Save Your Company Money In Monthly Bills Using Browser Caching

Bandwidth MoneyCompanies that operate heavily trafficked Web sites can save thousands of dollars every month by maximizing their use of browser-side caching.

Large Web sites pay for bandwidth at their Web hosting data center and also at their content delivery network (CDN, e.g. Akamai, LimeLight, CDNetworks). Bandwidth costs add up to huge monthly bills. On small-business or personal Web sites where bandwidth costs don’t go over, this is not an issue, but on large Web sites, this is important to address and monitor.

Companies operating large Web sites often have complex situations like the following:

  • An comprehensive and deep understanding of all technology cost drivers and their impacts on each other. For example, a programmer may think they are saving the company money by architecting an application in a way that it requires minimal hardware servers, but not realize that the same design actually results in even higher costs elsewhere like CDN bills.
  • Busy development teams working on multiple projects on tight timelines. This results in compromises between product features/timelines and technical/architectural best practices/standards.
  • Web content management and presentation platform(s) that have evolved over the years
  • Staff churn over the years and an uneven distribution of technical knowledge and best practices about the Web site(s)
  • The continued following of some obsolete “best practices” and standards that were established long ago when they were beneficial, but are now detrimental.

Tech teams at complex Web sites would likely find upon investigation that their Web sites suffer from problems that they either didn’t know about or didn’t know the extent of the damage they are causing.

One such problem is that certain static objects on the company’s Web pages that should be cached by the end users’ Web browsers are either not cached by the browsers at all or not cached enough. Some objects are at least cached by the CDN used by the company, but some perfectly cacheable objects are served all the way back form the origin servers for every request! An unnecessarily costly situation that can be avoided.

In addition to wasteful bandwidth charges resulting in high monthly bills, there are also other disadvantages caused by cacheable objects being unnecessarily served from origin servers:

  • They slow down your Web pages. Instead of the browser being able to use local copies of these objects, it has to fetch them all the way from your origin servers.
  • Unnecessary load on origin Web servers and network equipment at Web hosting facility. This can be an especially severe problem when a Web site experiences a sudden many-fold increase in traffic caused by a prominent incoming link on the home page of a high traffic like Yahoo, MSN or Google.
  • Additional storage in logs at the origin Web hosting locations’ servers and other devices.
  • Unneeded processing and work the origin servers, network equipment, CDN, the Internet in the middle all the way up to the client browsers have to do to transfer these objects from origin to the end user’s browser. Be environmentally friendly and avoid all this is costly waste.

The increase in bandwidth, load on servers and networking equipment and log file storage space increases caused by a few objects on Web pages being served by origin servers for every request may mistakenly seem like an insignificant problem, but little drops of water make the mighty oceans. Some calculations will show that for large Web sites, the cost of this can add up to tens of thousands of dollars a month in bandwidth costs alone.

How should companies operating large Web sites solve this problem?

For technology managers:

  • Make it a best practice to maximize the use of browser-side caching on your Web pages. Discuss this topic with the entire Web technology team. Awareness among the information workers is important so that they can keep this in mind for future work and also address what’s already in place. Show the engineers some sample calculations to illustrate how much money is wasted in avoidable bandwidth costs: that will prove this is not an insignificant issue.
  • If this problem is widespread in your Web site(s), make the initial cleanup a formal project. Analyze how much money you’d save and other problems you’d solve by fixing this and present it to the finance and business management. Once you show the cost savings, especially in this economy, this project will not be hard to justify.

For engineers:

  • Read the article about optimizing caching at Google Code for technical details on how to leverage browser and proxy caching. It explains the use of HTTP headers like Cache-Control, Expires, Last-Modified, and Etag.
  • Review any objects that are served by origin servers every time for legacy reasons that may now be obsolete.
  • Combine some JavaScript files commonly used by your Web pages so that the one unified and shared file would have higher caching probability. Do the same with external CSS style sheets.
  • Study a good book on Web site optimization like Even Faster Web Sites: Performance Best Practices for Web Developers. Share these recommendations and hold a discussion with your tech and production colleagues.
  • http://paradox1x.org/ Karl

    Hey – aren't you supposed to be on your honeymoon :)

    • http://intensedebate.com/people/rajivpant rajivpant

      Our honeymoon at Disney World was during the week before :-) We got back into NYC the weekend before this one. We plan to take another vacation later this year.

  • http://www.munisysllc.com Dinesh

    Rajiv, Great article. We are following the same in our projects and it is very cost effective. One drawback though, the projects that uses mainly AJAX / JQuery cannot use this technique (methodology).

  • http://www.teenbootcamps.com Kaira

    Nice article.

  • http://www.zoombits.fr/cadeaux-gadgets/ cadeau

    Thanks for sharing such nice information here. You described here very beautifully for different kinds of fields' people. My uncle has a big software company and this information is very useful for him.
    Thanks for this great post.

  • http://affilojetpacked.com affilojetpack

    interesting post…i never thought of bandwidth from a large company

  • http://www.squidoo.com/google-sniper-2-0 google sniper

    really great article

  • Pingback: Save Money On Hosting & CDN By Optimizing Your Architecture & Applications | Web Site of Rajiv Pant