What is Digg?Sponsored by Sony Pictures
Adam Lambert sings the 2012 theme song, “Time for Miracles” view!
whowillsurvive2012.com - Watch the Adam Lambert music video for the 2012 theme song. See 2012, in theaters Nov 13
98 Comments
- inactive, on 10/26/2007, -8/+78What else are you gonna call it? It first started off being called the "Slashdot Effect", now its picked up other names like Digg Effect, or Reddited. The common thread: web 2.0 and aggregation sites. So, why not call it the web 2.0 effect?
- AlphaPrime, on 10/11/2007, -4/+53Isn't that called the digg effect?
BTW, how hard is it to display some random text with a colored background. If they really wanted to test it they would be serving up video or something along those lines. Not just text. - inactive, on 10/11/2007, -2/+44u gotta add a better title though!! "20 tb of free streaming porn" or "man gets run over by google street view van PICS"
- KevinJB, on 10/11/2007, -5/+45A few problems with this experiment:
1 Each page load should connect to a database and perform a few queries.
2 I'd like to see a few more webhosts on there, notably Dreamhost (I've personally surivived Digg without noticing ANYTHING with them, YMMV), Site5, etc.
3 You're really getting the combined effects of Reddit, Digg, and anywhere else this gets picked up.
All in all a brilliant idea though. - rubberpants, on 10/11/2007, -1/+21FYI, the links at the bottom of the page all point to the first hosting company in the list.
- HigherLogic, on 10/25/2007, -1/+19This is the lamest ***** ever: "Hosting companies lie. They convince you that as long as you stay below your bandwidth and storage quota, you're site will stay up."
Hosting companies don't lie, hosting clients are just ignorant. You are required, as part of your agreement when you sign up, to read the TOS and AUP, which always include a RAP (Resource Abuse Policy). And this is true for virtually everything, people DO NOT read the fine print (but hey, every company also knows this too). When they're cut off for violating it (e.g. Google AdSense), they're the first ones to bitch about it.
This is the biggest misconception with most people when it comes to shared hosting, "Oh, they give me 500 GB of disk space and 5 TB of bandwidth, I think I'll run an ecommerce store now!" They never look at anything else. They know that there's no way in hell you could use 500 GB of disk space (legally). The majority of people on a server (we're talking like 99.9% here) use less than 1% of their disk space and bandwidth. The other percentage are host-hoppers. They'll switch till they're blue in the face.
They also know that you can't use all that disk space (and subsequently the bandwidth) because you'll hit the glass ceiling, so to speak, and use more resources than allowed on a (dun dun dun) SHARED hosting environment. That's just common sense. Use 100% of the resources, well guess what, you're not sharing! Get a dedicated if you want to use 100% of a server's resources.
Finally, they know that anyone who actually runs a website that should be on a dedicated server already has a dedicated server. Those numbers they offer, disk space and bandwidth, are for the ignorant. It's purely marketing. They know that 99% of the people out there don't care about the hardware of the server they are on, the load average, how many people are typically on a server, what kind of backup recovery is in place, etc. They also know you won't read the TOS/AUP, hardly anyone does.
Tip: when researching hosts, look at the TOS. Ignore the disk space and bandwidth. It's 100% irrelevant. If you really have to worry about how much disk space they offer, you should think about dedicated, not shared.
I've worked with some pretty big websites before, and this one client of mine had 6-7 websites running on a dedicated server, they were all sister sites part of a network/community of sites. Total uniques per month was in the 500k range (1M+ pageviews). There were 3 vB forums, a huge image gallery, and about 6 other databases for various sections of the site. With all that combined, they only had about 30-40 GB of disk space being used (the majority of which were images for the gallery). Could you run this on a shared environment? Hell ***** no. You would be stupid to think so.
People overestimate by a longshot how much disk space and bandwidth they actually need, but hey, more is better right? Nah, I don't want the V6, I want the V12. Don't give me 12 ounces, gimme 40 instead. I won't ever use that much HP, I won't even finish all 40 ounces, but just give it to me.
