If you plan to support all the RDBMS supported by Carbon you will have to write lot of db specific code. For example Oracle, DB2 and PostgreSQL will have lot of custom code. Using an ORM will help IMO. But I'm not much aware of the performance impact.
Regards, Prasad. On Thu, Jul 17, 2014 at 12:30 PM, Hasitha Hiranya <[email protected]> wrote: > Hi, > > If we use JPA, how will it affect the performance, and maybe database > level tuning? > > Anyway it will stop us writing MySQLBased Messagestore, > OracleBasedMessageStore, MSSQLBasedMessageStore, SQLLiteBasedMessageStore > etc. We will have only a RDBMSBasedMessageStore. > > Thanks > > > On Thu, Jul 17, 2014 at 12:16 PM, Dhanuka Ranasinghe <[email protected]> > wrote: > >> Please find my inputs for above questions inline. >> >> *Dhanuka Ranasinghe* >> >> Senior Software Engineer >> WSO2 Inc. ; http://wso2.com >> lean . enterprise . middleware >> >> phone : +94 715381915 >> >> >> On Thu, Jul 17, 2014 at 9:52 AM, Hasitha Hiranya <[email protected]> >> wrote: >> >>> Hi, >>> >>> Moving to Architecture@ >>> >>> Thanks >>> >>> >>> On Thu, Jul 17, 2014 at 6:59 AM, Srinath Perera <[email protected]> >>> wrote: >>> >>>> We need to d a initial design before the discussion. >>>> >>>> Please move this to arch@ >>>> >>>> >>>> On Wed, Jul 16, 2014 at 10:25 PM, Shammi Jayasinghe <[email protected]> >>>> wrote: >>>> >>>>> Hi Hasitha, >>>>> >>>>> I think , it would be great to have a meeting on this with all the >>>>> relevant people and come up with an idea on the design of the database >>>>> after explaining current data storage pattern in Cassandra. Then we ll be >>>>> able to verify that with wider audience and proceed further. >>>>> >>>>> Could you please schedule a meeting on this. >>>>> >>>>> Thanks >>>>> Shammi >>>>> >>>>> >>>>> On Wed, Jul 16, 2014 at 9:52 PM, Hasitha Hiranya <[email protected]> >>>>> wrote: >>>>> >>>>>> Hi All, >>>>>> >>>>>> We are ready to do the $Subject with more clean APIs etc now. >>>>>> We would like to get ideas from DB experts at WSO2 on the $Subject. >>>>>> >>>>>> We need to discuss how to >>>>>> >>>>>> 1. Save metadata >>>>>> 2. Save content chunks of messages >>>>>> >>>>> For this we will need to create proper database schema >> >>> 3. Read message metadata as groups of messages (between msg_id x and >>>>>> y) >>>>>> >>>>> Is it something like SQL BETWEEN? >> >>> 4. How to move message metadata efficiently >>>>>> >>>>> Better maintain status rather moving data across tables/databases >> >>> 5. How to deal with limited number of columns in RDBMS as opposed to >>>>>> Cassandra >>>>>> >>>>> Same schema structure we have applied to CQL can be used for RDBMS. >> >>> 6. Cassandra automatically order messages by message ID? What about >>>>>> RDBMS? >>>>>> >>>>> Is it something like SQL Order By? Indexing will help to optimize this >> >>> >>>>>> etc. >>>>>> >>>>>> As the first step we are going to embed H2 as the in-memory database. >>>>>> Thus we have following concerns. >>>>>> >>>>>> 1. Can we use H2 embedded in Carbon servers now. As it is file based >>>>>> can the same be used for memory based or should it be different? >>>>>> >>>>> We can try out >> http://www.cowtowncoder.com/blog/archives/2010/03/entry_374.html >> >>> 2. As I understand the best way to create a connection with the DB is >>>>>> to use masterDataSources.xml file and let the carbon framework to create >>>>>> the connection for us. Can we get the DB connection into a dependency >>>>>> using >>>>>> a look up by JNDI name? Is this possible for inmemory H2? >>>>>> >>>>> Actually ndatasource will register Datasource as JNDI entry, >> technically we should be able to regardless of inmemory or not. >> >>> 3. Is H2 the best DB for in-mememory implementation? Why H2? >>>>>> >>>>> There are other options as well, like JavaDB/Derby, worth to try out. >> >>> 4. We need to design this as there will be no drastic changes in >>>>>> on-coming releases. Otherwise we will need complex Data migration >>>>>> scripts. >>>>>> >>>>> How do we gonna support multiple databases? Any thought to use ORM >> framework ? If use JPA , provider will take care about schema generation no >> need of scripts. >> >>> >>>>>> Thanks >>>>>> >>>>>> -- >>>>>> *Hasitha Abeykoon* >>>>>> Senior Software Engineer; WSO2, Inc.; http://wso2.com >>>>>> *cell:* *+94 719363063* >>>>>> *blog: **abeykoon.blogspot.com* <http://abeykoon.blogspot.com> >>>>>> >>>>>> >>>>> >>>>> >>>>> -- >>>>> Best Regards, >>>>> >>>>> * Shammi Jayasinghe* >>>>> Associate Tech Lead >>>>> WSO2, Inc.; http://wso2.com, >>>>> mobile: +94 71 4493085 >>>>> >>>>> >>>> >>>> >>>> -- >>>> ============================ >>>> Director, Research, WSO2 Inc. >>>> Visiting Faculty, University of Moratuwa >>>> Member, Apache Software Foundation >>>> Research Scientist, Lanka Software Foundation >>>> Blog: http://srinathsview.blogspot.com twitter:@srinath_perera >>>> Site: http://people.apache.org/~hemapani/ >>>> Photos: http://www.flickr.com/photos/hemapani/ >>>> Phone: 0772360902 >>>> >>> >>> >>> >>> -- >>> *Hasitha Abeykoon* >>> Senior Software Engineer; WSO2, Inc.; http://wso2.com >>> *cell:* *+94 719363063* >>> *blog: **abeykoon.blogspot.com* <http://abeykoon.blogspot.com> >>> >>> >> > > > -- > *Hasitha Abeykoon* > Senior Software Engineer; WSO2, Inc.; http://wso2.com > *cell:* *+94 719363063* > *blog: **abeykoon.blogspot.com* <http://abeykoon.blogspot.com> > > > _______________________________________________ > Architecture mailing list > [email protected] > https://mail.wso2.org/cgi-bin/mailman/listinfo/architecture > > -- Prasad Tissera Software Engineer. Mobile : +94777223444
_______________________________________________ Architecture mailing list [email protected] https://mail.wso2.org/cgi-bin/mailman/listinfo/architecture
