Google's mod pagespeed Apache Module Comes Out Of Beta

[engine]

For developers and webmasters, making your page faster shouldn’t be a hassle, which is why we introduced mod_pagespeed in 2010. Since then the development team has been working to improve the functionality, quality and performance of this open-source Apache module that automatically optimizes web pages and their resources. Now, after almost two years and eighteen releases, we are announcing that we are taking off the Beta label.Google's mod_pagespeed Apache Module Comes Out Of Beta [googlewebmastercentral.blogspot.co.uk]

[tedster]

I'm very much in favor of using mod_pagespeed but so far I have not been able to convince others to make the leap. Maybe being out of beta will be the convincing factor.

There's a video at [googledevelopers.blogspot.com...] that shows side-by-side comparison of the same page loading with and without the module turned on. It's pretty convincing to me.

[mslina2002]

Two years ago, I asked my webhost when this first came out with beta whether they would add this. They didn't make any attempt to do anything about it. When this came out yesterday I asked again and only now they will do some testing. Looking around seems like others have already started implementing this (webhosts). It's probably time to switch if they don't get the ball rolling.

Was at a recent analytics conference and one of the major national brands (second spot on Fortune 500) said that for 1 second increase in speed, their conversion increases by 2%. So perhaps some actual data from others would convince others.

[scooterdude]

A lovely piece of Google software that does nothing other than advertised ?

hmm, this could be a different Google then,,


Goes to show, all you really have to do with the short of memory webmaster is wait till they;ve forgotten all they may have learned, then come bearing Gifts,

" this lovely wooden horse gonna make your lives better

[Roaming Gnome]

Goes to show, all you really have to do with the short of memory webmaster is wait till they;ve forgotten all they may have learned, then come bearing Gifts,

" this lovely wooden horse gonna make your lives better

Indeed.

I wouldn't doubt there will be stealth reporting built in.

[tedster]

So check any packets going out before you use it. Don't let paranoia poison you from using something that can be a benefit.

[Roaming Gnome]

Not that everyone is wearing tin foil hats, but FREE should always raise a flag.

The fact that one would even have to look for shady behavior is enough not to use it. Although always pushed under the rug, there are rumblings of other cool "tools" being used for grabbing user data.

[tedster]

Then how about Apache Web Server itself? Or Wordpress? Drupal? They're all free, right?

As I said, if in doubt check it out. Don't poison your own life with fear, uncertainty and doubt.

[scooterdude]

There is less uncertainty with Google nowadays, there is on the other hand lots of knowledge of previous events

We should have detector for the presense of this software on share servers for example, so webmasters can make an informed decision about whether they're happy to have their site on such a server

[levo]

Could be useful for generic hosting providers. As far as I can tell, some of the optimizations are 'dirty fixes.'

Personally, it is far more easy to optimize Apache (expire etc.) than install & maintain an extension. And I suspect all that processing (compressing/resizing), instead of doing yourself once, might have a very negative effect on server performance.

[TypicalSurfer]

The thing does chat with GA, actually it looks like you need GA to see results of any experiments. I'd need to see a "we don't track and never will" before exposing hosting clients to it, that could be a huge can of worms down the road even if you were unwittingly sending domain traffic data to google.

I guess google is responsible for any fear they put out there, it's not like they have a very clean record when it comes to data collection. I can't see them not wanting the data.

A good site build would probably accomplish the same thing. I guess if you preferred to be sloppy and rely on something like this you get what you get.

[edited by: TypicalSurfer at 8:47 pm (utc) on Oct 12, 2012]

[g1smd]

It apparently strips the meta copyright tag from pages sent out.

Kinda awkward when your page is copied otherwise verbatim elsewhere, or, ahem, 'cached' somewhere else.

[tedster]

it looks like you need GA to see results of any experiments

How so? As I understand it, what you need is actual page speed data taken from the visitors' browsers - and you can get that with boomerang.js

[xcoder]

Then how about Apache Web Server itself? Or Wordpress? Drupal? They're all free, right?

Yes they are all free but they do not appear to be 100% busy chocking free speech and competition on the web while clearly abusing their dominant powers. Google does. And this is more then a good reason for me to keep this mod well away OFF my servers.

I wouldn't touch this Apache mod with a 100-foot pole...

[mrguy]

Don't let paranoia poison you from using something that can be a benefit.

When it comes to Google, that's just no longer possible. If it doesn't benefit Google in some shape or form, they would not be giving it away.

[TypicalSurfer]

How so? As I understand it, what you need is actual page speed data taken from the visitors' browsers - and you can get that with boomerang.js

Just running with what Google Developers are showing:

This feature allows you to run experiments where segments of your traffic get the page rewritten with different settings in order to figure out which filters work best for your site. It reports to your Google Analytics account, storing data in a custom variable.

[developers.google.com...]

It's a little too integrated with Google stuff for me to consider it.

