Trader333
Moderator
- Messages
- 8,748
- Likes
- 1,026
timsk
Legendary member
- Messages
- 8,709
- Likes
- 3,413
Re. Onspeed -
I've decided to go for it and registered with them. There's a 14 day money back guarantee, so there's little to lose IMO. I did their 'pre' and 'post' test with the following results:
Loaded a colour pic' and text from server totalling 754,928 bytes.
Pre Onspeed: 144.75 seconds with a throughput of 42K on my 56K dial-up Modem.
Post Onspeed: 122.28 seconds with a throughput of 49K.
I can detect a noticable improvement just opening the pages of this BB so, assuming this continues, I'm happy. However, as the test clearly demonstrates, the improved performance isn't anywhere near as large as Onspeed suggest in their promotional hype.
HTH
Tim.
I've decided to go for it and registered with them. There's a 14 day money back guarantee, so there's little to lose IMO. I did their 'pre' and 'post' test with the following results:
Loaded a colour pic' and text from server totalling 754,928 bytes.
Pre Onspeed: 144.75 seconds with a throughput of 42K on my 56K dial-up Modem.
Post Onspeed: 122.28 seconds with a throughput of 49K.
I can detect a noticable improvement just opening the pages of this BB so, assuming this continues, I'm happy. However, as the test clearly demonstrates, the improved performance isn't anywhere near as large as Onspeed suggest in their promotional hype.
HTH
Tim.
Rognvald
Established member
- Messages
- 916
- Likes
- 15
timsk
under what internet conditions did you do your test - was the net busy - was your provider - do you have any contention in your connection - all these may affect the performance improvement which will probably be very variable. You can be sure the onspeed maketing claims will be based on conditions ideal for their product.
I am not trying to denigrate the product or put you off - merely to alert you to the likely variable nature of the improvement. Real time trading seems to me to be more dependent on a constantly fast transfer (and low latency) of a comparatively small amount of continuous data and not based on the one off large file downloads that I suspect this product is aimed at that.
under what internet conditions did you do your test - was the net busy - was your provider - do you have any contention in your connection - all these may affect the performance improvement which will probably be very variable. You can be sure the onspeed maketing claims will be based on conditions ideal for their product.
I am not trying to denigrate the product or put you off - merely to alert you to the likely variable nature of the improvement. Real time trading seems to me to be more dependent on a constantly fast transfer (and low latency) of a comparatively small amount of continuous data and not based on the one off large file downloads that I suspect this product is aimed at that.
Agree with Ron. Understand from another BB that Onspeed is compression software said (by the poster) to be aimed at email/large files downloads rather than data per se. Should still help though. UK-ISP claim that their dial-up MO tackles the contention (heavy traffic ) problem and recommend Onspeed on their site. Could be moving from a cable to a no-broadband area later this year myself (boo hoo) so must get a length of wire and try both of these out. Useful thread.
Rognvald
Established member
- Messages
- 916
- Likes
- 15
When I download large files from internet sites and over vpn I monitor my 100Mbps network utilisation with XP task manager and rarely achieve the full 5.12Mbps download speed (which in itself is only a small fraction of the LAN network bandwidth) - its usually just under 5Mbps and there can be quite short downturns to much lower speeds.
Using task manager to monitor my network utilisation when transferring data for charting and realtime trading I have found this ocuppies only about 10% of the potential broadband internet access bandwidth. The advantage of broadband access in rt trading/charting in my mind lies in low latency transfer and 24x7 low contention uninterrupted connection. There is also the ability to share the connection between a number of networked machines with no degradation at each machine. Of course it is very beneficial in rapid opening of web pages and rapid download of large files too!
Using task manager to monitor my network utilisation when transferring data for charting and realtime trading I have found this ocuppies only about 10% of the potential broadband internet access bandwidth. The advantage of broadband access in rt trading/charting in my mind lies in low latency transfer and 24x7 low contention uninterrupted connection. There is also the ability to share the connection between a number of networked machines with no degradation at each machine. Of course it is very beneficial in rapid opening of web pages and rapid download of large files too!
Rognvald
Established member
- Messages
- 916
- Likes
- 15
Paul,
Unfortunately I don't think its that straight forward.
I have found that, for example, Updata (who access data from Comstock)and who allow access to a large eod database as well as intraday downloading and upates, "broadcast" their database files at different speeds. Some download at a nominal 512k and others at a nominal 256k. This is for large file downloads. I asked them about this and was advised it was a commercial decision based on their isp charges. This may have changed recently though. For thr steaming download their bandwidth use is very low.
Most financial datafeed providers use complex compression algorithms to reduce their bandwidth requirement and increase their bandwidth usage efficiency - effectively compressing data at their end and decompressing at yours (your system processor/memory/disc read/write access speed is the limiting factor). After decompression the data has to be handled by your system. Result - some small delay (and loss of detail) over and above that caused by other internet factors such as those i first mentioned. Also there is an issue over how many instruments your software collects rt data for at anyone time.
