On Fri, Jul 24, 2009 at 11:10 AM, Alexander Klimetschek<[email protected]> wrote: > Gupta<[email protected]> wrote: >> Currently I am making use of RDBMS to keep the binary content. But this >> approach may have performance limitations in production environment. So I >> was contemplating JCR based solution for this where I can get best of both >> worlds i.e. RDBMS for meta data (which I am already maintaining in my >> schema) & file storage (which Jackrabbit can manage for me). > > JCR / Jackrabbit itself is already the best of both worlds (RDBMS and > file systems), as it allows you to store files along with metadata and > you get fulltext search etc. So you could completely switch to JCR as > well :-) >
Undoubtedly, JCR offer you a lot of features such as search, tree style storage and manage the persistence for you. But there are overhead on what you got. If you don't want to take advantage of those features, a simpler implementation could be much more scalable and efficient. Especially most of your content are multimedia in your case. For best performance, those multimedia content is better served from web server instead of app server IMO. --Guo
