Re: AOL web troubles.. New AOL speedup seems to be a slowdown
JC, I would encourage you to get more familiar with the HTTP 1.1 spec with regard to your claim of copyright infringement. I will summarize my interpretation of a part of it here. When someone provides HTTP content, they are agreeing to the protocols governing the transmission of that content, which includes caching and transformation of that content by proxy systems. Fortunately, the spec provides for netizens to send Cache-Control headers that can exclude their content from storage on and transformation by proxies. These headers are outlined in the spec, so I'm not going to detail them here. But as far as I have been able to tell, AOL is in compliance with the Cache-Control specs. I see it like this: Not including Cache-Control headers and claiming copyright infringement is like publishing a novel in English and distributing it in the U.S., but printing the copyright notice in Chinese. Clients accessing your content must be able to understand that you do not wish for it to be transformed; and since proxies speak HTTP, content providers need to include the appropriate HTTP headers so the proxies understand their wishes. Conversely, when a content provider excludes Cache-Control headers, ISPs have free reign to handle the content and deliver to the end user in whatever way they wish, as long as that way falls within the HTTP 1.1 spec. As an aside, I have a special folder on my Apache server for Jessica Stover's website where I keep images that I don't want to be compressed by all of these web accelerators (Earthlink and NetZero have them too, not just AOL). In that folder's .htacccess file, I have included instructions to send the "Cache-Control: No-Transform" header on all files requested via HTTP within that folder; those images are not modified in any way by the various web caching systems out there -- the end user gets the identical image to what is stored on my server. ~ The Gunn webmaster@jessicastover.com
AOL is copying and redistributing the image in a new format *without the permission of the copyright holder* in a way that A) makes AOL money and B) removes protections that the copyright holder had placed on the image to help keep third parties from reproducing the image without permission.
and in doing so:
IMHO they are infringing on the copyright of those who have placed the digital watermark in the image.
jc
I wouldnt have thought what happens at Layer7/8 (ie the production, copyright and distribution of the image) is related to whats going on at Layer4/5 (http) in this context. I recall this caching argument from a while back, but I dont think the subject of altering the data was in the discussion, it was around the legality of making and storing copies of copyrighted material without explicit consent. As I think the caching debate was settled without incident I think people publishing copyrighted material are happy with the situation, your suggestion that they can add in no-cache is probably not something they want to do (nor would ISPs want that in view of the performance effect it would have on their cache hit rate) Steve On Fri, 30 Jan 2004, webmaster@jessicastover.com wrote:
JC, I would encourage you to get more familiar with the HTTP 1.1 spec with regard to your claim of copyright infringement. I will summarize my interpretation of a part of it here.
When someone provides HTTP content, they are agreeing to the protocols governing the transmission of that content, which includes caching and transformation of that content by proxy systems.
Fortunately, the spec provides for netizens to send Cache-Control headers that can exclude their content from storage on and transformation by proxies. These headers are outlined in the spec, so I'm not going to detail them here. But as far as I have been able to tell, AOL is in compliance with the Cache-Control specs.
I see it like this: Not including Cache-Control headers and claiming copyright infringement is like publishing a novel in English and distributing it in the U.S., but printing the copyright notice in Chinese. Clients accessing your content must be able to understand that you do not wish for it to be transformed; and since proxies speak HTTP, content providers need to include the appropriate HTTP headers so the proxies understand their wishes. Conversely, when a content provider excludes Cache-Control headers, ISPs have free reign to handle the content and deliver to the end user in whatever way they wish, as long as that way falls within the HTTP 1.1 spec.
As an aside, I have a special folder on my Apache server for Jessica Stover's website where I keep images that I don't want to be compressed by all of these web accelerators (Earthlink and NetZero have them too, not just AOL). In that folder's .htacccess file, I have included instructions to send the "Cache-Control: No-Transform" header on all files requested via HTTP within that folder; those images are not modified in any way by the various web caching systems out there -- the end user gets the identical image to what is stored on my server.
~ The Gunn webmaster@jessicastover.com
AOL is copying and redistributing the image in a new format *without the permission of the copyright holder* in a way that A) makes AOL money and B) removes protections that the copyright holder had placed on the image to help keep third parties from reproducing the image without permission.
and in doing so:
IMHO they are infringing on the copyright of those who have placed the digital watermark in the image.