Tiscali have recently intoduced a cheap "broadband" connectionwhich I believe is about 128K. This is advertised at the same sort of monthly cost as the original 24x7 56k connections
It is difficult to say positively that 512 is OTT compared to 128 though. I think that 128 would probably be sufficient for most purposes. It also depends on where your 128 connection comes from. If it is a 128k ISDN line then it will almost certainly be better than a Tiscali type adsl limited offering, as you will always get 128K bandwidth on your connection but you still have no control over the other internet/data aspects. With the tiscali type option the max you get is 128 and the contention is probably high resulting in potentially very low speeds
Remember that the 512 down and 256 up is the maximum you will achieve. As use (contention) increases the bandwidth available to you goes down as you are sharing it with others. This is sometimes visible at busy times of day when down loads slow. However on the positive side the straeming data download is not so demanding (although it is always there)and so you maaaay be using far less bandwidth than the kid next door downloading for his homework - your use is constant - his is sporadic.
Not all 512 connections achieve the same speed on a routine basis not all provider achieve the same results and there can be substantial differences
Have a look at
http://www.adslguide.org.uk/
and try the speed test at
http://www.adslguide.org.uk/tools/speedtest.asp
There is also the question of the quality of transmission from your exchange and here the bandwidth can be further diluted in an effort to reach out further (rdsl) - your individual requirements and location etc are of some significance
For me reliability is just as great an issue as speed
Sorry if this has turned into a rant but you did ask
Unfortunately I don't think its that straight forward.
I have found that, for example, Updata (who access data from Comstock)and who allow access to a large eod database as well as intraday downloading and upates, "broadcast" their database files at different speeds. Some download at a nominal 512k and others at a nominal 256k. This is for large file downloads. I asked them about this and was advised it was a commercial decision based on their isp charges. This may have changed recently though. For thr steaming download their bandwidth use is very low.
Most financial datafeed providers use complex compression algorithms to reduce their bandwidth requirement and increase their bandwidth usage efficiency - effectively compressing data at their end and decompressing at yours (your system processor/memory/disc read/write access speed is the limiting factor). After decompression the data has to be handled by your system. Result - some small delay (and loss of detail) over and above that caused by other internet factors such as those i first mentioned. Also there is an issue over how many instruments your software collects rt data for at anyone time.
Tiscali have recently intoduced a cheap "broadband" connectionwhich I believe is about 128K. This is advertised at the same sort of monthly cost as the original 24x7 56k connections
It is difficult to say positively that 512 is OTT compared to 128 though. I think that 128 would probably be sufficient for most purposes. It also depends on where your 128 connection comes from. If it is a 128k ISDN line then it will almost certainly be better than a Tiscali type adsl limited offering, as you will always get 128K bandwidth on your connection but you still have no control over the other internet/data aspects. With the tiscali type option the max you get is 128 and the contention is probably high resulting in potentially very low speeds
Remember that the 512 down and 256 up is the maximum you will achieve. As use (contention) increases the bandwidth available to you goes down as you are sharing it with others. This is sometimes visible at busy times of day when down loads slow. However on the positive side the straeming data download is not so demanding (although it is always there)and so you maaaay be using far less bandwidth than the kid next door downloading for his homework - your use is constant - his is sporadic.
Not all 512 connections achieve the same speed on a routine basis not all provider achieve the same results and there can be substantial differences
Have a look at
http://www.adslguide.org.uk/
and try the speed test at
http://www.adslguide.org.uk/tools/speedtest.asp
There is also the question of the quality of transmission from your exchange and here the bandwidth can be further diluted in an effort to reach out further (rdsl) - your individual requirements and location etc are of some significance
For me reliability is just as great an issue as speed
Sorry if this has turned into a rant but you did ask
Last edited:
I enquired on pcpitstop and was directed to this discussion, http://pcpitstop.ibforums.com/index.php?act=ST&f=8&t=35591
I'm afraid a lot of the tech stuff is over my head :cheesy: but some may find it helpful.
I'm afraid a lot of the tech stuff is over my head :cheesy: but some may find it helpful.
timsk
Legendary member
- Messages
- 8,709
- Likes
- 3,413
Ron, Oatman et al,
I'm afraid I'm going to have to bow out of this thread as it's over the top of my head! I downloaded Onspeed just before Christmas and, in the short time that it's been operational, I have noticed a small improvement in speed. However, it's nothing like as fast as the marketing hype would have you believe. But, at £25.00 p/a - it is cheap and every little bit helps.
Tim.
I'm afraid I'm going to have to bow out of this thread as it's over the top of my head! I downloaded Onspeed just before Christmas and, in the short time that it's been operational, I have noticed a small improvement in speed. However, it's nothing like as fast as the marketing hype would have you believe. But, at £25.00 p/a - it is cheap and every little bit helps.
Tim.
I was just reading Fred Langa, a free newsletter, in which a subscriber has written in about pcbooster and "other" computer speed fixes. It's No. 5 on the list under "Fraud?"
http://langa.com/newsletters/2004/2004-01-12.htm#1
http://langa.com/newsletters/2004/2004-01-12.htm#1
