Case Study: How Hulu Scaled Serving 4 Billion Videos Using Redis via @gopivotal

Great post from hulu scaliing experience, relying a lot on redis !

  • starting with python + mysql + memcached
  • memcached failing regurlarly wit key loss
  • choose redis over MongoDB, Riak, LevelDB
  • has great doc
  • redis can solve lot of common cases with its data structure
  • use lua !  but mainly for avoid server connectiion trip
  • the lua script make may call on the redis side and return the while results
  • zset with timestamp score for sorting keys is great
  • redis is shared by user keys
  • have as many redis instance by servers as cpu core numbers
  • they use cassandra for write durability, and then load data back to redis when needed
  • do a lot of monitoring

What i think

  • more and more scaling real wolrd examples with redis
  • redis as simple queries and dataset so you know what cost any queries. no more black orm and magic. you give architecture complexity around this
  • i have already read: start with redis then go with cassandra: here they use both and i think its better (cassandra is tricky in prod)
0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *