Friday, July 29, 2011

Page Speed Service: Web performance, delivered.


By Ram Ramani, Engineering Manager

Update 7/29/11: We were notified of a bug in the measurement tool that sometimes causes incorrect measurements. If your results indicated a slowdown on your pages, please run the tests again, and make sure you specify a fully qualified domain such as www.example.com. We apologize for any inconvenience and confusion this may have caused.

Details:
Measurement tests run for bare domains (such as example.com, without the prefix www) previously indicated that pages were loading more slowly, rather than speeding up, when using Page Speed Service. The test results page now prominently notifies you of this when you visit this page, if this error applies to you. Please check your old measurement results page if this bug applies to you. Running the tests again with the fully qualified domain such as www.example.com usually fixes the issue and gives you the correct measurement.


Two years ago we released the Page Speed browser extension and earlier this year the Page Speed Online API to provide developers with specific suggestions to make their web pages faster. Last year we released mod_pagespeed, an Apache module, to automatically rewrite web pages. To further simplify the life of webmasters and to avoid the hassles of installation, today we are releasing the latest addition to the Page Speed family: Page Speed Service.

Page Speed Service is an online service that automatically speeds up loading of your web pages. To use the service, you need to sign up and point your site’s DNS entry to Google. Page Speed Service fetches content from your servers, rewrites your pages by applying web performance best practices, and serves them to end users via Google's servers across the globe. Your users will continue to access your site just as they did before, only with faster load times. Now you don’t have to worry about concatenating CSS, compressing images, caching, gzipping resources or other web performance best practices.

In our testing we have seen speed improvements of 25% to 60% on several sites. But we know you care most about the numbers for your site, so check out how much Page Speed Service can speed up your site. If you’re encouraged by the results, please sign up. If not, be sure to check back later. We are diligently working on adding more improvements to the service.

At this time, Page Speed Service is being offered to a limited set of webmasters free of charge. Pricing will be competitive and details will be made available later. You can request access to the service by filling out this web form.

Ram Ramani is an Engineering Manager on the Make the Web Faster Team in Bangalore, India. He is a believer in "Faster is better".

Posted by Scott Knaster, Editor

38 comments:

  1. How does this work for dynamic and personalised sites? Are you caching anything at all?

    Or is this more a preferred solution to static sites?

    ReplyDelete
  2. Doesn't that mean Google will be able to view in minute detail almost every interaction the user has with the site, including user's private data passed to and from the site? Of course not for encrypted connections, but even then, isn't it a rather perfect place for a man-in-the-middle attack?

    Oh yeah, I forgot, Google doesn't do evil, so it's OK.

    ReplyDelete
  3. optimized version was slower http://www.webpagetest.org/result/110728_2Q_3e7580f57d0bf156d364b2771c3efe94/

    ReplyDelete
  4. This is going to have a huge impact on the image vertical. How will Google ensure that we are not at loss as a result?

    ReplyDelete
  5. Of course the optimization is going to be faster if you don't load the images: http://www.webpagetest.org/result/110728_YY_ee2dd5b5ad27a678cdd1563efbe07d8e/

    However, I have images on my site for a reason: they're pretty.

    ReplyDelete
  6. optimized version was slower for me too

    http://www.webpagetest.org/result/110728_50_bbd13ed05b3cbdccac75fba7404888cc/

    ReplyDelete
  7. Great to see you guys stepping into another market, don't squish your competition too hard though - many of them are just getting off the ground.

    ReplyDelete
  8. Looking forward to trying this!

    ReplyDelete
  9. As unlikely as it is, what happens if Googles Server goes down, then you website doesn't load at all?

    Some people don't like losing control of simple stuff like that, I myself would, but my employer wouldn't.

    ReplyDelete
  10. Very interesting service. Look forward to use it

    ReplyDelete
  11. Don't be discouraged if your "optimized" site appears slower. We're under HUGE load today and our caches aren't primed. Try back again in a day or two - it'll look much better!

    ReplyDelete
  12. @Ram : in spite of test being slower, ill be going live for my blog after a while ... where is the cache located? Is it dispersed to all countries where google has servers? or its a centralized cache?

    ReplyDelete
  13. Also, will google sites like blogger automatically use this? That would be great.

    ReplyDelete
  14. I would really like to know the technique behind this - especially with focus on the questions that "velocity boy" asked:

    - How does this work for dynamic and personalised sites? Are you caching anything at all?
    - Or is this more a preferred solution to static sites?

