 3 different sites simultaneously and/or the same site multiple times simultaneously and see what the results are then. Those two things should shed some light as to why it is normal. Oh, and third, the software download sites probably also have rate limits on each upload (from their point of view) so that everyone gets the same level of service. All of these reasons are the driving force behind the development of bittorrent... Has anyone gotten download speeds any faster that what I have reported? What I am trying to determine is if my ISP only shows un-throttled speeds between me them, but then somehow throttles my bandwidth usage when I am using the Internet, or is it more probable that download speeds are being throttled from the download site itself? Other than by using `speed testers', I have yet to find a download site that pushes out more than 2-300KB/s? I have tried HTTP, FTP Bittorent and there is very little or no speed improvements as far as I can tell. Just wondering, Dan -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Testing upload/download bandwidth speeds for verification
On Friday 14 August 2009 18:23:41 Patrick O'Callaghan wrote: On Fri, 2009-08-14 at 12:42 -0400, Chris Tyler wrote: Yes. 3 megaBITs per second is just over 300 kiloBYTEs per second. There are 8 bits per byte, plus there's packet and protocol overhead, so a 10:1 ratio between the numbers is normal. Actually not. Even discounting protocol overhead, a 3Mbps connection is 3 million (3x10^6) bits per second. 300KB of data is 300 kilobytes (300x2^10) bytes. I don't understand your point. However you choose the base for M and K prefixes, the ratio is roughly 10:1. I mean, 3 Mbits = 300 KBytes because there are 8 bits in a byte (or say 10 if you simplify and/or count the overhead). It has nothing to do with prefixes. Besides, 1 Ki = 2^10 = 1024 = (roughly) 10^3 = 1 K , and similarly 1 Mi = 2^20 = (roughly) 10^6 = 1 M, (see http://en.wikipedia.org/wiki/Binary_prefix for details ;-) ...), so even if you mix them, 3Mbits = (roughly) 300 KiBytes, or any other combination you might think of. Even Gi:G is 1:1 within an 8% error. I would say Chris is completely correct. For the OP: don't worry, your ISP seems to be providing you with what you expect. HTH, :-) Marko -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Testing upload/download bandwidth speeds for verification
Hi Daniel, Daniel B. Thurman wrote: Of the several sites I have tried, they all more or less seemed to be in close agreement with one another in terms of the bandwidth speeds, i.e. my connection speed is quoted at 768KB/s up and 3MB/s down, and the farther away from central, the more reduced is the speeds are. The average speed tools says that I have measured speeds of 720-30 KB/s up and 2.0-5MB/s down. Why is it however, that when downloading software from the various Linux/M$ and other downloads sites I am seeing on average, speeds of 200-320(max) KB/s and never see anything much faster than that? Is this normal? Pardon me if this seems rather dumb or has been addressed by another post in the thread (I haven't gone through the whole thread), but are the speeds for your ISP KBps/MBps or Kbps/Mbps? Note the capitalized/small `B's. If the speeds are in Kbps/Mbps, then what you get is normal. KBps/MBps would be Kilobytes/Megabytes whereas Kb/Mb would be Kilobits/Megabits. It is common practice to quote bandwidth speeds in bits rather than bytes. To convert between the two just divide by 8, i.e. 2Mbps / 8 = 256KBps -- Suvayu Open source is the future. It sets us free. -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Testing upload/download bandwidth speeds for verification
On Fri, 2009-08-14 at 20:25 +0100, Marko Vojinovic wrote: [...] I would say Chris is completely correct. For the OP: don't worry, your ISP seems to be providing you with what you expect. I didn't suggest that what the OP is seeing isn't normal to within a reasonable margin of error. I was simply making a (pedantic) point. I should have read Chris's post more carefully (what he says is strictly speaking correct), but the distinction between the two meanings of K or M is worth bearing in mind, i.e. 300KBps does *not* mean 300 kilobytes (in the computing sense) every second, it means 300,000 bytes, which is close but not the same. That's all. poc -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Testing upload/download bandwidth speeds for verification
Suvayu Ali wrote: Hi Daniel, Daniel B. Thurman wrote: Of the several sites I have tried, they all more or less seemed to be in close agreement with one another in terms of the bandwidth speeds, i.e. my connection speed is quoted at 768KB/s up and 3MB/s down, and the farther away from central, the more reduced is the speeds are. The average speed tools says that I have measured speeds of 720-30 KB/s up and 2.0-5MB/s down. Why is it however, that when downloading software from the various Linux/M$ and other downloads sites I am seeing on average, speeds of 200-320(max) KB/s and never see anything much faster than that? Is this normal? Pardon me if this seems rather dumb or has been addressed by another post in the thread (I haven't gone through the whole thread), but are the speeds for your ISP KBps/MBps or Kbps/Mbps? Note the capitalized/small `B's. If the speeds are in Kbps/Mbps, then what you get is normal. KBps/MBps would be Kilobytes/Megabytes whereas Kb/Mb would be Kilobits/Megabits. It is common practice to quote bandwidth speeds in bits rather than bytes. To convert between the two just divide by 8, i.e. 2Mbps / 8 = 256KBps You are correct, over the wire Internet up/down speeds are Kb/s but when downloading via applications, i.e. via your download program, you see KB/s. I guess my eyes got crossed somewhere when I wrote :P -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Testing upload/download bandwidth speeds for verification
