Are Ad Servers Bogging Down the Web? 387
blackbearnh writes "The work of making high-volume web sites perform well is an ongoing challenge, and one that continues to evolve as the nature of web content changes. According to Google Performance Guru Steve Souders, fat JavaScript libraries and rich content are creating new problems for web site tuning, but one of the biggest problems lies outside the control of web site administrators — ad servers. In an interview previewing the upcoming Velocity Online conference run by O'Reilly, Souders talks at length about the real causes of poor web performance today, and in particular, the effect that poorly performing ad servers are creating. 'We adopted a framework of inserting ads, of creating ads, that's pretty simple. And because it's pretty simple, it's not highly tuned. That's one reason why we shouldn't be too surprised that we see performance issues in third party ads. The other reason is that ad services are not focused on technology. Certainly companies like Yahoo and Google and Microsoft, we're technology companies. We focus on technology. So it's not surprising that our web developers are on the leading edge of adopting these performance best practices. And it's also not surprising that ad services might lag two, three or four years behind where these web technology companies are.'"
Make it a statistic and they'll care (Score:5, Interesting)
Having worked for an ad-serving company, I'm pretty confident that the reason they don't care is that they're not measured on the speed at which they serve up ads.
If high-value websites started rejecting ad networks that served ads in less then x milliseconds after the rest of the page was downloaded, you'd see ad servers speed up, quick.
Re:Make it a statistic and they'll care (Score:5, Interesting)
Slow Ads by Contract? (Score:1, Interesting)
Even Slashdot is falling prey to slow ad servers. And to answer TheRaven64's presumption -- no, it used to be the case that ads loaded asynchronously, but today it seems that many, if not most sites delay loading the content you actually came to see until the ads load. I am guessing this is part of the contract between sites and advertisers. (Would any admin for Slashdot care to comment?)
-Coward
Re:Why? (Score:3, Interesting)
I realize that most websites run some version or another of "adverts", but generally speaking, most of those sites are marginal value to start. The sites I frequent usually use text ads, and not the flash (pun intended) graphical ads on some of the more questionable sites.
That actually reminded me of a short study I did in my English class a number of years ago. I wanted to know if you could get a quick feeling for the quality of a magazine based only on the number of advertisements/glossiness of the publication. Given the limited time and amount of money I was willing to spend I chose "Popular Science" and "Scientific American." PopSci had many more adverts than SciAm and, IMO, this means that SciAm is the better magazine. Yeah, it was a little subjective, but it was only a 5 page paper for an English class.
It's nice to see that the same thinking still applies to websites.
I'm looking at you, Slashdot (Score:3, Interesting)
I've mentioned the ad bottleneck before. Slashdot is an especially bad offender. Pages use several ad servers, and they use "document.write" to stall the page load until the ad comes up. Even if you have the ad images blocked, some of the junk JavaScript still needs to run.
Some sites are just slow at serving pages. Behind my SiteTruth [sitetruth.com] system there is a specialized web crawler which looks for a business name and address on each web site. It never looks at more than 20 pages, and it's looking for pages like "About", "Contact", and about 40 other words which might plausibly lead to contact info. This process runs about 5-15 seconds for a well-implemented site. I log sites where it takes more than 45 seconds. About 5-10% of sites run overtime. In the last hour, the slowest site is "www.airsmaxkey.com", at 159 seconds to read 10 pages. (Yes, they're a bottom-feeder. Not only is there no business address on the site (a criminal offense in the European Union), they have logos from Verisign, PayPay, Verified by Visa, and MasterCard SecureCode, none of which are actually clickable to do the claimed verification. Nor does their shopping cart checkout use SSL. The whole site may be a scam. SiteTruth gives them a "Do Not Enter" rating.)
Some of the social networking sites have so much Javascript that Firefox will time out. (Facebook had that problem for a while. They fixed it.)
Re:This isn't new (Score:3, Interesting)
That's why I use a targetted DNS black hole instead. I don't block ads until they cause a noticeable disruption in my browsing behavior. As soon as they add more than a second or so to a page load time, that particular ad server gets blocked permanently, and my caching DNS server returns a bogus response directed at 127.0.0.1 or a host not found, depending on the subdomain/host part. All of the various google-analytics domains are on my block list because they consistently fail to have adequate performance. Similarly, most of the larger ad networks are blocked for the same reason. The smaller ad networks, which usually have a more sane load average per server, are generally not blocked until they get too big for their breeches. This serves three purposes: reduces page load times, punishes ad servers that have slow performance, and promotes competition by encouraging the use of smaller ad networks.
Re:Flash Ads (Score:4, Interesting)
It is under your control (Score:4, Interesting)
Nonsense! I for one have chosen to keep my websites ad-free, hence no ad servers and no slowdown. The same goes for untold thousands of other webmasters.
