Forum Moderators: open
Showing in analog browser summary as Dual Proxy. Getting hundreds of files, no pages. I banned it, but I hate to just ban things without knowing what it is. Anybody?
You just banned a fellow wmw member from your site!
Dual Proxy is the User-agent used by Starband, a satellite internet service provider, and possibly others as well. It is a by-product of their efforts to minimize satellite-link latency; This can be considerable, since the satellite signal has to travel a distance of 22,300 miles four times (89200 miles) in order for a user to request a page and have it delivered to him. Doing the math, that's 480 milliseconds of latency due to air-travel-time of the signal alone. If each page requests several image elements and other client-side-includes, the latency would be horrible!
Here's how Dual Proxy works: The user's browser requests a page. This request is sent up to the satellite and then back down to a server program at the ISP's earth station. This server program then sends the request to the internet and waits for a response. If the returned document is HTML or SHMTL, the server program parses it, looking for any included elements such as images, external javascripts, etc. The server program then issues requests for those elements directly to the internet - before the user's browser has even requested them. When these elements have been received by the server program, the whole page and all of its includes are sent up to the satellite and back down to the requesting user. This earth station client is thus acting as the user's
browser's proxy in requesting page elements. On the user's machine, a client program temporarily caches the as-yet-unrequested page elements until the user's browser actually asks for them.
What you should see in your logs for legitimate Dual Proxy users is an initial page request with a normal browser User-agent. Subsequent requests for elements included on that page will show a Dual Proxy UA. Note that you may see subsequent Dual Proxy requests without the initial "normal" page request if the earth station caches your page but not all of the included elements. You can control this (to some extent) by making sure your server provides cache control information in the headers served with your files.
I live out in the country, get about 18400 bps throughput with an analog modem, and have no opportunity for DSL, cable, LMDS, or any other high-speed internet service. For me, satellite is the only choice. And if I try to load your page, you will likely see a User-agent of "Dual-Proxy". Please examine your logs and see if this UA is really a "bandwidth hog". And if not, please don't ban me!!!
Thanks! - Jim
On the user's machine, a client program temporarily caches the as-yet-unrequested page elements until the user's browser actually asks for them.
That's the whole problem. :(
>You just banned a fellow WebmasterWorld member from your site!
Sorry, Dual Proxy was not requesting webpages, just hundreds and hundreds of files. I have a website. It's there for people who want to view pages.
JDM: "On the user's machine, a client program temporarily caches the as-yet-unrequested page elements until the user's browser actually asks for them."
KM: "That's the whole problem."
No, these are elements of the page that was requested. Basically, the earth station server parses the page and requests all page elements before sending them to me over the high-latency satellite path. From the page server's viewpoint, this is no different than what a "real" browser does. The difference is that it takes place at my ISP's end - and saves me from waiting a half-second for every single graphic image on your page. It makes an enormous difference on graphics-heavy sites that use hundreds of 1x1-pixel image spacers and such.
This can be best appreciated if you consider that my download speed is around 680kpbs, but the best ping time I can ever get is 500 millisconds - half a second MINIMUM.
Now there is the same potential for abuse here as with any ISP. If the user is doing something bad to your site, then ban him by IP (they are static). But consider that you may cut off a considerable number of rural internet addicts (and shoppers), several of the Hawaiian Islands, and much of the state of Alaska before you ban by UA...
Also, running a robot, web server, or any automated agent on the client end of a Starband link is expressly forbidden by their TOS, so report these abuses if you see them!
I understand your concern, since I discovered Dual Proxy's significance when I banned it myself using .htaccess! I thought some agent was following me around requesting elements of pages I had requested - I assumed that I had those elements already in my browser or local network proxy cache - and thought some "weirdo" was following my activities. So I banned myself! :)
I'd be interested in knowing if your pages contain a lot of preloads or stuff like that which would trigger Dual Proxy into loading a bunch of "extra" stuff. When I look at my own accesses to my own sites, I never see it loading anything except the page I request and the graphic and script files that go with that page. AOL's cache is a far worse abuser of my site than Dual Proxy... I just don't want Dual Proxy lumped in with Larbin and Indy Library in wmw member's minds - otherwise, I'm back to 18400 bps analog modem surfing and own a $700+ pile of satellite techno-junk!
Dual Proxy is my proxy, and I'm stuck with it. If *I* abuse your site, ban my
IP. But if not, please don't.
Anyway, thanks for considering it.
Jim