There's three things people look at: price, disk space, and bandwidth. You can make up the numbers as high or as low as you want for all three, it's just a matter of coming up with a reasonable number. Preferably a low price and high DS and BW (the higher the better).
And seriously, if you're website is a crucial part of your business, then shared hosting is obviously not the route you need to take. If you're spending the equivalent of a 6-pack of beer a month on hosting, you're not really serious about your website.
If you write articles that are frequently hitting Digg, spending $5, $10, hell even $20 a month is just plain stupid. Either go halfway and spend $50-70 for a semi-dedicated solution, or go all out and spend $110+ for a dedicated environment. This is especially true if you have an online business. The last thing you need to be on is a shared environment. People THINK their business is important, and they're the first ones to submit a ticket and raise hell because their site is down, "This is my business! I'm losing money every hour the site is down!" Yeah, sweet ***** business.
Anyways, there's a lot of information to be desired here. What are the technical specs of the server these sites are on? What's the average load? How many users are on the server? So many things that need to be factored in here.
And serving up just straight text is ridiculous. If your host can't even handle traffic to a static file, you have more problems to worry about. Let's see some dynamic content, some MySQL queries, some actual test scenarios.
I also think it's funny that IX has a high rating, but the majority of the reviews for IX are about downtime. And another gripe, why does HostGator get to have their reseller plan listed? I thought these were entry-level plans, that would mean we should be looking at their Hatchling plan.
Sorry for the long post, been a long day. - iamshades, on 10/11/2007, -3/+20@alpha
Random text? That's an insult to Lorem Ipsum.. - essjay, on 10/11/2007, -0/+14I'm completely with you. Out of the nine hosts, 8 are up 100% of the time with one 85%. Anyone who visits digg regularly knows that sites that use this sort of hosting stay up for percentages anywhere near that amount. The real test, especially for the content digg gets, would be to throw a wordpress installation on them and then run the test. Plain html is a completely pointless test and holds very little merit.
- sedition, on 10/11/2007, -1/+14You forgot the best one. Penny Arcade linked sites are "Wanged"
- Rhino2, on 10/11/2007, -0/+12Yea, the web pages are kind of small. I'd throw some images up, may some simple DB queries, some Javascript, banner ads and other crap... stuff you would find on a "typical" web page.
- Abennobashi, on 10/11/2007, -11/+23I prefer the term "Digg-*****"
- k4st, on 10/11/2007, -1/+13*great* domain name.
- gforce42, on 10/11/2007, -1/+11"Rankings on Last 25 Visits"
- schnitzi, on 10/11/2007, -1/+10It has a name: the Slashdot effect. That's what it was first called. Why not keep the historical name? Are we to change it every time a new popular website starts generating a lot of hits for a site?
- thephotoman, on 10/11/2007, -0/+6You know, "using too many resources" usually refers not to bandwidth or disk space, but processor capacity and memory. Read the terms of service for your host: most shared websites will not allow you more than a certain percentage of your server's total capacity.
Where I work, we'll try to just kill the script or database that's causing the server load issue, but if we can't isolate it to one script, the entire account will be suspended for abuse. If you manage to keep your site simple, though, bear in mind that a Commodore 64 once survived a Slashdotting. - opusaz, on 10/11/2007, -0/+6Don't forget farked. They're sensitive about that. :-)
- carguy84, on 10/11/2007, -2/+8Digg effect troubles have very little to do with hosting and more to do with coding. I'm sure Wordpress sites have plenty of server horsepower to handle text based blogging, but the code behind the scenes is what always gets them into trouble. lighthttpd can handle hundreds of html requests per second, but if you throw crappy programming in front of it, bye bye website.
- GrendelT, on 10/11/2007, -2/+7Don't forget everyone's favorite shoddy webhost, dreamhost.com !!!
- masterc, on 10/11/2007, -0/+5Why do people always post the exact error message they get? That's kinda stupid...
- GrendelT, on 10/11/2007, -2/+7I still prefer to call it Slashdot'd, since they were the first to really create problems of mass visitors.