If you've chosen differently then ... well, I suppose it's your website and your decision — but please don't come whining to us about the consequences.
Re:Kind of Fitting (Score:3, Interesting)
No, the main reason Flash is used is because it's "flashy" and draws more eyeballs to that space. Any additional tracking is just a side-benefit.
And everyone knows you can block ads if you really want to. Although Flash allows them to overlay ads over video and that kind of thing.
Re:Make it a statistic and they'll care (Score:5, Interesting)
JavaScript loading is usually blocking the rendering whereas img loading usually not.
NYT loads an astounding amount of JS. At home I use an especially slow dial-up and turn off image loading, so I was surprised to spend so much time waiting for "graphics8.nytimes.com". Then I looked in Firebug's Net panel. NYT home page launches 41 requests for 141K of data:
HTML: 5 requests, 31KB
CSS: 4 requests, 13KB
Flash: 2 requests, 37KB
JS: all the rest, 30 requests, 60KB
(Flashblock is allowing those 2 requests for some reason. I don't use AdBlockPlus.)
So next for me is to find or write an extension to block JS per-site.
Re:Kind of Fitting (Score:3, Interesting)
I got it earlier this year too, I'm under the impression it's some sort of high karma perk though.
Notify content-site owners (Score:2, Interesting)
When you hit a web site that loads slow because an ad-server or its DNS is slow to respond, report it to the content-owner web site.
They will be annoyed on multiple counts:
*Their advertisers aren't getting eyeballs they want
*Their own content is being devalued due to their site appearing "sluggish"
*They are getting complaints
By the way, a well-run ad network can give better performance than a poorly-run in-house network.
Old, old story (Score:5, Interesting)
In 1995, columnist and Ethernet-inventor Bob Metcalfe was again going on about a topic that eventually had him literally eating his words (he had to chop up a column in a blender with water and chug it) - that the Internet was going to collapse from all the heavy bandwidth demands of its exponentially-expanding clientele.
So I did a "View Source" on the Infoworld page with his column on it. I've lost the E-mail now, but the stats were something like his column being 2000 bytes and the sum of all the advertising around it, mostly GIF images at the time, was over 20,000 bytes. The Ad/Content ratio even then was over 10:1.
Metcalfe, who'd been railing against irresponsible bandwidth consumption in the column, could only plead that he had no control over the magazine's decisions on what went around it.
The web has always been the reverse of TV, where the ad/content bandwidth is about 1:4 or even 1:5. It's not far different from some magazines, though, where I swear there are 3 pages of ads for every page of content. And if you digitized the magazine, the ads would mostly be images, the content mostly text, and the ratio would be at least 10:1.
This is all prologue to new web content where you are slowed down not so much by download times as the start-up times for various Flash and JavaScript programs that make the ads so much more intrusive, zipping back and forth over the text you're trying to read, or just dancing in the corner of the page.
This is all necessary: they do what they MUST to get response from the ads. If the stats don't show a response, they stop buying them and the business model fails.
Everybody says "Nobody will pay for content on the Internet". Yes, they will. The put up with all that crap rather than pull out a credit card. They just pay with their time and attention instead of actual cash.
Rod Serling, one of the great TV writers of all time, once commented that it is hard to tell a story when you must work it around being interrupted every ten minutes by dancing rolls of toilet paper. I wonder what he'd think of writing for a medium where the toilet paper literally dances all over your words until you click on it to make it go back to the lower right frame.
Re:Make it a statistic and they'll care (Score:3, Interesting)
I can't count them either, because I can't count things I don't see. Call it what you want, but AdBlock works, and I think it upholds a tenet of free speech: Your right to say something doesn't burden me with an obligation to listen. I block the online ads, which in my view is no different from fast forwarding through them on a recording, or not paying attention to them during a live broadcast. As an added bonus, the ad servers save on bandwidth, making ads faster for those who wish to view them.
The only real question is whether blocking ads will decrease the value of ads and harm the earning potential of a given site, to which I can only say that if it does, then ads were overvalued to begin with. If someone is making a living (or supporting their endeavors) by selling an overpriced commodity, and the value crashes, well.. that's life.
Comment removed (Score:4, Interesting)
Re:Re flash cookie blocking (Score:3, Interesting)
Why not just go to Adobe's site and block them at the source [macromedia.com]?
That link doesn't personally appeal to me because it (and Flash cookies in general) is a default-allow policy. I greatly perfer a default-deny policy. I also dislike the idea that I would need Adobe's blessing in order to fully control the behavior of my own computer and the applications on it. That BetterPrivacy addon for Firefox is a better way to deal with this. So is making ~/.adobe/ a symlink to /dev/null, or deleting everything in it and then making it read-only (and root-owned if the plugin tries to modify the permissions). None of those depend on the goodwill of Adobe or its decision to have a default-allow policy for a proprietary and manifestly insecure plugin.
