Okay, there is a "super alexa booster" tool going around that is garbage. Use a packet sniffer, and see what it does... nothing. It pings alexa once, pings your site once, then puts on a show all day long making you think it is doing something. It is not. In fact, from wireshark, here is what the program does with alexa:
That would be pinging google. Version 7 huh? Who is "dexim-20" with that amazon id?
Looks pretty fishy. Well, I installed sparky for firefox, and took a look at what the real toolbar does. I have results for the same google request, and also a search for whitehouse, then to whitehouse.gov
First, google:
I haven't got to alexa's "ping" yet, but in a new browser session, those results are duplicated with the following change:
Notice the change to the WID. You will notice in following samples, that the WID stays the same during a browsing session. If a program were to imitate the toolbar program, it would have to have unique WID for each request and each IP.
The Unique Computer ID, is a Unique ID given to each computer after an install. It would also have to be changed. Whereas the WID is a simple 5 digit numeric number, the computer ID may have some sort of algorithm to it. It's not Base64, if anyone has any ideas, let me know.
That should also be changed every time. See the problems so far with the magical super alexa booster? Even though they have different IP's, it has outdated alexa tool bar codes, and the id's are consistently the same.
We haven't even got to the more interesting stuff in my opinion yet.
Here is an aol.com example for shits and giggles.
Very similar. Format is nothing crazy, should be easily imitated if the id's and what not could be sorted out.
The whitehouse example, this shows the google request, and then the actual page request, and shows google as the "referrer"
Some changes when using a referrer, but nothing that couldn't be handled for more simple requests like saying dmoz sent you.
Now, let's move onto the alexa "ping"
recovery.alexa.com is where this request is sent to.
The rnd= is a long decimal number, and is started at the beginning of a browser session and at some random interval.
My hesitation with this, is depending on what this number is and how it's stored (could be just seeded random sure) how does that associate with the above WID and does this even matter?
Lastly, it goes to widgets.alexa.com to what appears purely to get the information to display in the toolbar.
Get request, not a problem, doesn't even send an ID or anything, no reason not to do that just in case it somehow tracks that information.
However, the POST only happens on a search. Would appear to simply track keywords of the search and parameters sent to google. If you were to attempt to use this to do fake searches towards google and track them at alexa, this post would certainly be required. It would make sense to me this would be extremely valuable as it may help with certain keyword rankings potentially even on google depending on the real relationship between the two of them.
I posted a work around that I use on a forum PM poster to bypass time limits. This hopefully would give a headstart on a REAL alexa booster program.
I would love to help create it, this doesn't look hard... not at all.
A simple winsock program, connects to a proxy, sends out the multiple tcp http requests, doesn't even have to do anything with the data, and moves onto the next one. With winsock, this could be multithreaded very easily.
How do we get a huge list of proxies to use for this? samair and what not have tiny lists that don't scrape easily. If there were 10,000 proxies available, you could easily do 10,000 pings per hour if your connection was fast enough. It wouldn't use tons of bandwidth because you can see the requests above, they are tiny. You could just trash the information it sends back to you, that's not important.
It's also not really pinging your website, so adsense or any advertising agencies couldn't be mad because the impressions are not artificially inflated. This only would talk to alexa. There is no way for alexa to actually know if someone visited your site or not. There is no way unless you had an alexa tracker or something built in. As far as I am aware, there is no such thing.
So, if a few people would be interested in helping me test and develop, I'll work on this. You wouldn't need a major connection or anything, just windows and some development knowledge.
Let me know if this helps.
Code:
Get Request to data.alexa.com:
Pq3b012ef000L8?cli=10
&dat=snba
&ver=7.0
&cdt=alx_vw=20
&wid=9521
&act=20000000000
&ss=1024x768
&bw=1008&t=0
&amznid=dexim-20
&ttl=0&vis=1&rq=0
That would be pinging google. Version 7 huh? Who is "dexim-20" with that amazon id?
Looks pretty fishy. Well, I installed sparky for firefox, and took a look at what the real toolbar does. I have results for the same google request, and also a search for whitehouse, then to whitehouse.gov
First, google:
Code:
data.alexa.com
GET /data/UNIQUECOMPUTERID?cli=10
&ver=spkyf-1.4.5
&dat=ns
&cdt=rq=undefined&wid=30546&s=200&ref=
&url=http://www.google.com/ HTTP/1.1\r\n
I haven't got to alexa's "ping" yet, but in a new browser session, those results are duplicated with the following change:
Code:
changes
&cdt=rq=undefined&wid=19922&s=200&ref=
Notice the change to the WID. You will notice in following samples, that the WID stays the same during a browsing session. If a program were to imitate the toolbar program, it would have to have unique WID for each request and each IP.