[tedster]

One of the features of mod)pagespeed that I don't know how to duplicate myself - but it looks extremely useful - is this functionality:

Recompressing an image when its HTML context changes to serve only the bytes required

[googledevelopers.blogspot.com...]
This sounds to me like a real boon. First, image size is one of the most common areas that I see slowing down page loads. For many websites, especially those with a frequent publication schedule, optimizing images creates a real snag in the workflow. In fact, many just abandon any effort at good image compression and let their authors upload very fat files.

But even more useful as I see it, is catering to the Apple Retina displays at the same time as you cater to conventional displays and even small screens where you still want a small image. It seems to me you could upload just one image file optimized for Retina screens, and let the server step it down for other users.

========

In terms of usage, I think many will find this section good to know, whether you want to avoid the module or embrace it:

We're working with Go Daddy to get mod_pagespeed running for many of its 8.5 million customers... We're also working with Cotendo to integrate the core engine of mod_pagespeed as part of their Content Delivery Network (CDN) service.

[incrediBILL]

It's about time they got this thing ready for prime time!

First, image size is one of the most common areas that I see slowing down page loads.

I agree about image size being a problem. However, the problem here is you have no real control over the image quality if you let some random tool apply some random compression. It's obviously better than nothing, but I prefer compressing my own images so I know what the visitor actually sees.

Furthermore, how to you know their compression doesn't ditch any watermarks or meta data in the file? Would have to test it before trusting this aspect of the product.

If it doesn't benefit Google in some shape or form, they would not be giving it away.

Of course it benefits Google. They crawl the web and it speeds up page delivery and reduces bandwidth usage, therefore making crawls faster as well as speeds up visitor page loads and Google page preview screenshots.

The whole thing is Open Source so anyone can see what it's doing so any paranoia is just silly at this point. I'd install it if I didn't think it might interfere with my control panel software. I'll probably just wait for them to include it in a future control panel build.

[phranque]

this module looks like it's trying to be the swiss army knife of apache modules.
it looks like it's ok if you are starting from scratch but i sure wouldn't try to shoehorn this into an existing site that was doing anything extraordinary.
for example if you have a wordpress cache plugin that works great for you, you probably won't be using that with this apache module.

this part is absolutely hilarious!
https://developers.google.com/speed/docs/mod_pagespeed/configuration#lower_case

you can sometimes achieve a modest improvement in the size of compressed HTML by enabling this feature with:

 ModPagespeedLowercaseHtmlNames on

(i never knew lower case used fewer bits.)

[levo]

First, image size is one of the most common areas that I see slowing down page loads.

For someone with technical knowledge, it should be easier to process/resize (even apply unsharp filter to make them look sharper, strip meta data, compress) with ImageMagick.

Sounds great for a generic service provider, but doing all that on-the-fly should not be the way to go for a serious developer.

[g1smd]

i never knew lower case used fewer bits.

With all lower case, the file might compress slightly more than one with mixed case.

[phranque]

shorter bits?

[brotherhood of LAN]

I can see why Google would want to offer this, as it'll help Googlebot in bandwidth & time.

It reminds me of GoogleGuy's thread Are you using If Modified Since? [webmasterworld.com]

[lucy24]

It apparently strips the meta copyright tag from pages sent out.

WHAT?

i never knew lower case used fewer bits.

Funny, last time I looked, lower-case letters came after upper-case letters in every script that has both.

Converting to lower-case HTML would save page speed IF browsers start by converting any upper-case HTML they meet into lower-case before they can start to parse it. (Possibly that's why the w3c validator demands lower-case for HTML 4 even though it isn't in the spec.) But that's hypothetical, and doesn't seem to be what they're talking about.

[incrediBILL]

It apparently strips the meta copyright tag from pages sent out.

More than likely it's the graphics library, not mod_pagespeed, that strips the meta data. However, that's a simple bug that can be fixed, should be fixed. It's also possible it's just a setting or parameter that needs to be used when calling the graphics library to maintain meta data.

Either way, both are Open Source and anyone here with the skills should feel free to rectify what is most likely a trivial problem. If this bug were fixed once and for all, especially if it's in the graphics library, it would solve the problem for everything that also uses the library.

If I had the time I'd offer to look into the situation, probably no more than a couple of hours from start to finish.

[ergophobe]

>> shorter bits?

No, longer bits ;-)

Or rather, longer chunks. The more repetition, the more compression. Try creating a two text files of all ASCII (so all capable of being represented with seven bits), but one is random and one is just one letter repeated. Huge difference obviously. But now take the large random file and convert it all to lower case. You have a much great chance of a repeated characters, so better compression.

>> i sure wouldn't try to shoehorn this into an existing site

That's what I've heard from early testers. If you're running Wordpress, you'll get better results with WP Super Cache. With mod_pagespeed I've heard there can be some issues with inline Javascript and things like that.

It's better than nothing, but it's not as good as a solution that's customized for the site itself (I suppose that's obvious). So I've heard, if you can't do something else, do mod_pagespeed, then monitor your site very closely, because there are incompatibilities.

[phranque]

i get it.
it's not "lower case" per se - but rather "same case".

[ergophobe]

Yes, plus lower case letters are smaller ;-)

[phranque]

i knew that! =8)

[brotherhood of LAN]

What if my lowercase text are in a H1 tag, are they still small ? ;o)