The Google Web Accelerator made a brief appearance in May 2005 before downloads were cut off due to a „maximum capacity of users” being reached. The Accelerator was criticized by many, with privacy and security concerns frequently expressed. Google Labs has made a reworked Accelerator available for download.
Aimed at broadband Internet users in Europe and North America, Google Web Accelerator is a tool to speed up web browsing. Available for MS Windows users, Google’s Web Accelerator (GWA) works, with a few differences, like similar accelerators available from leading ISPs.
In the following article, we review the Web Accelerator’s functionality and consider its potential impact on web sites, with a particular focus on Web Analytics reporting.
What is a Web Accelerator?
A Web Accelerator is a layer of software which works behind Internet applications to speed up the user’s Internet browsing experience. User requests, such as those from Internet Explorer or Firefox are intercepted by the accelerator which then locates the closest valid copy of the requested content. The copy might already be on the user’s computer, on the accelerator’s servers or at the content-creator’s web servers. Gains in delivery speed are achieved by
- reducing the network distance by delivering a copy of the document closest to the user
- compressing HTML pages before delivery, reducing the network bandwidth requirements
- compressing images, but this usually involves data loss, reducing image quality.
Friends call me cache
The technique of storing copies of web content closer to an end user in order to facilitate content delivery is called caching (from the French cacher, to press, hide ¹). By storing copies of content closer to users, content can be delivered more quickly and efficiently. Many people are already familiar with their Internet browser’s cache. The instantaneous response a user sees when clicking on the back button is a demonstration of caching in action. Occasional browser messages about content having expired or the need to POST something are signs of additional rules, called directives, being employed to ensure the user is receiving a valid, up-to-date, copy of the content. Accelerators are a variant on what is more commonly called a client proxy cache.
Internet service providers (ISP) have long taken advantage of caching to improve web performance and reduce costs. Image compression has not been wide spread due to quality degradation issues. Many ISPs currently offer local caching software to their dial-up customers².
Not everything is cacheable
Some Internet content is never cacheable. Examples include content traveling over a secure connection (https://), streaming media and most compressed content such as audio and video formats. Other content may be labeled as non-cacheable in http headers.
How is Google’s Web Accelerator different?
The primary innovation of Google’s Web Accelerator is the addition of logic to proactively pre-fetch content linked from a user’s current web document. GWA aims to anticipate what link a user might visit next, based on factors including where the user has moved the mouse and what linked documents others have requested. The goal is to reduce latency, the lag time between a user requesting a document and the document actually appearing, by utilizing idle network connection time to retrieve documents in advance of a user’s request.
Support for pre-fetching first appeared in the Mozilla and Firefox browsers. While the Mozilla standard requires web developers to enable each link for pre-fetching, Google’s Web Accelerator considers any link a candidate for pre-fetching.
Unlike ISP Web Accelerators, a Google Web Accelerator user does not have a direct connection to the cache server – so it is more difficult to make the case that Google can serve a new page faster than the originating site as it is not clear that Google is closer than the originating site. If the pre-fetching feature is disabled, the only advantage of browsing with the Google Web Accelerator is the compression of textual content before it arrives on the user’s computer.
Standalone web accelerators are usually aimed at the dial-up market. Google says its Web Accelerator is optimized for broadband users – yet latency, page load time, isn’t usually a significant issue for broadband users.
Improvements in release 0.2.62.80
Current Web Accelerator issues to consider
Privacy: „Please read this carefully”.
Privacy was one of the significant concerns expressed in May. To Google’s credit, privacy implications are clearly set forth during the Accelerator installation. A message in red notes: Please read this carefully. This is not the usual Yada Yada and is different from the Google Toolbar Yada Yada you may have seen before. Following this warning, Google notes:
- 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 web page requests in order to improve performance.
- In order to speed up delivery of content, Google Web Accelerator may retrieve web page content that you did not request, and store it in your Google Web Accelerator cache.
In short, the end user conveys most of their browsing information, including personal data, to Google. The notable exception is https:// connections.
Distortion of Web Analytics reports
Mass adoption of Google’s Web Accelerator could potentially impact a site’s Web Analytics reporting. While the exact impact will depend on a series of factors including a site’s web server configuration, Web Analytics reporting system configuration and general site user demographics, we can give a high-level view of what some of the potential issues are:
- Some user traffic will appear to originate from a Google host (for example 64.233.17x.*, 72.14.19x.*) rather than the user’s own ISP. This is only relevant if geographic location information is being utilized. The phenomenon is not new, many sites notice an inordinate number of users from Virgina. AOL’s proxy servers are in Virgina.
- Visitor numbers may decline. Low-end or mis-configured Analytics systems may rely solely on a user’s host to determine visitor uniqueness. If multiple users only arrive through the GWA, the site will have a decline in visitors. This issue is not new nor unique to the GWA.
- There will be an increase in page traffic as GWA pre-fetches pages which a user may not ever view. If a site is static, the site may be highly cacheable. In this particular scenario, the effect would be temporary.
- The site may have a decline in traffic as Google serves up a copy of the site’s pages instead of routing traffic directly to the site.
Distortion mitigation techniques
Forums and blogs have discussed various mitigation techniques, from banning pre-fetching (based on the X-moz: prefetch http header) to outright banning of the Google Web Accelerator based on it’s IP range. We think this is a particularly bad idea. Contact Antezeta to find out what mitigation technique would best be suited for your web site, web site users and Analytics reporting.
Site management issues
One site management issue warrants immediate attention. Site pages beyond a user login should be marked as uncacheable using appropriate web server http header cache management directives. Note: HTML http-equiv meta tags for cache control are not an appropriate solution – they often don’t work. Contact Antezeta to find out more about proper site caching techniques as a tool to improve site performance, accuracy and Web Analytics reporting.
Search Engine Optimization Implications?
The Google Web Accelerator gives Google increased visibility into Internet browsing habits. Presumably this insight will be used as one of the many factors which determine a web site page’s relevance to a particular search engine query.
Google has already been able to track browsing habits by users who have installed the Google Browser Toolbar with PageRank™ enabled. The recently launched Google Analytics also offers Google visibility it might not otherwise have. Thus, the Google Web Accelerator will augment existing methods which verify which sites users are really visiting.
Worth the trouble?
While we’re fans of many of Google’s innovations, the disadvantages of the accelerator as presently conceived seem to far outweigh any potential advantages to site owners and end-users.
Google Web Accelerator Details
|Operating system:||MS Windows|
|Browsers:||Internet Explorer, Mozilla Firefox. Others are supported with manual configuration|
|Version:||googlewebaccclient.exe: version 0.2.62.80-pintail.a googlewebaccwarden.exe: version 0.1.62.80-dogcatcher.a Firefox extension: version 18.104.22.168|
Further Information on Google Web Accelerator
- ¹ As defined by the Merriam-Webster Online Dictionary.
Google Web Accelerator
- Google Web Accelerator home page.
- Google Web Accelerator Help.
- What Webmasters Need To Know About Google Web Accelerator.
- Google Web Accelerator Support Forum.
Other Web Accelerator resources
Improper cache management, all too common, can impact both correct content delivery and web statistics.
- Duane Wessels, Web Caching (O’Reilly Media), ISBN: 1-56592-536-X, 318 pages.
- Caching Tutorial for Web Authors and Webmasters
- Brian D. Davison’s Web Caching and Content Delivery Resources
- Microsoft post and pre-check cache commands
- Tracking Search Engine Cache Page Views with Web Analytics
- Google, the spy we love