To me, if a site gets "dugg" that means I cast a vote, or someone did. If it gets swamped and goes offline, it's slashdotted. Not farked, not reddited, slashdotted. - Tabou, on 10/11/2007, -0/+5Who's hosting the main page?
Could you include the total number of hits on that page as well? - spudnic, on 10/11/2007, -0/+4Farked is easily my favorite
- ayeroxor, on 10/11/2007, -1/+5kill it with fire!
- bgolat, on 10/11/2007, -0/+4Make it worthwhile and add a big image
- DiggzDE, on 10/11/2007, -1/+4I laughed at that right there. Great "experiment"
- senixon, on 10/11/2007, -0/+3Ok this is a great idea to test the hosting out there but C'mon 3.5kb of data, this isn't a real life example of a webpage.... add a few images and bump it up to at least 30-50kb and do it while this story is still on the front page!!!!!!!!!!!!!!!! Also 25 minutes is not enough time span ... do last 24 hours!!!!!!
- user777, on 10/11/2007, -0/+3if u're going to test it again, preferably with wp installation, php or some more code intensive program to push everyone to the limit.
Please put BLUEHOST on there also. - iamnos, on 10/11/2007, -1/+4Agreed, its very simple for a web server to hand out static html pages, with no images. Load a couple average size images (maybe 20K-50K) and has KevinJB suggested, run a couple basic queries, they don't even have to be against a actual table, just format the current date or something. This will also help to make sure the web server isn't doing any caching.
The fact that a couple of hosts are having problems with just these static pages does say something though. - ayeroxor, on 10/11/2007, -0/+2Actually, I created these hosts to see how many stupid little marketing tools would use the term "web 2.0" when testing their bandwidth.
You fail it. - HigherLogic, on 10/11/2007, -0/+2Why the hell would you need to include GoDaddy? I mean my ***** god, they might as well add Yahoo! to the list. There's probably a reason why DreamHost isn't on the list, it's a given that it's going to be down. This from a company where load averages of 700+ have been reported.
The author should also post the data center for each of these places, my guess is 7/10 of them are using the same data center. - ElfWord, on 10/11/2007, -2/+4Please add Dreamhost!
- KevinJB, on 10/11/2007, -0/+24 That's not very close at all to the size of the average webpage, it would be nice to at least triple the amount of text on there (probably about five more times)
5 Stuff like images and css adds up to a lot more requests to the server per page, but when you only have text it only takes one. Adding some tiny images (5 - 10) would also help make it more realistic.
While typing these last two up I realized how dumb both of those (and most of the above) are. Just set up a WordPress blog, then copy it and the database on the other eight servers.
Edit - whoops, sorry, didn't see iamnos' comment, which basically says the above two things. The admin of the site said he was dissapointed he couldn't get a Dreamhost account on Reddit... why not? - iamlost456, on 10/11/2007, -0/+2ahh, I didnt see that. I feel really smart now.
- Kostik, on 10/11/2007, -0/+2Can someone please do the same with dedicated and virtual servers?
- garbs, on 10/11/2007, -0/+2Midphase acquired AN Hosting a while ago so they are essentially the same company.
- Diggtatorship, on 10/11/2007, -0/+2I agree with essjay, you need to throw a wordpress installation on each server. I'd also like to see Verio, Dreamhost, and Bluehost added to the lineup.
Another really important piece of information here is the cost. Unless you're getting all these accounts for the same price there's no use comparing them. - inactive, on 10/11/2007, -0/+2all different iframes though
- FPSDavid, on 10/11/2007, -0/+2AN Hosting 1337 Timeouts
- ascott9, on 10/11/2007, -0/+2@russvirante
Are you going to post up a write up after this phase about which host did the best overall, etc? - Jyaif, on 10/11/2007, -0/+1a bit late, but maybe that could help: http://deskowner.googlepages.com/desk3.html
(it's a web page that reloads the page) - kyl3Full3r, on 10/11/2007, -0/+1cron 1 * * * * wget http://web2.0effect.org/
Now every minute your computer will download the html page calling all of the iframes?
