Thanks. But, it seems to me that Twitter does a fine job for now, its gone out for a few minutes here and there, but it gets the tweets across. Revou.com sounds more like a new "facebook" creator than just Tweeter. Its cool stuff though, have you used it?
only mentioned Revou as all these little scripts are a joke really.
if someone used that, and traffic took off, wonder what they'd
actually do? give up probably. you need to get to grips with
not just stuff like memcache which is easy enough, but how to
handle persistent connections, else you'd run out of servers
with all the socket opens/closes/waits. Look like they've come
up with 'starling', not too different in philosophy from the shmproxy
yahoo developed back in late 1990's for ads, and everything else.
mind you, when I jumped ship two years ago, we were peaking at
around 500M ads/day in europe, and US was magnitudes larger.
it focuses the mind.
never used twitter either, still can't see the point in it all!
If anyone made Revou really work for them, they probably would be smart enough to set up something more long-term. But why re-invent the wheel?
I didnt get Twitter either, until i tried to search in it. Sear.twiter dot com. You can find conversations that are going on in cyberspace. Pick a word, like microsoft, or a city name, ect.. you can actually start connecting with people about topics that they care about. Right now we are actually driving a bit of business through Twitter because we can listen. We listen to what people want, what they are looking for, what they dont like etc.. Its a great tool for listening, which is, in my opinion, the larger part of selling and keeping a customer.
I'd assume places like twitter have dozens of servers in different datacenters. Downtime for them really would cost a lot for real, it's not like when a $5 hosting customer says they are losing hundreds by the minute.
I know that they have servers everywhere. I'm more curious about how many tweets accounts can a server generally handle. I would assume they do grid computing for the entire thing though... I guess It's all probably pretty secretive.
I actually don't think that under their current no-income model it would cost much. Even so, it just means that a tweet is delivered a few minutes late. Vs. a website where not finding a page can get pretty annoying.
IMHO It became popular because it finally allowed for blog and facebook-like communications to become conversations. You can search any keyword and twitter and find people talking about it, then reply back, and voila! a conversation is begun. Its really hard to do that with a blog, or a FB or MS account. I use tweetdeck, it constantly searches up to 10 keywords of my interest and I can reply in real-time to someone in Turkey who mentions something about servers i'm interested in. When I began searching like that, Twitter suddenly made sense. A lot of sense.
Tell me if there is a twitter like site that is as successful as twitter itself?
Tell me a Microsoft Exchange clone that is as successful as MS Exchange itself? There isn't. That doesn't mean there isn't a market for it though.
If you setup a Twitter clone which is exactly the same as Twitter, there is indeed not a point in doing so. However you could for example setup a twitter for a specific branche, hobby or activity, or you could make a Twitter for a specific region, and so on.
Twitter is just generic. If you try to compete with a successful generic product, you will indeed fail. That's why you don't make a 100% identical clone, but you add a touch, which could make it successful.
Natwest Streamline guys still use an IBM mainframe (or did last time I spoke to them a few years back) - wonder how many cc billing transactions they process overnight, every night, from Tesco, Sainsburys, Asda, and just about every major department store and shopping chain in UK.
Come to think of it, it's probably just idling.
The IBM 3090 I used at Glasgow Uni yonks ago, had so many flops, it was unbelievable, then along came the cheap and nasty technology we all
The issue with mainframes is that apps you write for it don't tend to be very portable or very scalable (at least within decent time frames). Writing an application for a cluster of servers is a much better paradigm.
What do you do if you find your mainframe server can't handle the load? Buy a new mainframe? What do you do with the original... it's pretty wasted...
Instead, if you write an application to use a cluster of servers, you can simply at additional capacity at will.
Darren Lingham - UK Webhosting Ltd. - (0800) 024 2931 Tsohost.co.uk - Quality UK Windows and Linux hosting since 2003 UK WordPress Hosting - Fast, easy, cloud based WordPress Hosting
And they are using Amazon's AWS S3 for storage of images like profile backgrounds and others. It's all on scalable utility computing. So really can't guess how much resources it needs unless you are a Twitter employee
# Over 350,000 users. The actual numbers are as always, very super super top secret.
# 600 requests per second.
# Average 200-300 connections per second. Spiking to 800 connections per second.
# MySQL handled 2,400 requests per second.
# 180 Rails instances. Uses Mongrel as the "web" server.
# 1 MySQL Server (one big 8 core box) and 1 slave. Slave is read only for statistics and reporting.
# 30+ processes for handling odd jobs.
# 8 Sun X4100s.
# Process a request in 200 milliseconds in Rails.
# Average time spent in the database is 50-100 milliseconds.
# Over 16 GB of memcached.
Well one of my associates owns a website that had a link to it through a youtube vid. Some guy in germany posted his video with its link into a "tweet" and the traffic to my associate's site went crazy! If twitter can generate the kind of traffic he got, then I think that it's got have one insanely powerful server.