Only MOZILLA/FF stuff protected YOUR way: Try this (Score:4, Interesting)
"Top 3 addins for privacy: Better Privacy, AdBlock Plus, and NoScript, hands down imo." - by MollyB (162595) on Monday November 30, @01:37PM (#30272526)
Per my subject line above? How about a GLOBAL solution, instead, & one that extends to ALL of your "webbound apps", instead, AND acts as "layered security" in combination with the FF/Mozilla only methods you use (which slow your browser down, use CPU cycles & more... where this solution does not & covers ALL webbound apps, globally)??
Ok, well then - Here we go, & on that note, specifically:
Here is a GOOD SOLID & GLOBAL WORK-AROUND, CALLED A HOSTS FILE!
(It works for more speed online, AND SECURITY ESPECIALLY... Also, it works for your money, because you pay for your linetime out of pocket most likely as I do, you can get back your speed, AND, gain security easily, & from a single easily edited file & a file eats no CPU cycles like a local DNS server can (& are not as security vulnerable either if you protect write access to a HOSTS file also)... Anyhow/anyways - Here goes:
SO - "that all said & aside"? Well, per your reply??
Hey - NO PROBLEM, 110% agreement here on that account... & more (like more speed online AND more security, via a SINGLE EASILY EDITED + POPULATED FILE, called a HOSTS file that extends to EVERY WEBBOUND APP YOU HAVE):
I use a custom HOSTS file, in addition to the tools others here in this thread have noted (which MANY like FF addons only really function for FireFox/Mozilla products, but don't extend globally to all other webbound applications, & that is part of what HOSTS files give you above the methods you extoll + utilize: "GLOBAL COVERAGE", & of ALL webbound apps, not just FireFox/Mozilla ones via the addons you noted + use yourself...).
HOSTS files can be used to blockout KNOWN "bad" adserves, maliciously coded sites or adbanners, and "botnet C&C servers" too!
You can obtain reliable HOSTS files from reputable lists for more security online, but also for speed!
(More on that later & WHY/HOW (I use reliable lists for that, such as these HOSTS @ Wikipedia.com -> http://en.wikipedia.org/wiki/Hosts_file [wikipedia.org] or those from mvps.org (a good one this one))
I also further populate & keep current my custom HOSTS file with up to date information in regards to all of those threats, via:
----
A.) Spybot "Search & Destroy" updates (populates HOSTS and browser block lists)
B.) Sites like ZDNet's Mr. Dancho Danchev's blog -> http://ddanchev.blogspot.com/ [blogspot.com]
C.) Sites like FireEye -> http://blog.fireeye.com/ [fireeye.com]
D.) SRI -> http://mtc.sri.com/ [sri.com]
----
My HOSTS file incorporates ALL of the entries from the HOSTS files shown @ wikipedia as well... gaining me speed online (by blocking adbanners, which have been compromised many times the past few years now by malscripted exploits (examples below)).
(I combined ALL reputable HOSTS files with one of my own (30,000 entries), & I removed duplicates removed via a Borland Delphi app I wrote to do so called "APK HOSTS File Grinder 4.0++". That program also functions to change the default larger & SLOWER 127.0.0.1 blocking 'loopback adapter' IP address to either 0.0.0.0 (for VISTA/Windows Server 2008/Windows 7, smaller & thus faster than 127.0.0.1 default) or the smallest & fastest 0 "blocking 'IP ADDRESS'" (for Windows 2000/XP/Server 2003 which can STILL use it (& it was added in a service pack on Windows 2000, only on 12/09/2008 MS patch tuesday was it removed for VISTA onwards (& now all these "phunny little bugs" are showing up as FLAWS in this new NDIS6 approach via WFP as well in the firewall, which ROOTKIT.COM has stated (with code too no less on how it is done) ->
Re:Make it a statistic and they'll care (Score:2, Interesting)
Re:Make it a statistic and they'll care (Score:3, Interesting)
That happens when ad/analytics Javascript makes use of document.write to get its job done. What the industry needs is a move towards dynamically loading that content using XMLHttpRequest, so that the normal page content loads first and only when it's finished do the ads/analytics tags do their thing.
The only place this might be troublesome is if you're using the ad server to do geo-location. But even that isn't a big deal if it happens a split second after page load.
Re:Kind of Fitting (Score:4, Interesting)
From what I understand, it is. It's shown to those who have a high Karma, moderate, and meta-moderate. So, the good users. :)
I don't moderate or meta-moderate, and I don't think I have high karma, but I do get the check box for disabling ads. I think it might be related to the age of your account.
Style-sheet servers too!! (Score:3, Interesting)
I'm quite willing to custom-hosts blacklist CSS-servers, too.
FWIW.