Google Windows Web Accelerator

##########
Google Web Accelerator sends requests for web pages, except for secure web pages (HTTPS), to Google, which logs these requests. Some web pages may embed personal information in these page requests.

·Google receives and temporarily caches cookie data that your computer sends with webpage requests in order to improve performance.
##########
More data more data more data, wherever they can get it from. -- So far I saved 0.2 seconds. Ahhh... finally all that leisure time computers have been promising us!

I wonder how it "prefetches" in IE? Same algo as Mozilla? (just great - more ways to figure out how to suck our bandwidth...)

also, [webaccelerator.google.com...]

I think this has huge implications for webmasters - but not just bandwidth

Could be a big issue for a complex dynamic site - such as a forum - every link prefetched, every prefetched page being rebuilt on the server* - all that extra load...

(* I've got server caching built in to my sites where ever possible but there are limits...)

If this gets to be the norm how much extra load are we going to have to deal with?

Wondering if prefetched queries show up in the logs differently...

IP 72.14.194.33
OrgName: Google Inc.
OrgID: GOGL
Address: 1600 Amphitheatre Parkway
City: Mountain View
StateProv: CA
PostalCode: 94043
Country: US

NetRange: 72.14.192.0 - 72.14.207.255
CIDR: 72.14.192.0/20
NetName: GOOGLE
NetHandle: NET-72-14-192-0-1
Parent: NET-72-0-0-0-0
NetType: Direct Allocation
NameServer: NS1.GOOGLE.COM
NameServer: NS2.GOOGLE.COM
Comment:
RegDate: 2004-11-10
Updated: 2004-11-10

TechHandle: ZG39-ARIN
TechName: Google Inc.
TechPhone: +1-650-318-0200
TechEmail: arin-contact@google.com

OrgTechHandle: ZG39-ARIN
OrgTechName: Google Inc.
OrgTechPhone: +1-650-318-0200
OrgTechEmail: arin-contact@google.com

Time to rewrite my 403 error pages to let people know why they are blocked....

I setup a test domain here and did a cursory check of all the usual suspects and couldn't see any telltale signs.

+++GET 1+++
GET / HTTP/1.1
Accept: image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, application/vnd.ms-excel, application/vnd.ms-powerpoint, application/msword, application/x-shockwave-flash, */*
Accept-Language: en-us
Accept-Encoding: gzip, deflate
User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; .NET CLR 1.1.4322)
Host: www.webmasterworld.com
Cookie: lastvisitinfo=22-; WebmasterWorldRocks=.
Connection: keep-alive

+++RESP 1+++
HTTP/1.1 200 OK
Date: Wed, 04 May 2005 20:55:58 GMT
Server: Apache/1.3.26 (Unix) FrontPage/5.0.2.2510
Cache-Control: max-age=0
Pragma: no-cache
X-Powered-By: BestBBS v3.33
Transfer-Encoding: chunked
Content-Type: text/html

+++GET 2+++
GET /web/phdss/WebmasterWorldgfx/dlogo.png HTTP/1.1
Accept: */*
Referer: [webmasterworld.com...]
Accept-Language: en-us
Accept-Encoding: gzip, deflate
User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; .NET CLR 1.1.4322)
Host: showcase.netins.net
Connection: keep-alive

+++GET 3+++
GET /web/phdss/WebmasterWorldgfx/genie_knows.jpg HTTP/1.1
Accept: */*
Referer: [webmasterworld.com...]
Accept-Language: en-us
Accept-Encoding: gzip, deflate
User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; .NET CLR 1.1.4322)
Host: showcase.netins.net
Connection: keep-alive

+++RESP 2+++
HTTP/1.1 200 OK
Date: Wed, 04 May 2005 20:55:58 GMT
Server: Apache/1.3.26 (Unix) PHP/4.3.10
Last-Modified: Sun, 12 Jan 2003 07:57:41 GMT
ETag: "f4493-bb5-3e211ff5"
Accept-Ranges: bytes
Content-Length: 2997
Content-Type: image/png
+++CLOSE 2+++

+++RESP 3+++
HTTP/1.1 200 OK
Date: Wed, 04 May 2005 20:55:58 GMT
Server: Apache/1.3.26 (Unix) PHP/4.3.10
Last-Modified: Mon, 21 Mar 2005 14:17:19 GMT
ETag: "7545d4-1d51-423ed76f"
Accept-Ranges: bytes
Content-Length: 7505
Content-Type: image/jpeg
+++CLOSE 3+++
+++CLOSE 1+++

provides a nice fast proxy for site downloaders/rippers

Yep, my first thought as well...

Please don't use any automated query scripts on our website.

Google does:

Implements prefetching of pages through automated processes to everyone else's website.

I don't understand why they need a tag to let you explicitly define a page for prefetching if that's the default?

I'd rather have some way to explicitly block prefetching.

I liked it until it messed with my cookies for our phpBB forums. I couldn't stay logged in until I stopped the accelerator.

Yep, does NOT handle cookies well as far as I can see. So any site out there depending on cookies for logins, shopping, etc. are going to have to come up with a way to keep people from thinking your site is broken when it's really their browser...

Anyone tried it on a dial up modem yet? (yes some of us still live in the sticks and the dark ages)

Also, anyone noticing unusual heavy activity by the Googlebot? Seems to be running fast and furious last few days.

How am I supposed to troubleshoot this?

Are they just using this to track browsing habits?

I wonder. Is it possible they'll pull pages from their own cache? If so then it could be fairly trivial to block it. Just wondering. I'm not really as smart as I look. ;)

If you have blocked anything (like some ads servers) you will start seeing them again with the Google Web Accelerator.

(speculation..)
This, like a ton of other things that google releases, is a one man engineering project - done in the fabled 20% of RD time that every engineer has. Two days of marketing and half an hour of testing...
(/spec)

Costs to google - a few proxy servers - probably 10 whiteboxes in the corner of a datacenter somewhere.

Benefits to google - everybody talking about google again - "google's made the internet faster - they're great!"

Benefits to user - faster browsing - 0.2secs? (anyone saved more than that? - thats almost 60 secs a year!)

Benefits to web publishers...

[edited by: gethan at 9:35 pm (utc) on May 4, 2005]

Now that is useful ;) - plus the latest batch of viri/pharming/phishing are editing lmhosts to spoof websites more effectively.

Storing copies of frequently looked at pages to make them quickly accessible

This may help Google know which pages are popular / a replacement for "link popularity" in PageRank algo.

Index 1: googlebot
index 2: google web accelerator

if Page A(index 1)!= Page A(index 2)

then Page A is CLOAKING

then Nuke Page A

now Googlebot can see pages as 'Humans' see. a new era.

Why would Google put this out? How does this help their business? Are they just using this to track browsing habits?

it's called brand building darling. Put out a few cutting edge products with no ads, get known for cool.

Google says:

Please don't use any automated query scripts on our website.

Google does:

Implements prefetching of pages through automated processes to everyone else's website.

Classic.

If so then it could be fairly trivial to block it.

deny from 72.14.192.0/20 added to the rest of banned IP's works well ;)

The users are then sent to a page explaining the Google webaccelerator is not allowed on the site. The reason why I banned it? Because as I said earlier it can be used by site rippers/downloaders (and it will be) along with other unscrupulous acts that people use proxies to hide behind. Sometimes I wonder how "out of touch" they can be with the "real" world in which they have built their business on.

Do I think they have the best SE out there? You bet, but these kind of tools when not implemented well cause more grief than good. I would think they would have been a little more in tune with things the way they are.