Or will wget not view the other websites using the iframes, just the main page? - inactive, on 10/11/2007, -1/+2@higherlogic
I ran the experiment and I agree with several of your points, although many of them were deliberate...
1. I really wanted to test whether or not the sites throttled bandwidth, not specifically up and down time, which could be caused by quite a few things. I can change a lot of things about my site, including the mysql or postgreSQL set up, caching, etc. but I can't change the host's internal policies of when you are using up too much bandwidth. That is why I focussed on this directly.
2. I realize that HostGator Reseller is not quite accurate. To be straightforward, I ran out of money that I could budget out but had a friend with a HostGator Reseller account who let me set up on one of them. I did make sure to say that it was HostGator Reseller because I wanted to be honest, but it, admittedly, is not as good as if I were able to purchase another hosting account.
3. In my honest opinion, Hosting Companies do lie. Pretty much every service industry has some sort of exclusion in their Terms of Service that lets you know that at any time the company can refuse service. This is a common, understandable and unavoidable addition to protect against malicious users. However, I have yet to find an industry that so readily and regularly utilize those policies to temporarily refuse service to an existing client as is the case in the hosting business. Refusing service (ie: "temporarily disabled due to..." errors) has become a part of the business plan, yet another way for them to stuff more clients onto a server without revealing directly, up front, the exact throttle.
Shared hosts do have it tough, though. It is competitive and people are always asking for more and more and more. The reason that people ask for more, in my opinion, is that they become confused into believing they need more bandwidth/space when their site goes down on a shared host when, in reality, they just need the same or less bandwidth and a host that doesn't strip their service when traffic starts pouring in.
If shared hosts openly described bandwidth, storage, and bandwidth rate (1 gig/ hour max for example), we would be a lot better prepared as consumers to find the right host. - kyl3Full3r, on 10/11/2007, -1/+2You can't add a new host half way through, dreamhost wouldn't get as many views making this test unfair.
- kyl3Full3r, on 10/11/2007, -0/+1I did a whois and the domain is using sitelutions.com name servers. Looks like http://sitelutions.com/ is hosting the main site.
- manfmnantucket, on 10/11/2007, -0/+1yes I'd love to see a comparison with MediaTemple grid vs DV
Also - what's with this Burton company? I don't recall seeing it on the list when I looked at the
earlier version of this test page. The page was obviously edited during the test.
I just tried going to burtonhosts and the hosting site itself doesnt even respond - it times out! - HigherLogic, on 10/11/2007, -0/+1>> 3. In my honest opinion, Hosting Companies do lie.
Well I guess you were right about that after all...
http://www.devindra.org/tech/2007/07/06/the-web20effectorg-hoax-how-one-marketing-firm-fooled-the-web - moltar2, on 10/11/2007, -0/+1Add slhost.com to the list as well. Check the WHT forums for their benchmarks on VPS -- they are significantly better than any other VPS.
- sygyzy, on 10/11/2007, -0/+1The more I read the comments, the more I realized how flawed (and lame) this experiment was. Please try again because it's actually quite an interesting idea.
- trouble916, on 10/11/2007, -0/+1Agreed... many of us got in on Dreamhost's $7.95/yr deals or whatever coupons are being offered. They're actually one of the few hosts to explain overselling and be a little bit honest about it.
I thought they were being kind of reasonable about it, limiting it only to high cpu users that were impacting perfromance of other customers... but lately it seems like there's NO possible way for someone to use up 100% of their allocated (paid for) bandwidth.
I agree with the guy running this story, the hosters aren't being honest about what they're offering. I'd rather plan my purchases and usage against definitive numbers of what is and is not allowed. I would like to see a bigger-scale test next time including Dreamhost and some sort of database query.
Nice job though! - trouble916, on 10/11/2007, -0/+1Could you please add the prices you're paying for each, and what they claim to be offering (99.999999999% uptime, 4billion gig transfer, etc. etc.) to the chart. That would be really useful!
-
Show 51 - 98 of 98 discussions
Browsing Digg on your phone just got easier with our enhancements to the 
© Digg Inc. 2009
— Content created and posted by Digg users is