You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@airavata.apache.org by Suresh Marru <sm...@apache.org> on 2011/07/12 02:46:35 UTC
Embedded Database choice for Airavata
Hi All,
To refresh us all, we discussed earlier on this list about mysql alternatives [1] and the conclusion was to go with embedded database for default configuration and provide documentation on how to change the database to mysql at deployment time. Derby was discussed as a viable option, but before we jump into implementation I want to solicit feedback from every one. I do not have any recent first-hand experience with any of them, so my vote will be +1 to who ever convinces me that they are convinced about one over other :)
The three most popular in-memory databases Derby, H2, HSQLDB all have license compatibilities. db40 is GPL. Java DB seems to be based on Derby and ALV2 but could not figure out why we should use Java DB and why not just the Derby.
In any case, do folks have any practical experiences and preferences over any option? Just note that Airavata has more scalability and reliability requirements over performance.
Cheers,
Suresh
[1] - http://markmail.org/thread/3bzlnwr75mylh7mu
Re: Embedded Database choice for Airavata
Posted by "Mattmann, Chris A (388J)" <ch...@jpl.nasa.gov>.
+1 to checking out Derby. I haven't used it extensively, but have heard
that it scales...
Cheers,
Chris
On Jul 12, 2011, at 2:02 AM, Ross Gardler wrote:
> On 12 July 2011 01:46, Suresh Marru <sm...@apache.org> wrote:
>
>> In any case, do folks have any practical experiences and preferences over any option? Just note that Airavata has more scalability and reliability requirements over performance.
>
> The only practical experience I have of Derby is as the default
> database for Apache Wookie. As far as I am aware it has not been
> tested at scale in this project, certainly I have not done so.
> However, it is I that did the migration from MySQL to Derby and I
> found it an easy process with no surprises around the corner.
>
> In general I would recommend using an Apache project over another
> third party project. Mainly because it's easier to be heard if we need
> to engage with the project community. Over the coming years Airavata
> contributors are almost certain to find themselves at the same events
> as Derby contributors. Furthermore, even though merit does not
> transfer from one ASF project to the next, it does send a strong
> signal that you are a real open source contributor and thus are likely
> to attract attention to your communications with the project (not that
> Wookie has needed to engage with the Derby community in its use).
>
> Ross
>
> --
> Ross Gardler (@rgardler)
> Programme Leader (Open Development)
> OpenDirective http://opendirective.com
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Chris Mattmann, Ph.D.
Senior Computer Scientist
NASA Jet Propulsion Laboratory Pasadena, CA 91109 USA
Office: 171-266B, Mailstop: 171-246
Email: chris.a.mattmann@nasa.gov
WWW: http://sunset.usc.edu/~mattmann/
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Adjunct Assistant Professor, Computer Science Department
University of Southern California, Los Angeles, CA 90089 USA
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Re: Embedded Database choice for Airavata
Posted by Lahiru Gunathilake <gl...@gmail.com>.
Hi all,
I have some performance comparison experience with Derby vs H2 but I
strongly agree with Ross's point to use another Apache project rather moving
to another open source project unless there is a very big difference.
I believe going with Derby will be a nice approach !
Lahiru
On Tue, Jul 12, 2011 at 5:02 AM, Ross Gardler <rg...@opendirective.com>wrote:
> On 12 July 2011 01:46, Suresh Marru <sm...@apache.org> wrote:
>
> > In any case, do folks have any practical experiences and preferences
> over any option? Just note that Airavata has more scalability and
> reliability requirements over performance.
>
> The only practical experience I have of Derby is as the default
> database for Apache Wookie. As far as I am aware it has not been
> tested at scale in this project, certainly I have not done so.
> However, it is I that did the migration from MySQL to Derby and I
> found it an easy process with no surprises around the corner.
>
> In general I would recommend using an Apache project over another
> third party project. Mainly because it's easier to be heard if we need
> to engage with the project community. Over the coming years Airavata
> contributors are almost certain to find themselves at the same events
> as Derby contributors. Furthermore, even though merit does not
> transfer from one ASF project to the next, it does send a strong
> signal that you are a real open source contributor and thus are likely
> to attract attention to your communications with the project (not that
> Wookie has needed to engage with the Derby community in its use).
>
> Ross
>
> --
> Ross Gardler (@rgardler)
> Programme Leader (Open Development)
> OpenDirective http://opendirective.com
>
Re: Embedded Database choice for Airavata
Posted by Suresh Marru <sm...@cs.indiana.edu>.
On Jul 12, 2011, at 5:02 AM, Ross Gardler wrote:
> On 12 July 2011 01:46, Suresh Marru <sm...@apache.org> wrote:
>
>> In any case, do folks have any practical experiences and preferences over any option? Just note that Airavata has more scalability and reliability requirements over performance.
>
> The only practical experience I have of Derby is as the default
> database for Apache Wookie. As far as I am aware it has not been
> tested at scale in this project, certainly I have not done so.
> However, it is I that did the migration from MySQL to Derby and I
> found it an easy process with no surprises around the corner.
This is very good to hear.
> In general I would recommend using an Apache project over another
> third party project. Mainly because it's easier to be heard if we need
> to engage with the project community. Over the coming years Airavata
> contributors are almost certain to find themselves at the same events
> as Derby contributors. Furthermore, even though merit does not
> transfer from one ASF project to the next, it does send a strong
> signal that you are a real open source contributor and thus are likely
> to attract attention to your communications with the project (not that
> Wookie has needed to engage with the Derby community in its use).
+ 1 for all of above.
I found interesting differences when I googled through. The third party comparisons widely varied and nothing conclusive [1][2][3]. As expected each of the database claimed its the best [4][5][6].
So I do not have any bias and Ross's and Lahiru's arguments are strong enough for me to pick Derby. Since I did not see any objections to Derby or other suggestions, I will not call for a vote. I will create JIRA tickets for this porting.
Suresh
[1] - http://www.encorewiki.org/display/encore/Open+Source+Databases+Comparison
[2] - http://database-management-systems.findthebest.com/compare/6-15-16/Apache-Derby-vs-HSQLDB-vs-H2
[3] - http://www.sukhvinder.co.uk/articles/design-concepts/database-comparison-table-h2-hsqldb-derby-postgresql-mysql
[4] - http://www.h2database.com/html/performance.html
[5] - http://db.apache.org/ojb/docu/guides/performance.html
[6] - http://hsqldb.org/images/imola_retrieve.jpg
Re: Embedded Database choice for Airavata
Posted by Ross Gardler <rg...@opendirective.com>.
On 12 July 2011 01:46, Suresh Marru <sm...@apache.org> wrote:
> In any case, do folks have any practical experiences and preferences over any option? Just note that Airavata has more scalability and reliability requirements over performance.
The only practical experience I have of Derby is as the default
database for Apache Wookie. As far as I am aware it has not been
tested at scale in this project, certainly I have not done so.
However, it is I that did the migration from MySQL to Derby and I
found it an easy process with no surprises around the corner.
In general I would recommend using an Apache project over another
third party project. Mainly because it's easier to be heard if we need
to engage with the project community. Over the coming years Airavata
contributors are almost certain to find themselves at the same events
as Derby contributors. Furthermore, even though merit does not
transfer from one ASF project to the next, it does send a strong
signal that you are a real open source contributor and thus are likely
to attract attention to your communications with the project (not that
Wookie has needed to engage with the Derby community in its use).
Ross
--
Ross Gardler (@rgardler)
Programme Leader (Open Development)
OpenDirective http://opendirective.com