The Unique Computer ID, is a Unique ID given to each computer after an install. It would also have to be changed. Whereas the WID is a simple 5 digit numeric number, the computer ID may have some sort of algorithm to it. It's not Base64, if anyone has any ideas, let me know.
That should also be changed every time. See the problems so far with the magical super alexa booster? Even though they have different IP's, it has outdated alexa tool bar codes, and the id's are consistently the same.
We haven't even got to the more interesting stuff in my opinion yet.
Here is an aol.com example for shits and giggles.
Code:
data.alexa.com
GET /data/UNIQUECOMPUTERID?cli=10
&ver=spkyf-1.4.5
&dat=ns
&cdt=rq=null&wid=30546&s=200&ref=
&url=http://www.aol.com/ HTTP/1.1\r\n
Very similar. Format is nothing crazy, should be easily imitated if the id's and what not could be sorted out.
The whitehouse example, this shows the google request, and then the actual page request, and shows google as the "referrer"
Code:
GET /data/UNIQUECOMPUTERID?cli=10
&ver=spkyf-1.4.5
&dat=ns
&cdt=rq=null&wid=19922&ref=&url=http://www.google.com/#hl=en&source=hp&q=whitehouse&aq=f&aqi=n1g6g*****=&fp=c807e9cc***** HTTP/1.1\r\n
whitehouse in same req:
[truncated]
GET /data/UNIQUECOMPUTERID?cli=10
&ver=spkyf-1.4.5
&dat=ns
&cdt=rq=1&wid=19922&s=200&ref=http://www.google.com/url?sa=t&source=web&ct=res&cd=1&ved=0CAo****&url=http%253A%252F%252Fwww.whiteh
Some changes when using a referrer, but nothing that couldn't be handled for more simple requests like saying dmoz sent you.
Now, let's move onto the alexa "ping"
recovery.alexa.com is where this request is sent to.
Code:
HEAD /ping/?aid=UNIQUECOMPUTERID&plugin=spkyf-1.4.5&rnd=0.792***** HTTP/1.1\r\n
HEAD /ping/?aid=UNIQUECOMPUTERID&plugin=spkyf-1.4.5&rnd=0.143***** HTTP/1.1\r\n
The rnd= is a long decimal number, and is started at the beginning of a browser session and at some random interval.
My hesitation with this, is depending on what this number is and how it's stored (could be just seeded random sure) how does that associate with the above WID and does this even matter?
Lastly, it goes to widgets.alexa.com to what appears purely to get the information to display in the toolbar.
Code:
GET /traffic/sparky/?v=1&url=google.com HTTP/1.1\r\n
also posts there
POST /traffic/rankr/?ref=http://www.google.com/#hl=en&source=hp&q=whitehouse&aq=f&aqi=n1****&oq=&fp=***** HTTP/1.1\r\n
Get request, not a problem, doesn't even send an ID or anything, no reason not to do that just in case it somehow tracks that information.
However, the POST only happens on a search. Would appear to simply track keywords of the search and parameters sent to google. If you were to attempt to use this to do fake searches towards google and track them at alexa, this post would certainly be required. It would make sense to me this would be extremely valuable as it may help with certain keyword rankings potentially even on google depending on the real relationship between the two of them.
I posted a work around that I use on a forum PM poster to bypass time limits. This hopefully would give a headstart on a REAL alexa booster program.
I would love to help create it, this doesn't look hard... not at all.
A simple winsock program, connects to a proxy, sends out the multiple tcp http requests, doesn't even have to do anything with the data, and moves onto the next one. With winsock, this could be multithreaded very easily.
How do we get a huge list of proxies to use for this? samair and what not have tiny lists that don't scrape easily. If there were 10,000 proxies available, you could easily do 10,000 pings per hour if your connection was fast enough. It wouldn't use tons of bandwidth because you can see the requests above, they are tiny. You could just trash the information it sends back to you, that's not important.
It's also not really pinging your website, so adsense or any advertising agencies couldn't be mad because the impressions are not artificially inflated. This only would talk to alexa. There is no way for alexa to actually know if someone visited your site or not. There is no way unless you had an alexa tracker or something built in. As far as I am aware, there is no such thing.
So, if a few people would be interested in helping me test and develop, I'll work on this. You wouldn't need a major connection or anything, just windows and some development knowledge.
Let me know if this helps.