Suvayu Ali wrote: Hi Daniel, Daniel B. Thurman wrote: Of the several sites I have tried, they all more or less seemed to be in close agreement with one another in terms of the bandwidth speeds, i.e. my connection speed is quoted at 768KB/s up and 3MB/s down, and the farther away from central, the more reduced is the speeds are. The average speed tools says that I have measured speeds of 720-30 KB/s up and 2.0-5MB/s down. Why is it however, that when downloading software from the various Linux/M$ and other downloads sites I am seeing on average, speeds of 200-320(max) KB/s and never see anything much faster than that? Is this normal? Pardon me if this seems rather dumb or has been addressed by another post in the thread (I haven't gone through the whole thread), but are the speeds for your ISP KBps/MBps or Kbps/Mbps? Note the capitalized/small `B's. If the speeds are in Kbps/Mbps, then what you get is normal. KBps/MBps would be Kilobytes/Megabytes whereas Kb/Mb would be Kilobits/Megabits. It is common practice to quote bandwidth speeds in bits rather than bytes. To convert between the two just divide by 8, i.e. 2Mbps / 8 = 256KBps To clarify, a capital B is used to indicate bytes per seconds (Bps), a lower case b to indicate bits per second (bps). bps was also known back in the day as the baud rate (although that's not completely accurate). (wistful) I remember my 300 baud (300bps) acoustic coupler modem. :-p On a more serious note, some sites limit download bandwidth on each connection made so the sites can handle more connections. Example, if a site has a 2Mbps connection and one person starts a download, that person could suck the entire 2Mbps pipe, thereby blocking others from downloading. If the site puts in a 256Kbps limit on each connection, they can handle eight simultaneous connections. kernel.org does this, for instance (I think they limit to 128Kbps). Some sites also vary the limit. They start out with a large limit, but narrow it down the longer the download takes. The idea is that if you need a download of something small, you get it fast. If you're downloading a bunch of stuff, they keep squeezing down until you hit the minimum they allow and the remainder of the download continues at that limited rate. These are some of the reasons things like BitTorrent were created. -- - Rick Stevens, Systems Engineer ri...@nerd.com - - AIM/Skype: therps2ICQ: 22643734Yahoo: origrps2 - -- - Millihelen, adj: The amount of beauty required to launch one ship. - -- -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Testing upload/download bandwidth speeds for verification
Do remember that your final throughput can be influenced by many factors. One that hasn't yet been covered is the type of physical wiring you have, the age and condition of that wiring, and whether or not it is twisted pair (as in unshielded twisted pair or shielded twisted pair Category 5e network cable. Replacing the 40+ year old phone cable between my Branch Exchange Protector and my DSL splitter with Cat 5e cable resulted in a huge improvement in the quality of the voice signal my wife was getting over the phone and the data throughput rate we were getting over DSL. I didn't do any measurements but downloads completed much more quickly. Bob On 08/14/2009 11:29 AM, Daniel B. Thurman wrote: I have been testing my residential ISP/DSL-Landline connections and wanted to make sure that I was getting what I am paying for. Supposedly, one can use the various website based speed test tools to determine their upload and download speeds. Are these speed test tools credible and can they be trusted? Of the several sites I have tried, they all more or less seemed to be in close agreement with one another in terms of the bandwidth speeds, i.e. my connection speed is quoted at 768KB/s up and 3MB/s down, and the farther away from central, the more reduced is the speeds are. The average speed tools says that I have measured speeds of 720-30 KB/s up and 2.0-5MB/s down. Why is it however, that when downloading software from the various Linux/M$ and other downloads sites I am seeing on average, speeds of 200-320(max) KB/s and never see anything much faster than that? Is this normal? Has anyone gotten download speeds any faster that what I have reported? What I am trying to determine is if my ISP only shows un-throttled speeds between me them, but then somehow throttles my bandwidth usage when I am using the Internet, or is it more probable that download speeds are being throttled from the download site itself? Other than by using `speed testers', I have yet to find a download site that pushes out more than 2-300KB/s? I have tried HTTP, FTP Bittorent and there is very little or no speed improvements as far as I can tell. Just wondering, Dan -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Testing upload/download bandwidth speeds for verification
On 08/14/2009 08:29 AM, Daniel B. Thurman wrote: I have been testing my residential ISP/DSL-Landline connections and wanted to make sure that I was getting what I am paying for. Supposedly, one can use the various website based speed test tools to determine their upload and download speeds. [snip] Other than by using `speed testers', I have yet to find a download site that pushes out more than 2-300KB/s? I have tried HTTP, FTP Bittorent and there is very little or no speed improvements as far as I can tell. Just wondering, Dan Just a little aside ... If you're looking for a site that can saturate your connection, try newshosting.com as a usenet provider. They seem to be able to push as much bandwidth as you can take. They max out my 12 Mbit cable with a single connection, and they allow 20 connections per client. Of course, they aren't free. Regards John -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Testing upload/download bandwidth speeds for verification
On Fri, 2009-08-14 at 12:45 -0430, Patrick O'Callaghan wrote: * You may be using your download link for several other things that you might not even be aware of, e.g. mail updating, browser page reloads, or even other parallel downloads (not forgetting other machines on your local net which share your ISP connection). This is more important that people give it credit for. You often want to do more than one thing at once, even just browsing two websites at the same time counts. If you limit yourself, by buying a slow internet connection, you will probably regret it when it comes to trying to do your own things while the computer is downloading updates, and the like. -- [...@localhost ~]$ uname -r 2.6.27.25-78.2.56.fc9.i686 Don't send private replies to my address, the mailbox is ignored. I read messages from the public lists. -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines