BBC News Under The Bonnet
diodesign writes "BBC News has revealed that Linux and Apache power its popular news website, along with a modified DNS server and machine farms in New York and London. At peak times, the site serves over 4 million users and 50 million page impressions a day. It's a pretty well explained guide to producing a regularly updated content based website that scales well." From the article: "The technology which serves the site is designed to be as simple as possible. The simpler the site, the cheaper it is to run. There are fewer elements which can malfunction on big days; and there are fewer parts which can be compromised by someone trying to gain unauthorised access."
none of those machines were bought with linux pre installed on them so it doesn't really count towards anyone's market share!
Now perhaps the CBC will learn from this.
Did anyone else read that as "BBC under the botnet" ?
According to Netcraft, they're the 9th most popular site on the web. That's after several variations of Google, and toolbar.netcraft.com... so take with heaps of NaCl.
To /.'s daily load?
antipaucity
Getting late. I thought this was a followup to:t id=172&tid=220&tid=95
http://it.slashdot.org/it/05/06/03/1516247.shtml?
I leave them on. I want to see the sites that insist on putting up pop-up ads to whoever comes by. So I will remember never to use them. BBC has never done this to me, and I use it for news even though I am a USian. Go BBC.
Transcend Humanity. Please.
Bonnet is to the British what Hood is to Americans.
See? It makes sense now.
They also have a better name for speed bumps - "sleeping policemen".
Yeah, that looks pretty good, but just in case... here's a mirror
not sure... but they've been around a while.
about them
Ah, if I could only watch BBC news at my local bar without someone asking why I hate America!
The revolution will NOT be televised.
Aww.. I was more interested knowing how they've managed the traffic within database systems. Or caching capabilities.
Why don't you post that on hotornot.com .. so we can give you a little reality treatment!
kg/am
maybe it's just me, but i'm never putting physical addresses on ANY network map with any company i work for, especially maps that will be posted publicly.
...British rappers are from da Bonnet?
Weaselmancer
rediculous.
Do we really need it revealed? It's not hard to figure out, just send a bad HTTP request:
bbc.co.uk
HTTP/1.1 400 Bad Request
Date: Fri, 03 Jun 2005 19:07:25 GMT
Server: Apache/2.0.51 (Unix)
"The crows seemed to be calling his name, thought Caw."
Do we know what distro the Linux servers run? Just interested...
Guy asked me for a quarter for a cup of coffee. So I bit him.
Post it on Face The Jury... they're ratings are a little more realistic.
It takes more than showing some skin to get a 10ish rating.
Hotornot you get a 10 for anything whore-ish.
Note: I'm at work so I didn't check the picture.
Apache is the most common web server around. But Apache on its own does not deliver content. Apache + Linus is not news any more. Apache and web servers in general are commodities today. On top of Apache a content management system runs for sure. It would be more intresting to read how this system works, if it is proprietary or free, etc.
vaginal bloodfart!
I wonder if they'll ever shake their heads and say - how did we ever think we could put all our eggs in those two high-profile terrorist target cities. I'll set up a mirror for them on my ADSL line.
Get your own free personal location tracker
Banu
I recall an article about CNN's website handling the load of traffic on 9/11 but have never been able to find it again. I think it talked about how their systems automatically reduced the complexity of page content in direct response to the number requests coming in that morning. Anyone else remember seeing an article like this?
The simpler the site, the cheaper it is to run. There are fewer elements which can malfunction on big days; and there are fewer parts which can be compromised by someone trying to gain unauthorised access."
They obviously subscribe to the "Keep It Simple Guv'ner" methodology at the BBC.
Thanks to this article, now I can make a content based website, too! Maybe now people will come to my site. All I had before were blank pages.
- None can love freedom heartily, but good men; the rest love not freedom, but license. -- John Milton
You can't compare the BBC to Slashdot. While the BBC website sometimes serves 50 million page views per day, Slashdot is unlikely to serve that many in a month!
I know that you think Slashdot is the ultimate in websites but, the rest of the world does not share your view. In a ranking of the busiest sites on the internet, Slashdot is VERY low on the list. Sites like the BBC, Google, MSN, Associated Press, eBay and a hundred others, all blow Slashdot away.
Slashdot's "awesome power" to crush webservers isn't very impressive when one realizes that the web servers are on DSL connections. Whereas, places like the BBC are on multiple gigabit pipes.
People who try to label anything outside of the US with the US version of the word "liberal" need to be smacked down with a cricket bat.
Reality is liberally biased.
Remember to US citizens:
liberal=socialist=communist
In the article it explains how they use SSI (Server Side Includes) to put their sidebar items into each story. Seems like a bit of an antiquated method for these days, no?
I'm at work so I didn't check the picture
Good thing. I've seen less grease in a deep fryer.
I really have nothing to say. I'll go now.
Ciao
More to the point.... better tell IBM that the simpler the solution, the cheaper it is to run.
That should put a kink in their global services based business model...
Simple, elegant designs = less $$$ required for consulting services.
Safe for Work, just NSFL (Not Safe for your Lunch). Actually, nothing horrific, just a very unflattering picture of a not too pretty person.
the preceding comment is my own and in no way reflects the opinion of the Joint Chiefs of Staff
Thanks for wasting my time again, Taco.
The Beeb are heavy users of Perl, too. Back when I decided to make the move from Perl programming to info security (around the time of Code Red and Nimda, also around the time our supposedly solid & profitable employer went tits-up) several of my ex-colleagues ended up hacking Perl at the Beeb. Apparently that chunk of their IT was just out-sourced to Siemens (German conglomerate) who are hardly ever referred to as "semens". Apparently.
"None are more hopelessly enslaved than those who falsely believe they are free." -- Goethe
How the heck did the netcraft toolbar appear so high up the list? That's a nerd site!
My dear fellow, what is "chippin"?
Reality is defined by the maddest person in the room
Jes' wondering if they use IRRd hooked into their DNS to provide resolution based on best service to incoming IP.. And if IRRd or something like it is part of an opensource DNS daemon..
Here's the blurb from the NTK link above:
"None are more hopelessly enslaved than those who falsely believe they are free." -- Goethe
"BBC News has revealed that Linux and Apache power its popular news website ... "
And Solaris.
The servers themselves are running Apache web server software on either the Linux or Solaris operating system.
'The servers themselves are running Apache web server software on either the Linux or Solaris operating system'
http://milkshake.dexy.org
maybe it's just me, but i'm never putting physical addresses on ANY network map with any company i work for, especially maps that will be posted publicly.
What makes you think that map's complete? There's a lot more resiliency in the system than it would suggest, and there are extensive disaster recovery plans, alluded to in that article.
(An SBS employee...)
Does anyone have any links for their bandwidth usage?
I have always used the text site as it loads almost instantly - any interesting story that requires pictures I then head over to the 'graphic' site.
BBC text news
The next step is to get them to report the news unbiasedly (during the last Iraq war, BBC was known here in the UK as the 'Baghdad broadcasting Corp.'); and we all know what their technical expertise is like explaining computer issues.
IBM GS'er here. Unfortunately for our numerous critics who don't have our experience, designing and building a web site like the BBC is childs play compared to the kinds of things we do. Have you ever built an airline reservation system with LAMP? How about something simpler like the software needed to manage a 3000 bed hospital? Would it scale to handle 1,000 hospitals just like it? Ever design and implement a mixed media storage and retreival system for multiple petabytes of data? We have. The stuff we do is complex because the problems we solve are complex.
I noticed on the diagram that the location of the London and New York server farms are called "Telehouse". Is this a fancy British word for a telecommunication building or am I just a stupid American? ;)
[(me)@localhost security]$ wget -S http://news.bbc.co.uk/2/hi/uk_news/magazine/460671 9.stm1 9.stm0 c06003cbcac5048d522bee0Wget%2f1%2e8%2e2; expires=Tue, 02-Jun-09 20:23:36 GMT; path=/; domain=bbc.co.uk;
--13:23:36-- http://news.bbc.co.uk/2/hi/uk_news/magazine/46067
=> `4606719.stm'
Resolving news.bbc.co.uk... done.
Connecting to news.bbc.co.uk[212.58.240.41]:80... connected.
HTTP request sent, awaiting response...
1 HTTP/1.1 200 OK
2 Date: Fri, 03 Jun 2005 20:23:36 GMT
3 Server: Apache
4 Cache-Control: max-age=0
5 Expires: Fri, 03 Jun 2005 20:23:36 GMT
6 Set-Cookie: BBC-UID=1472caf04b1c04e8e9678f1ef1604aff54f48fcef
7 Connection: close
8 Content-Type: text/html
[ <=> ] 40,846 92.76K/s
13:23:37 (92.76 KB/s) - `4606719.stm' saved [40846]
And all this time, I got my news from the BBC only because news is their primary goal - not selling ad space (unlike CNN, MSNBC, NBC, ABC, oh, hell - every "American" news source).
I'm in the wrong damn country!
"The large print giveth, and the small print taketh away" -- "Step Right Up", Tom Waits
They should really consider removing nations from the EU that don't want to ratify the new constitution.
