Message boards :
Number crunching :
A temp fix for our bandwith problems ?
Message board moderation
Previous · 1 · 2 · 3 · 4 · Next
Author | Message |
---|---|
OzzFan Send message Joined: 9 Apr 02 Posts: 15691 Credit: 84,761,841 RAC: 28 |
I dont know what guestimating is. Maybee some flagwaving will explain. Guestimating is a portmanteau of "guess" and "estimating". Even if we can't calculate the actual effect, does it not stand to reason that we shouldn't be abusing the system and/or other users by even a small amount? |
Lint trap Send message Joined: 30 May 03 Posts: 871 Credit: 28,092,319 RAC: 0 |
Without facts to help us make a determination, there is no direct evidence that any user abuse or DDOS is being caused solely by use of the Retry button. Lt |
OzzFan Send message Joined: 9 Apr 02 Posts: 15691 Credit: 84,761,841 RAC: 28 |
So it's completely impossible that even a portion of users or computers have ever been affected by someone hitting the retry button. Like I said, people will quickly rationalize that their actions don't affect the system and what they are doing isn't hurting anyone else because they don't see it. |
janneseti Send message Joined: 14 Oct 09 Posts: 14106 Credit: 655,366 RAC: 0 |
I dont know what guestimating is. Maybee some flagwaving will explain. You are right. Abusing a system is a bad thing. But there are many users like my self who does not use their computers 24/7 for this project. I prefer to conclude downloads from SETI by pressing the retry button if necessary when shutting down the system. I have a long experience using the FTP protocol (25 years) and I am sure (dead sure) that these extra requests are not giving any extra load on the system. It is exactly what the rescheduler program is doing but in an other time. |
Lint trap Send message Joined: 30 May 03 Posts: 871 Credit: 28,092,319 RAC: 0 |
So it's completely impossible that even a portion of users or computers have ever been affected by someone hitting the retry button. No, not completely impossible. I said there's no direct evidence. I feel it's more likely that some comm problems are caused by hosts that are in need of work and are repeatedly asking for work. My 6.10.60 pc does that every five minutes. And has been doing that all day, every day since it hasn't been able to fill its 4 day cache. How many hosts are doing that continuously?? Setting NNT or disabling Network Activity will stop those requests. Lt |
OzzFan Send message Joined: 9 Apr 02 Posts: 15691 Credit: 84,761,841 RAC: 28 |
You are right. Abusing a system is a bad thing. I agree that there's a lot of users who don't leave their computers on 24/7, but that doesn't automatically translate into being required to press the retry button to force downloads through. During the summertime when the heat is high I usually turn off all my machines during the day. If a download is stuck, it will pick up on its own after I boot my system. I don't need to press retry just to get it to work. I have a long experience using the FTP protocol (25 years) and I am sure (dead sure) that these extra requests are not giving any extra load on the system. SETI@Home isn't using the FTP protocol though, they are using HTTP via TCP/IP. And by the way, I have 20 years of experience as a tech, 10 of those as a SysAdmin. I can assure you that these extra requests are having a negative effect on the servers. Dead sure. |
OzzFan Send message Joined: 9 Apr 02 Posts: 15691 Credit: 84,761,841 RAC: 28 |
No, not completely impossible. I said there's no direct evidence. I think there is direct evidence in that people aren't getting through and they feel the need to force it through by pressing retry. Put enough of these users together and they can have a significant impact. I feel it's more likely that some comm problems are caused by hosts that are in need of work and are repeatedly asking for work. My 6.10.60 pc does that every five minutes. And has been doing that all day, every day since it hasn't been able to fill its 4 day cache. How many hosts are doing that continuously?? There may be some bug in the comms logic in BOINC, I'm not disputing that. My only argument is that people should not willingly make the situation worse, even if it is only effecting a small portion of users by abusing the retry button. |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13746 Credit: 208,696,464 RAC: 304 |
I can assure you that these extra requests are having a negative effect on the servers. Dead sure. Once again, i agree. Once again i point out that the effect is too small to be significant. ie- if the retry button on all hosts was to be disabled, it would have no effect; there would be no improvement in uploads, downloads, scheduler requests. Grant Darwin NT |
janneseti Send message Joined: 14 Oct 09 Posts: 14106 Credit: 655,366 RAC: 0 |
[/quote] SETI@Home isn't using the FTP protocol though, they are using HTTP via TCP/IP. And by the way, I have 20 years of experience as a tech, 10 of those as a SysAdmin. I can assure you that these extra requests are having a negative effect on the servers. Dead sure.[/quote] I am aware that these extra requests are having a negative effect on the servers. But how much? Bring this to a test. I can see that in the cricket graph if you are right. |
OzzFan Send message Joined: 9 Apr 02 Posts: 15691 Credit: 84,761,841 RAC: 28 |
