About Archive Tag Cloud Translations RSS
  1. Time to Load Redis Snapshots

    by Will Larson
    August 23, 2010 redis

    Here are some quick notes on the time to load Redis snapshots, perhaps useful when investigating Redis as an architectural component.

  2. Notes on Redis Memory Usage

    by Will Larson
    August 22, 2010 redis

    Yesterday I spent some time running memory usage tests on Redis. Specifically I was interested in storing roughly comparable data in the different data structures and comparing costs. This article contains the numbers and notes from that process.

    Comment by antirez on August 26, 2010

    Hey, there are a few very important changes to this memory benchmark to actually spot the interesting memory optimizations:

    1) Also run all the tests against Redis master. You should see moderate improvements for all the data types and huge improvements for the list data type. 2) Make sure for Hashes and Lists to test with more granularity, for instance hashes (and also lists in Redis master) of 10 / 20 elements will take 5 times less memory than you would expect just dividing your 100 items figure. 3) Editing redis.conf you can change the waterlevel for zipmap->hash conversion. Read my latest post on antirez.com for more information.

  3. SuperFamily Relationships with Lazyboy

    by Will Larson
    May 23, 2010 python Cassandra lazyboy

    In this article I take a look at using Cassandra and Lazyboy for modeling a trivial task management application. Intended to contain examples of iterating over a subset of keys and storing relations between SuperColumns.

  4. Tailing in Python

    by Will Larson
    May 16, 2010 python

    A quick and pointless look at implementing tail in Python. Something of a koan.

  5. Callbacks Are An Informal Pipeline

    by Will Larson
    March 7, 2010 javascript Node.js

    Spending some time with Node.js I've begun to realize what a pain callbacks can be when they become sufficiently deep. That said, after sufficient nesting it's pretty easy to see how callbacks become very similar to a pipeline (in the simplest case, anyway).