performance - How does CouchDB perform for a regularly updated dataset? -


I am planning to use CouchDB on a project. But as the questionnaire system involves writing ideas (which are very much like the sequence of regular RDMMS) I was thinking that if the document database is updated very much (write heavy writing), then the CEHDB is better than regular RDBMS Performs the way? Or should we have to compact / re-index the system occasionally so that it can be managed faster?

You can think of the professionals / ideas of the visual model. (CouchDB hackers may disagree but IMO is accurate enough for this user.)

  1. A visible function always does a complete "table scan" when it is created first (an RDBMS As BTW)
  2. They do not have any side effects, reduce mappings and functions, can be arbitrarily complex
  3. reduce each document and map / result Cache is done and never counted
  4. If you Adding or changing a document, this (and only then) will be re-calculated (and cached) for that view

Looking at them, you get some conclusions about CochinBee performance

  • There is no re-index phase for the entire data set, just incremented according to the documentation update
  • Changing a view function index
  • Since both CouchDB and RDBMS are indexed for new data , So it is worth thinking that performance will be very similar to the use of heavy update / pouring.

Obviously YMMV and standard police-out, "You must test your own weight" applies though I will add some more ideas.

  • I say that RDBMS is better for the query-style of your data inquiries. When you also do not know which questions to ask for your data, you do not actually structured for the language query Can beat
  • However, once you define what you want to know, CouchDB (and perhaps Hadop) provides the most rich query system because you are just writing code.
  • If your data set is large, the NoSQL database will be scaled more easily. For example, allows a cluster of coaches for parallel processing. Hadop does this, then it will come in secondary ideas: familiar, maintenance, Sokdev is a web server, but a bit more DIY is required; Hadoop internalizes more cluster management at the cost of complexity, foreign, etc.

I hope that helps to put some light on your decision!


Comments

Post a Comment