Once again i point out that the effect is too small to be significant. I think the effect would be > 0 and there would likely be less failures to reach the SETI servers by some participants. |
OzzFan Send message Joined: 9 Apr 02 Posts: 15691 Credit: 84,761,841 RAC: 28 |
I am aware that these extra requests are having a negative effect on the servers. But how much? Bring this to a test. I can see that in the cricket graph if you are right. If I could provide you with specific numbers, I would, but it should be obvious that the numbers are greater than zero. So while it will not solve all of the bandwidth problems, it would prevent more than a few clients from being unable to contact the SETI servers. |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13746 Credit: 208,696,464 RAC: 304 |
Once again i point out that the effect is too small to be significant. There are roughly 155,000 active users (250,000 hosts). Lets say there are a couple of hundred active users of the message boards (unlikely to be that high) & all of them hit the retry button whenever possible (even less likely). That's 0.0013% of the user base (0.0008% of the hosts base). There is no way something so small could have any significant effect. Grant Darwin NT |
OzzFan Send message Joined: 9 Apr 02 Posts: 15691 Credit: 84,761,841 RAC: 28 |
Once again i point out that the effect is too small to be significant. What about the number of users that don't visit the message boards but are still savvy enough or anxious enough to press the retry button? |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13746 Credit: 208,696,464 RAC: 304 |
What about the number of users that don't visit the message boards but are still savvy enough or anxious enough to press the retry button? I would suggest an even smaller percentage. Grant Darwin NT |
rob smith Send message Joined: 7 Mar 03 Posts: 22220 Credit: 416,307,556 RAC: 380 |
I would beg to disagree - the number of users who will hit the retry button and not visit the message boards will far outweigh those who will visit the message boards. I've seen this so many times in the "outside world" where users will try and use the equivalent to the retry button, but not pick up the phone and listen the the message that would tell them that the service they are trying to use if off-line and will be off-line for hours. Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
William de Thomas Send message Joined: 15 May 99 Posts: 17 Credit: 15,501,592 RAC: 0 |
I have a cache set for 10 days and it seems that it's never enough. Right now I am out of work and the server or router problem has not been out for 10 days. |
OzzFan Send message Joined: 9 Apr 02 Posts: 15691 Credit: 84,761,841 RAC: 28 |
What about the number of users that don't visit the message boards but are still savvy enough or anxious enough to press the retry button? Like Rob, I have to disagree. I still think the percentage is closer to .5-1.0% which is enough to cause significant delays in comms for a lot of users when the servers are already overloaded with requests from other clients. |
Ichobod Send message Joined: 20 Apr 08 Posts: 6 Credit: 2,948,560 RAC: 0 |
Please avoid suggesting that our fellow seti users have malicious intent in using the retry button. They are not causing the problem. They are reacting to it. |
OzzFan Send message Joined: 9 Apr 02 Posts: 15691 Credit: 84,761,841 RAC: 28 |
Please avoid suggesting that our fellow seti users have malicious intent in using the retry button. They are not causing the problem. They are reacting to it. Please re-read what I've posted. I never said they have malicious intent, and I agree they are reacting to it. However, I am trying to point out that in doing so, they are potentially making the problem worse. It only becomes malicious if in the light of a new view they choose to ignore the warnings and continue with their behavior in their own need to force things through. My point is that sometimes what we think is the natural and right thing to do is exactly the wrong thing to do. For example, when I worked as an IT Administrator for a small nursing home, I had to undergo protective training. One piece of that training explained that if someone is biting you, our natural reaction is to rip our flesh out of their teeth causing a tear. In fact, the best thing to do is to push your body further into their mouth to that point that it overwhelms their jaws, and to provide a para-sympathetic response which automatically forces the jaws to release. |
W-K 666 Send message Joined: 18 May 99 Posts: 19078 Credit: 40,757,560 RAC: 67 |
In 99% of cases I would agree that unnecessary button pushes usually make situations worse. But by accident a few weeks ago, after not getting any tasks for hours, I was going to switch the computer off as I was going to be out all day. Only to find it had just been allocated a few tasks, but these were very quickly into a project back-off. So I hit the retry button and although a couple failed, the first time, all were downloaded in about 5 minutes. So since then I have done it a few more times and on most but not all, the majority of tasks have downloaded with only one or two pushes. Therefore in this case I would say hit the retry button, but don't hit it repeatedly. |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.