jc
your suggestion that they can add in no-cache is probably not something they want to do (nor would ISPs want that in view of the performance effect it would have on their cache hit rate)
Steve, perhaps you misinterpreted my posting, or missed the quote from JC that I referenced at the bottom. If you re-read it, you'll see that I made no such suggestion (and I did not even mention "no-cache" specifically, only no-transform). What I am saying is that a specific allegation made by another poster to this list (quoted again below) has no basis because the HTTP spec provides mechanisms for ensuring copyrighted material is not transformed or stored by proxy networks. I would never suggest to anyone to arbitrarily put a Cache-Control: no-cache header on their content. Cacheing by ISPs is a great thing for everyone and a content provider who uses no-cache is only costing themselves money in bandwidth. But, on the flip side, there are a myriad of situations that necessitate the need to control the way in which content is cached or (more commonly) compressed/transformed by a proxy; for example, high-res medical x-rays and other confidential information, consumer purchased high-res images and other copyrighted information purchased by the end user, or trademarked logos. For situations like these, I do think that it's important for content providers to know they have the ability to directly limit what the caches can do by reasonably implementing the appropriate Cache-Control headers. ~ The Gunn webmaster@jessicastover.com
AOL is copying and redistributing the image in a new format *without the permission of the copyright holder* in a way that A) makes AOL money and B) removes protections that the copyright holder had placed on the image to help keep third parties from reproducing the image without permission.
and in doing so:
IMHO they are infringing on the copyright of those who have placed the digital watermark in the image.
jc
<snipped since its kinda long> Just got done working with my mother's machine again, and have been watching her and a bunch of other people who use AOL 9.0 and some who use 8.0. Something over the past week alone has definately happened in regards to the AOL TopSpeed stuff. I've got a situation with more then 75% of the people I've tested, that they have problems running java applets (including AOL's own link into pogo games) in AOL 9.0 GM (that they are distributing to end users). When the user switches to AOL 8.0, the problem exist. When the user uses IE separate from AOL, the problem does not exist. There are other issues developing as well - random freezing of java games for example. Once again, this only happens in 9.0. This was working fine two weeks ago on all of these people's machines. Of course, this is increasing my daily workload, as I now have users having problems that I need to sit and try and diagnose. I've been telling people to use AOL 8.0 or IE if they want to play games. But, yes, there appears to be a problem somewhere with this TopSpeed stuff that people have been noting complaints about. Sorta off topic, but alot of people here also do support for this kind of stuff, and would like to get some feedback as to what others are seeing with their end users. I have a sinking feeling that when I take the time to file an official bug report/issue, they will tell me 'reformat and reinstall'. -- Brian Bruns The Summit Open Source Development Group Open Solutions For A Closed World / Anti-Spam Resources http://www.sosdg.org The AHBL - http://www.ahbl.org
Brian, I have some friends in the web proxy group at AOL, if you can send me (or post to this list) some urls that are breaking, they can take a look for you. According to them, if the java problem is happening on AOL 8.0 as well as 9.0, then it's not a TopSpeed issue (TopSpeed is just an executable that runs in tandem with 9.0), but could be some other client-related problem.. And in response to Rob's suggestion to use SSL instead of implementing cache-control, it would be a pretty wasteful implementation of SSL if its purpose is solely to prevent a proxy from recompressing your images. -- The Gunn webmaster@jessicastover.com
<snipped since its kinda long>
Just got done working with my mother's machine again, and have been watching her and a bunch of other people who use AOL 9.0 and some who use 8.0. Something over the past week alone has definately happened in regards to the AOL TopSpeed stuff. I've got a situation with more then 75% of the people I've tested, that they have problems running java applets (including AOL's own link into pogo games) in AOL 9.0 GM (that they are distributing to end users). When the user switches to AOL 8.0, the problem exist. When the user uses IE separate from AOL, the problem does not exist. There are other issues developing as well - random freezing of java games for example. Once again, this only happens in 9.0.
This was working fine two weeks ago on all of these people's machines.
Of course, this is increasing my daily workload, as I now have users having problems that I need to sit and try and diagnose. I've been telling people to use AOL 8.0 or IE if they want to play games.
But, yes, there appears to be a problem somewhere with this TopSpeed stuff that people have been noting complaints about.
Sorta off topic, but alot of people here also do support for this kind of stuff, and would like to get some feedback as to what others are seeing with their end users. I have a sinking feeling that when I take the time to file an official bug report/issue, they will tell me 'reformat and reinstall'.
-- Brian Bruns The Summit Open Source Development Group Open Solutions For A Closed World / Anti-Spam Resources http://www.sosdg.org
The AHBL - http://www.ahbl.org
"The Gunn" <webmaster@jessicastover.com> writes:
And in response to Rob's suggestion to use SSL instead of implementing cache-control, it would be a pretty wasteful implementation of SSL if its purpose is solely to prevent a proxy from recompressing your images.
To clarify (for list readers who are out of the loop because I made the point in a private communication) I suggested use of SSL in the cited hypothetical case of "for example, high-res medical x-rays and other confidential information, consumer purchased high-res images and other copyrighted information purchased by the end user". Given the HIPAA and e-commerce implications of the two named cases, using SSL would seem to be a no-brainer, and effectively renders the issue of cache-control moot. I also suggested that "The Gunn" read http://www.nanog.org/aup.html item #7 and begin posting with an account that has a real name on it. ---Rob
participants (5)
-
Brian Bruns -
Robert E. Seastrom -
Stephen J. Wilcox -
The Gunn -
webmasterï¼ jessicastover.com