Also you forget the fact that normally to produce the kind of complex document that could be stored in as a single record in couchdb involves query many tables in mysql (or other relational database) not only one.
> However, we think that this difference could be overcome by using CouchDB on a heavily clustered environment.
Let me clarify this: you're suggesting to buy more than fifty hardware servers to replace a single MySQL one :)
But, of course, it depends on the application. I.e. noone will use MySQL for Twitter... It's unfair to compare such databases. You can compare memcachedb with CouchDB the same way.
Glad to finally see some performance comparisons. It would be nice to compare other types of queries as well to see how it performs on things such as date ranges, string comparisons, etc.
Also you forget the fact that normally to produce the kind of complex document that could be stored in as a single record in couchdb involves query many tables in mysql (or other relational database) not only one.
RispondiElimina> However, we think that this difference could be overcome by using CouchDB on a heavily clustered environment.
RispondiEliminaLet me clarify this: you're suggesting to buy more than fifty hardware servers to replace a single MySQL one :)
But, of course, it depends on the application. I.e. noone will use MySQL for Twitter... It's unfair to compare such databases. You can compare memcachedb with CouchDB the same way.
Glad to finally see some performance comparisons. It would be nice to compare other types of queries as well to see how it performs on things such as date ranges, string comparisons, etc.
RispondiEliminaIt's a pitty that you did not compare the Results of inserting, indexing and querying for 500K, 1M, and 10M MySQL server as well.
RispondiElimina