My Heart Twitters When I See This Chart

Congratulations to the Twitter team on their phenomenal growth.

Twitter’s popularity is exploding. But not their infrastructure costs. They use Accelerators and our Scale services to achieve great performance for the right price.

A twittering heart is a very good thing.

19 Comments

  1. Posted March 10, 2007 at 10:53 pm | Permalink

    How many accelerators?

  2. Posted March 10, 2007 at 11:22 pm | Permalink

    Congrats

  3. Jacob
    Posted March 11, 2007 at 5:38 am | Permalink

    How does twitter make money?

  4. Posted March 11, 2007 at 6:02 am | Permalink

    I just signed up. Interesting service, I’m also interested as to how they make any money. It is pretty well done, though.

  5. Henry Todd
    Posted March 11, 2007 at 9:02 am | Permalink

    I don’t mean to be overly negative here, I use both Twitter and Textdrive and didn’t realise until reading this that there was any connection. That said, Twitter has one of the worst downtime problems of any site I use regularly enough to notice. It’s down right now as I type this for instance. Their IM bots seem to go offline (or be online but unresponsive) a little too much for comfort as well.

    I’m sure like all services these are just the signs of early growth, and all will be fine, but I just thought I’d say that in my experience Twitter is not a good example of a service that’s scaling smoothly.

  6. Posted March 11, 2007 at 1:09 pm | Permalink

    I have to agree – twitter is down a lot.

    Don’t get me wrong, I want to see more and more rails infrastructure success stories but touting about a success before it’s actually a success does more harm that good.

    Get it sorted then shout it out to the world.

  7. Posted March 11, 2007 at 9:24 pm | Permalink

    @Henry/Michael: what clients are you using?

  8. Posted March 11, 2007 at 9:26 pm | Permalink

    @Caleb: it varies. That’s one of the benefits of Accelerators. It’s an on-demand service. But for specifics, I’ll let the Obvious guys tell that story.

  9. Posted March 12, 2007 at 2:59 am | Permalink

    @David: I’m using Adium. Have tried both AIM and Gtalk. Although my comment about downtime is for web moreso.

    I can’t get the IM bot to show online at all… not sure what’s going on there.

  10. Henry Todd
    Posted March 12, 2007 at 10:03 am | Permalink

    David: I am also using Adium (1.0) to connect to the AIM bot. To be fair, it’s been a few days since I recall the bot being offline when I wanted to send a message. How much do the bots differ? Does one have a better uptime record than the others?

  11. Posted March 12, 2007 at 2:42 pm | Permalink

    @Michael/Henry: the sheer number of people using bots means that the clients (Adium, Twitterific, etc.) don’t account for the timeouts properly. If your bot isn’t working, the web page will.

  12. Posted March 12, 2007 at 9:56 pm | Permalink

    I hate to join on the negativity-band-wagon but the site has been down intermittently for me for the past 5 minutes and when it did finally load, it took over 90 seconds for the page to load after logging in.

    I think it’s great that Twitter has pro-actively tackled scalability but I think the phrase “great performance” in your entry needs a just a tiny reality check ;-)

  13. Posted March 12, 2007 at 10:55 pm | Permalink

    I have to agree with the voices here. I love Twitter and use it all the time (user:drewm), but I wouldn’t cite it as an example of a service free from scalability problems.

    Twitter is down or unacceptably slow a lot. We don’t care because we love it, it’s not important data, and hey, it’s free. Also we all know they’re growing at a massive rate. Hopefully those problems will get ironed out.

    It’s great that it’s not costing them a bomb, but let’s not get carried away and pretend there are no problems.

  14. Greg
    Posted March 13, 2007 at 4:13 am | Permalink

    And has anyone yet figured out how Twitter plans to continue operating since it’s unclear how they are generating revenue.

  15. Posted March 13, 2007 at 2:33 pm | Permalink

    I think some are missing the point. Twitter has been able to grow (service interruptions aside) because of Accelerators. Accelerators have been combined with our Scale services to help Twitter achieve the great performance (4000+ requests per second) that has allowed them to serve more customers. Have things been perfect? No. But Twitter isn’t complaining. They know what’s been going on behind-the-scenes to achieve their growth on Accelerators. I’ll work to get one of them on our podcast to talk about how it was done.

  16. Henry Todd
    Posted March 13, 2007 at 5:42 pm | Permalink

    David: I’m sure I’m not the only one that would like to hear from the Twitter people about how they’ve been handling their evidently very rapid growth and keeping costs down, the podcast is a great idea.

    I also echo Drew’s sentiment: “We don’t care because we love it”. :-)

  17. Henry Offenganger
    Posted March 28, 2007 at 6:14 pm | Permalink

    Yeah… have to agree with the sentiment here…. Twitter is balls slow.

    I’m sure things will work out in the end but clearly they are having scaling troubles. This post makes you sound like the old GWB.

  18. asdf
    Posted April 3, 2007 at 12:26 am | Permalink

    asdf

  19. Cheryl Merkowski
    Posted April 11, 2007 at 4:09 pm | Permalink

    Twitter is sooooo slow! It is down almost every day… one would think that they would fix these problems—they have been ongoing for quite awhile now.


Post a Comment

You must be logged in to post a comment.
Follow

Get every new post delivered to your Inbox.

Join 49 other followers