Why do you hate the rest of the world?
The revolution will NOT be televised.
I'm astounded. Theres been a long debate about the BBC over its history that will probably never go away. Both Labour and the Conservatives have decided its against them and for their foes at different times. As have the Liberals, the Greens, the Scots Nationalists, the parties in Ulster and obviously the socialists and the fascists harbour their own grievances.
I can't think of a more compelling proof that the BBC does its best to put out the truth; after all its the only type of news that could be equipotential in its power to upset politicians of starkly varying political persuasions.
But I must say I'm amazed that the Beeb is selected as "anti-Bush" and the "anti-USA" thing is outrageous. What possible grounds could one have for thinking that?
Plays violent online games as: Nerfherder76
Woah! A big site on the web is based on Apache and Linux! What a news!
Come on. Linux is one of the most popular server OS, and Apache is the number one web server. Come out from the 1993, please.
most people who critisize this article are spoonfed MCSE's who really 'tinker' with the idea of a non-microsoft OS, rather than take the time to actually learn to use it (properly), or reap the rewards it can offer (when properly used).
But that's a good thing too me, I can just laugh at them when their customers come to me instead. =)
the only permanence in existence, is the impermanence of existence.
The BBC seems like a relatively simple site to optimize for heavy loads -- mostly static pages, and probably 99% of the load comes from hits to the top 20 stories.
/ 6257
LiveJournal has to deal with millions of user accounts and jillions of posts, plus photos and audio files. Their database, application-level caching, image/audio storage, and Web serving *all* scale horizontally -- little or no "big iron", just lots of commodity boxen. The result: lots of college kids can post "OMG LOL!1!". *That's* democratizing the Internet.
Anyway, they're a Perl/MySQL shop, and Brad Fitzpatrick gave details on their setup (more recent than his OSCON slides once posted on Slashdot): http://www.mysqluc.com/cs/mysqluc2005/view/e_sess
I'd like to hear information about Amazon, too -- they use Perl and HTML::Mason.
The code behind the CBC "zed" site has been open sourced as well, under the Apache license no less!
STFU about slashdot bias.
The BBC television and radio news are both far and away the best in the world, but the errors that routinely find their way into the online version are shocking.
$ nmap -sS -p 80 -v -O news.bbc.co.uk
[snip]
Running: Sun Solaris 8
OS details: Sun Solaris 8
Uptime 251.064 days (since Sat Sep 25 14:50:31 2004)
[snip]
I used to read Caltizzle. I was a lot cooler than you.
Not much gets past the nerds of /. org
Timing makes Google an April Fool
...but that doesn't mean I'm stupid.
And please, don't let our president influence your stereoty^H^H^H^H^H^H^H^H impression of Americans.
i'm really pleased that half the world is benifiting from a bloody service I pay for with my license fee "tax".
So tell me why if there's an easy way to do something and a complicated way to do something -- even something simple -- IBM will invariably pick the complicated way?
/etc/filesystems in AIX? The traditional fstab or vfstab is easily parsed, and the stanza format gains nothing.
Point in case:
- What possible reason is there to use a "stanza" format for
- Ever hear of someone didn't have a hard time getting Tivoli working properly?
Complicated software sells services.
That said, there are some things IBM has done very well... the disk management subsystem in AIX is second to none.
I always find that when working with IBM software, take how long you think it *should* take, and double it. At least.
Nobody called it that. The reason being that it wasn't the case.
This has also been the subject of a very detailed study:
Far from revealing an anti-war BBC, our findings tend to give credence to those who criticised the BBC for being too sympathetic to the government's pro-war stance. Either way, it's clear to accuse the BBC of an anti-war bias fails to stand up to any serious or sustained analysis
Was the BBC really biased against the war?
A fix for this Apache problem on Fedora: /etc/httpd/conf.d/php.conf: /etc/httpd/modules/libphp4.so into server: /usr/lib/libcurl.so.3: undefined symbol: tld_strerror
e xport LD_LIBRARY_PATH /etc/init.d/httpd and then you can re-enable PHP in /etc/httpd/conf.d/php.conf and do "service httpd restart".
/usr/local/lib that installing Jabber put on your system from the path that Apache searches.
Starting httpd: Syntax error on line 6 of
Cannot load
is to add these two lines:
LD_LIBRARY_PATH=/usr/lib:/usr/lib/mysql
(or edit path as appropriate for your system)
near the beginning of
This removes the libidn files in
Posting here in the hopes that the search engines will find it; don't want to sign in at innumerable sites...