    Also, if Google Servers cache the content, how often are the caches refreshed, and will login etc. even work?

    ReplyDelete
  15. @sajal - the caches are distributed globally.
    @Carlos Martins - Blogger & Sites don't use it today. I can't comment on future plans.
    @velocity boy, Martin DJ - it should work for both dynamic and static sites, but the gains depends on the site itself. Not all sites will show gains, so its best to run measurement tests & find out. As the blogpost says, we are working on more enhancements.

    ReplyDelete
  16. I tested about a dozen of my sites, in every instance the "optimized" version was significantly slower..

    I do have to add that I have spent a lot of time optimizing these sites already, they run on dedicated servers that are optimized as well and most serve a significant percentage of their "stuff" using a CDN so the benefits of this program might not be geared at sites such as mine.

    ReplyDelete
  17. Google announces Page Speed Service which is kind of like the start up CloudFlare. The thing is, how do you know that Google isn't going to shut down their project in 1 year if they find it doesn't receive much traction (or any number of reasons)? In the past 6 months they have shut down a number of services as well as many public APIs and Google Labs.

    In this case, it makes more sense to trust your services to a start up who is at least going to _try_ to stay around and at least _try_ to give you customer support. To tell you the truth, I think Google has lost a lot of consumer confidence in their services.

    ReplyDelete
  18. @Julian: Well, if that ever happens, you'll get an early notification like the other shut down services. You can then move back to CloudFlare or other CDN services you like, I guess

    ReplyDelete
  19. Can we get a sneak peek idea of how much this will cost? I'd hate to fall in love with this thing just to later find out I can't afford it. :-\

    ReplyDelete
  20. @Leigh Harrison - please put in your request in the signup web form. We will be approving sites gradually.

    ReplyDelete
  21. Very interesting service, but how do we (or more importantly my customers) refresh stored content when content changes (all my sites are content managed) especially for dynamic sites like blogs with comments updating often. Or is this service only good for static sites, or sites with infrequent content change?
    Thanks

    ReplyDelete
  22. Creo que no es una buena idea, esto va en contra de la neutralidad de la Red, no se si se puede confiar luego que estas web no esten mejor posicionadas, y que se les de ventajas sobre otras web. no me agrada esta idea, es mas me parece lamentable. Supongo que si no es google lo haria otro.. pero asi y todo lo considero negativo.

    I think it is a good idea, this is against net neutrality, if you can not trust these websites after not better positioned, and to give them advantages over other websites. I do not like this idea, but I find it regrettable. I guess if I would do is google the other .. But even so I consider it negative.

    ReplyDelete
  23. Quite the improvement for JSM Protective's site. We've signed up.

    ReplyDelete
  24. Does the rewriting of the code also make it pretty for search engine spiders? :) speed is good.. but if it also makes code more friendly then that is a double +. Would love to try it.

    ReplyDelete
  25. Your innovation is superb.

    ReplyDelete
  26. Tested my LiveJournal. -7% to page load time, -63% to repeated view. Nice :)

    ReplyDelete
  27. Sign Up page says We're sorry.
    You can't access this form because it is in violation of our Terms of service. If you feel this is in error, please contact us.

    ReplyDelete
  28. @Ram Ramani - Because the content would be distributed across Google's CDN, presumably, does this provide a valid solution against a DDoS attack? We're currently paying a certain well-known host $3.5k per month for a lot of metal, including a dedicated Preventier filter, in the off chance we get attacked again. Could you elaborate what would happen if a site using the new Page Speed service is DDoSed?

    Also, I'm with @Mohamed Fasil - your signup form is throwing a Violation of Terms of Service error.

    ReplyDelete
  29. @Mohamed Fasil - the sign up page should be back up again. Sorry for the inconvenience.

    @sstringer - we currently don't promise DDoS protection for individual sites, although this is turning out to be a popular feature request. We can handle spikes in your traffic, up to reasonable limits - such as due to a promotion for your site.

    ReplyDelete
  30. What do you do for JPEGs? Anything like this - www.jpegmini.com?

    ReplyDelete
  31. i want to know more about it who can help?

    ReplyDelete
  32. Look at this

    http://www.webpagetest.org/result/110818_PQ_dbbcc5d6d6d5d5cea8a7d91d89dca80e/

    ca. 25% slower and not faster.

    ReplyDelete
  33. Does the rewriting of the code also make it pretty for search engine spiders? :) speed is good.. but if it also makes code more friendly then that is a double +. Would love to try it.

    ReplyDelete