News for hackers. Stuff that matters.
Re:How Node.js has revolutionised the Mail Online (Worlds No... by petercooper on Monday November 25, 2013 @08:07AM (#6793817)
The entire backlog of 1.2 million articles are stored in elastic-search, which is used to drive the site

Elasticsearch is awesome and I'd heard rumors of people using it as a data store rather than just a search index but it's awesome to see it being used as a data store at such scale.

Re:How Node.js has revolutionised the Mail Online (Worlds No... by lost_my_pwd on Monday November 25, 2013 @08:27AM (#6793911)
At the last place I worked, we used ES for both search and storage of 2.5MM+, and quickly growing, items (the data was also stored in DynamoDB for safety). It worked quite well, saved us some development time, and reduced the latency of search requests a little since there was no 2nd network call to fetch the documents found by ES from separate storage.
Re:How Node.js has revolutionised the Mail Online (Worlds No... by petercooper on Monday November 25, 2013 @10:25AM (#6794491)
What are the basic tradeoffs at that scale? I'm only using it for search at a very low scale but it almost sounds like a dream database if it works well at scale and for canonical storage.