Ben Laurie blathering

Splash! Startup Fixed

For those that don’t know, Splash! is a distributed masterless database, based on Spread, which is a group communication system.

Anyway, it has long suffered from a problem: if the database gets big, then when a new instance of Splash! starts it has a tendency to die. This is because Spread will simply drop a client that sends messages too fast. Wrongheaded, IMNSHO, but then, Spread does a fair few wrongheaded things. As of yesterday, its fixed: what I did in the end was make the database refresh over a unicast link. The sender of the refresh forks so they don’t block, and the receiver doesn’t care about blocking, coz its dead in the water until it gets refreshed anyway.

Seems to work.

I have a long-term ambition to replace both Spread and Splash! with something much cooler. So do other people I know – but I’d love to hear from people with similar ambitions.


  1. Cool!

    Comment by Tony Finch — 16 Sep 2005 @ 15:26

  2. Wet!

    Drop in next time you’re in the Welsh borders for a glass of wine and we could discuss your ideas.. sounds very intersting…

    Comment by Argon — 16 Sep 2005 @ 17:59

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

Powered by WordPress